Description
The price variable is among the most powerful instruments in the arsenal of the executives to achieve entry deterrence objectives. There are two main pricing strategies that firms may use to defend against a competitive market entry.
THE ROLE OF PRICING STRATEGY IN MARKET DEFENSE
A Dissertation
Presented to
The Academic Faculty
by
Can Uslay
In Partial Fulfillment
of the Requirements for the Degree
Doctor of Philosophy in Management
Georgia Institute of Technology
April, 2005
THE ROLE OF PRICING STRATEGY IN MARKET DEFENSE
Approved by:
Dr. Naresh K. Malhotra, Advisor
College of Management
Georgia Institute of Technology
Dr. Fred C. Allvine
College of Management
Georgia Institute of Technology
Dr. Pat H. Dickson
College of Management
Georgia Institute of Technology
Dr. Jagdish N. Sheth
Goizueta School of Business
Emory University
Dr. Richard D. Teach
College of Management
Georgia Institute of Technology
Date Approved: April 6, 2005
iii
ACKNOWLEDGEMENT
This work is dedicated to my parents, Dr. Yuksel and Nezihe Uslay.
I would like to express my heart-felt thanks to all of my dissertation committee
members. This dissertation would not have been possible without the ample guidance and
support from Professors Fred C. Allvine, Naresh K. Malhotra, Richard D. Teach, Pat H.
Dickson, and Jagdish N. Sheth. I cannot thank them enough, but let me attempt to capture
my gratitude with a few words here. Every student is lucky to have a mentor but I was
lucky to have not one but five of them!
Dr. Dickson provided the kind of constructive and useful feedback every doctoral
candidate needs. Dr. Teach always had time for me when I needed guidance and advice.
Dr. Allvine’s enthusiasm on the use of price in competitive interaction was critical for my
choice of the dissertation topic, and his industry contacts proved crucial in securing the
data. He has always trusted, supported me and served as a role model. Every minute spent
with the great Dr. Sheth has been inspiring and his vision influenced my work and life in
more ways than even he could imagine. Last and certainly the most, I would like to thank
my advisor, Dr. Malhotra. He simply is a true mentor.
I would also like to thank Dr. Goutam Challagalla, Nancy Wong, Charles Parsons,
Robert G. Hawkins, Francis M. Ulgado, Alka Citrin, Koert van Ittersum, and my friends
Tracey King, Brian Murtha, Leslie Vincent, Lan Wu, and Tim Quey who have provided
useful comments, feedback and camaraderie along the way. Naturally, I take full credit
for any errors that may remain in this work.
iv
TABLE OF CONTENTS
ACKNOWLEDGEMENT iii
LIST OF TABLES viii
LIST OF FIGURES xi
SUMMARY xiii
CHAPTER 1: INTRODUCTION 1
CHAPTER 2: THEORETICAL BASE AND CONCEPTUAL
UNDERSTANDING 5
2.1 Competitive Signaling 5
2.2 Resource Advantage (R-A) Theory 8
2.3 Game Theory 11
2.4 Marketing Strategy 13
2.4.1 Reputation effects from a marketing perspective 16
2.4.2 Inter-type and intra-type competition 22
2.4.2.1 Mutual Forbearance 24
CHAPTER 3: AN INQUIRY OF THE USE OF PREDATORY PRICING IN
MARKET DEFENSE: EVOLUTION, REVIEW,
AND MARKETING SYNTHESIS 27
3.1 Introduction to the Review 27
3.2 Pricing 32
3.2.1 Predatory Pricing 36
3.2.2 Predatory Pricing versus Predation 39
3.3 Evolution of the Theory and Practice of Predatory Pricing 41
3.3.1 A Historical Perspective 41
3.3.2 Regulatory Acts Regarding Predatory Pricing 45
3.3.3 Early Years and the Populist Era (1890-1975) 48
3.3.4 Areeda-Turner Era (1975-1982) 49
v
3.3.4.1 Counter Cost and Non-cost Proposals 51
3.3.5 Augmented Average Variable Cost Rule Era (1982-1992) 55
3.3.6 Recoupment Requirement Era (1993- ) 56
3.3.6.1 Brooke Case 57
3.3.7 The Emerging School of Thought 59
3.3.8 Differing Practices among States 62
3.3.9 Insights from European Union Regulation 63
3.4 Critical Cases for the Evolution of Case Law 64
3.4.1 Standard Oil (1911) 65
3.4.2 Utah Pie (1967) 66
3.4.3 Matsushita (1986) 67
3.4.4 Rose Acre (1989) 68
3.4.5 Brooke (1993) 69
3.4.6 American Airlines (2001) 71
3.5 Legal Elements of Proof 73
3.6 The State of the Debate on Predatory Pricing 78
3.6.1 The Chicago School 78
3.6.2 The Post-Chicago School of Thought 84
3.7 Synthesis and a Marketing Perspective 92
3.7.1 The Need for a Marketing Perspective 92
3.7.2 A Focus on Deregulated Industries 93
3.7.3 Epilogue 95
3.7.4 Reaction Framework for Competitive Price Reductions 104
CHAPTER 4: NETWORK PRICE COMPETITION FRAMEWORK 107
4.1 Network Industries 111
4.2 Hypotheses 114
4.2.1 Pre-emptive Defense 114
4.2.1.1 Market Power 114
4.2.1.2 Barriers to Entry 118
4.2.1.3 Strategic Assessment 121
4.2.2 Entry Decision 122
4.2.3 Post-Entry Defense 126
4.2.4 Exit Decision 128
4.2.5 Policy Consequences 129
4.2.5.1 Consumer Welfare 131
vi
CHAPTER 5: CONTEXT: AN INQUIRY OF THE DYNAMICS OF
COMPETITION IN THE U.S. AIRLINE INDUSTRY 137
5.1 Historical Perspective 138
5.1.1 The Propeller (Early) Era 140
5.1.2 The Jet (Post-war) Era 146
5.1.3 Deregulation Era 147
5.1.4 Consequences 151
5.1.4.1 Service and Quality 152
5.1.4.2 Fare Levels 155
5.1.4.3 Market Concentration and Power 159
5.1.4.4 Hub-and-Spoke Networks 161
5.1.4.5 Hub Premiums 164
5.1.5 Marketing 167
5.1.5.1 Yield (Revenue) Management 167
5.1.5.2 Frequent Flier Programs 169
5.1.5.3 The Rise (and Fall) of Travel Agents 170
5.1.5.3.1 Computer Reservation Systems (CRS) 172
5.1.5.4 Discounting and Competition 174
5.1.5.5 Price Wars 181
5.2. Competitive Outlook 185
5.2.1 Industry Snapshot 185
5.2.2 The Role of Barriers to Entry 196
5.2.3 After September 11, 2001 200
5.2.3.1 Anticipating Trends 202
5.2.4 European Landscape 206
CHAPTER 6: DATA AND METHODOLOGIES 208
6.1 Data Set 208
6.2 Methodologies 209
6.2.1 Logit Models 210
6.2.2 Multi-Level Mixed Coefficient Model (HLM) 212
6.2.3 Event-History Analysis 215
6.3 Measures 219
CHAPTER 7: DATA ANALYSIS AND RESULTS 224
CHAPTER 8: CONTRIBUTIONS 242
8.1 Managerial Implications 246
vii
8.2 Public Policy Implications 248
8.3 Limitations and Future Research 251
8.4 Conclusion 256
APPENDIX A: Survey of State Law regarding Predatory Pricing 258
APPENDIX B: Comparison of HHI and Proposed MP Index 260
APPENDIX C: U.S. Regional and Major Airline Hubs 265
APPENDIX D: Gate Usage at Large Hubs 1998 266
APPENDIX E: Transformations and Other Analyses 267
APPENDIX F: HLM Analysis Supplements 272
APPENDIX G: Logit Analysis Statistics 274
APPENDIX H: Event-History Analysis Supplements 276
APPENDIX I: Overview of Relevant Empirical Research in the
Airline Industry 280
APPENDIX J: Graphical Representation of Inter-Type Competition 288
REFERENCES 293
viii
LIST OF TABLES
Table 2.1: The Pedigree of Resource Advantage Theory 8
Table 2.2: Characteristics of Intra-type versus Inter type Competition 23
Table 2.3: Non-competitive Pricing among Intra-type Competitors
in the Airline Industry 24
Table 3.1: A Summary of the Evolution of Predatory Pricing in the U.S. 42
Table 3.2: Criticisms regarding the Areeda-Turner rule and the Chicago School 96
Table 4.1: Dimensions of Consumer Welfare and Marketing Insights 131
Table 4.2: Summary of the Framework Premises 136
Table 5.1: The U.S. Airline Industry Timetable 139
Table 5.2: Airline Mergers in the 1985-87 Period 160
Table 5.3: Single Carrier Market Shares at Major Airports
Pre and Post Deregulation 163
Table 5.4: Changes in Hub Premiums over Time 165
Table 5.5: Sample HHI Index for the 10 largest US Airports 1985-1996 166
Table 5.6: A Comparison of Major Frequent Flier Programs 169
Table 5.7: 1999 Market Share for CRS 173
Table 5.8: Average Fares in Markets With and Without Low-Fare Competition 178
Table 5.9: Average Fares Sorted by Distance 178
Table 5.10: Atlanta Markets with AirTran Entry 179
Table 5.11: Atlanta Markets with AirTran Exit 179
Table 5.12: Airports with Most U.S. Destinations Served by Low Fare Airlines 180
ix
Table 5.13: The Cyclical Nature of the U.S. Airline Industry 186
Table 5.14: Aggregate Industry Snapshot Over time 188
Table 5.15: American Airlines’ Actions Relevant for the DOJ Lawsuit 193
Table 5.16: Percentage of Domestic Air Carrier Slots Held by Selected Groups 199
Table 5.17: Direct Aids to Major Airlines by the Federal Government 200
Table 6.1: Contents of O&D databases 209
Table 6.2: Comparison of Alternative Methodologies 218
Table 7.1: Multiple Linear Regression Results 227
Table 7.2: Main Logit Model Results 228
Table 7.3: Logit Model Results for Alternative Cut-off Points 229
Table 7.4: Main HLM Results 232
Table 7.5: Logit Model Results for Hypotheses 8 and 9 233
Table 7.6: Logit Model Classification Table 234
Table 7.7: MLR Results for Hypothesis 10 235
Table 7.8: Logit Analysis Results for Hypotheses 11 and 12 236
Table 7.9: Bivariate Correlations for Hypotheses 11 and 12 238
Table 7.10: Cox Regression Results for Hypothesis 11 and 12 238
Table 7.11: Summary of Results 239
Table 8.1: Overall Contributions through Key Marketing Concepts,
Methods, and Data Sources 244
Table 8.2: Issues/Research Perspectives on the Use of Pricing Strategy
in Market Defense 255
Table A1: Predatory Pricing Survey 258
x
Table B1: Bivariate Correlations with Full Sample 262
Table B2: Bivariate Correlations with Final Sample 263
Table E1: Correlation Table 270
Table E2: MLR Model results with full sample 270
Table E3: Reduced MLR Model Results 271
Table G1: Results with Dummy Variables 274
Table H2: Event-History Analysis 279
Table I1: Overview of Relevant Empirical Research in the Airline Industry 280
xi
LIST OF FIGURES
Figure 2.1: A Simple Game-Theoretic Construct in the Airline Context 13
Figure 2.2: A Typology of Reputation and Image 18
Figure 2.3: Framework for Assessing the Incumbent’s Reputation for 21
Predation and its Impact on Market Entry
Figure 3.1: Perception of Predatory Pricing under Current Law 38
Figure 3.2: Approximated Predatory Pricing Plaintiff Court Success over Time 44
Figure 3.3: Cost Relationships 50
Figure 3.4: Decision-Making Framework in the U.S. Federal Courts 83
Figure 3.5: Framework for Deregulation and Predatory Pricing Relationship 93
Figure 3.6: Incumbent’s Reaction Framework for Competitive Price Reductions 106
Figure 4.1: Pre-entry and Post Entry Network Price Competition Framework 110
Figure 4.2: Marketing Extension for Consumer Welfare Assessment 130
Figure 5.1: Demonstration of fare level differences 1998Q2 157
Figure 5.2: Demonstration of fare level differences: Top 1000 markets, 2001 Q3 158
Figure 5.3: US Travel Agency Overall Airline Commission Rates
and the Number of Agency Offices, 1976-2000 172
Figure 5.4: New Entry Applications 177
Figure 5.5: Entry/Exit Patterns in the Airline Industry 177
Figure 5.6: A Comparison of Generic and Industry Specific Barriers to Entry 197
Figure 7.1: Graph for Survival Function 237
Figure 7.2: Modified Network Price Competition Framework 241
xii
Figure B1: Comparison Graph 1 263
Figure B2: Comparison Graph 2 264
Figure B3: Comparison Graph 3 264
Figure E1: Histogram for Raw Yield Data 267
Figure E2: Histogram after Ln Transformation on Yield 268
Figure E3: Histogram after Log10 Transformation on Yield 268
Figure E4: Histogram after final transformation 269
Figure E5: Normal Probability Plot for Yield 269
Figure H1: Survival Function 278
Figure H2: One Minus Survival Function 278
Figure J1: Incumbent: Northwest; Route MSP-MCI; Entry: 1995 Q2 289
Figure J2: Market Category comparison (1997) 290
Figure J3: Market Category comparison (1998) 290
Figure J4: Market Category comparison over time (0-200 miles) 291
Figure J5: Market Category comparison over time (200-400 miles) 291
Figure J6: DFW-Wichita Fare Patterns 292
Figure J7: Detroit –Boston 1996 Q1- 1997 Q1 292
Figure J8: Detroit –Philadelphia 1996 Q1 – 1997 Q 293
xiii
SUMMARY
The price variable is among the most powerful instruments in the arsenal of the
executives to achieve entry deterrence objectives. There are two main pricing strategies
that firms may use to defend against a competitive market entry. The first of these
options, limit pricing (or entry deterring price), may be utilized prior to competitive entry.
The second option, aggressive (predatory) pricing, may be executed post-entry. The
effectiveness of both of these options is still controversial. For example, the Chicago
School proponents argue that these strategies are anecdotal in nature. On the other hand,
the rationality of such conduct has been reliably simulated by Post-Chicagoans in game
theoretic settings. The potential contributions of the marketing discipline have been
recognized and called upon to help resolve the conflict.
With this dissertation, I attempt to shed light on the role that price plays in
preemptive and post-entry market defense of firms. As such, the questions tackled
include but are not limited to: how effective is price as an entry-deterrence tool; in
conjunction with firm and market specific barriers to entry; and as a post-entry retaliation
mechanism? What are the facilitating conditions for limit, aggressive (predatory),
competitive and supra-competitive pricing? What are the (long-term) consequences of
these strategies? Following a multi-disciplinary literature review, I present a dynamic
process model and test my hypotheses in a key network industry – the airline industry.
Building upon the advantages of multiple methods a la triangulation, I find that both limit
pricing and predatory pricing can serve as effective strategies for the incumbents’ market
defense. Predatory use of pricing in network industries may diminish consumer welfare.
xiv
Results also suggest that firm specific barriers have a more significant role in market
defense than market specific barriers. Insights and frameworks based on the marketing
philosophy are also presented with the hope of advancing the ongoing debate between the
Chicago and Post-Chicago Schools of thought.
1
CHAPTER 1
INTRODUCTION
The contemporary competitive process resembles military campaigns. Firms need
to advance and protect their strategic positions often at great cost. Conflict with too many
competitors quickly diminishes the scarce resources and depresses the viability of firms.
Resource advantages need to be fortified, exploited, and enhanced before competition can
emulate (cf. Barney 1992). Therefore, the attention to market defense mechanisms has
been increasing (Kuester et al. 1999), and managers consider market entry deterrence to
be a major strategic issue (Smiley 1988). Scholars from economics, industrial
organization, and marketing have acknowledged the important managerial and public
policy implications of market defense and provided guidance for this problem (e.g., Bain
1956; Gatignon et al. 1989; Gruca et al. 1992; Han et al. 2001; Kuester et al. 1999;
Milgrom and Roberts 1982a; Ramaswamy et al. 1994; Sullivan 1977).
A firm may retaliate to competitive market entry with several options such as
increased promotional spending (to consumers or trade members), new product launch
and price reductions, simultaneously or otherwise. There is evidence that as the number
of the marketing mix instruments used (breadth of reaction) increases, market defense
becomes less successful (Gatignon et al. 1997). The most common reaction pattern is
reaction by a single variable (Robinson et al. 1988). The price variable is among the most
powerful weapons in the arsenal of the executives to achieve entry deterrence objectives.
It is inherently flexible, can be modified quickly, and is directly associated with
profitability. In one survey, pricing was deemed to be “extremely important” by seventy-
2
eight per cent of the respondents and ranked third among fifteen key marketing issues (cf.
Srinivasan et al. 2000). Chen and McMillan (1992) reported that the likelihood of
competitive response is higher, the response delay is shorter, and the likelihood of a
matching response is higher for price cuts than they are for other competitive actions.
Incumbent firms often reduce price when they encounter new market entry (Calantone
and di Benedetto 1990).
There are two main pricing strategies that firms may use to defend against a
competitive market entry (LeBlanc 1992). The first of these options, limit pricing (or
entry deterring price (Porter 1980)), may be utilized prior to competitive entry. The
second option, predatory (aggressive) pricing, may be executed post-entry. The
effectiveness of both of these options is still controversial. The Chicago School
proponents argue that they are “more anecdotal than actual” (Gilbert 1989, p.125) or an
outright myth (DiLorenzo 1992; Koller 1971; Lott 1999). This has been the basis that
Federal courts have used for not attributing much credit to unfair pricing claims for the
past three decades. On the other hand, the Post-Chicagoans approach the same facts with
different assumptions (e.g., information asymmetry) and deem the same set of options
rational, alive and well (Jung et al. 1994).
1
Lively debates between Chicago and Post-
Chicago scholars continue (e.g., Edlin 2002; Edwards 2002; Elhauge 2003; ten Kate and
Niels 2002).
1
Price competition is the core element of free markets. Even though it has become much
more complicated during the past two decades, it still remains the variable that is the least under
the control of the firm under (perfectly) competitive market conditions. However, when the
marketplace is not competitive the reverse becomes true and one or few firms may enjoy the
benefits of controlling the going price in a market through the use of their market power.
Ironically, drastic price cuts and price wars that are ingredients for healthy competition are also
associated with unfair competitive conduct (Gundlach and Guiltinan 1998). The abuse of market
power in terms of price manipulation is considered illegal and is subject to antitrust sanctions.
3
The disciplines of economics, law, public policy, and more recently strategic
management have been heavily involved in this stalemate debate between Chicago and
Post-Chicago proponents as to how intense (if any) should the antitrust sanctions be. The
potential contributions of the marketing discipline have been recognized and been called
upon to help resolve the conflict (e.g., Bloom and Gundlach 2001b; Grewal and Compeau
1999; Guiltinan and Gundlach 1996b; Gundlach 1995; Helgeson and Gorger 2003; Ursic
and Helgeson 1994). Foer (2002, p.227) suggested that marketing and strategic
management could serve as “the third leg that gives the antitrust stool stability.”
With this dissertation, I attempt to shed light on the role that price plays in pre-
entry and post-entry market defense of firms. As such, the questions I tackle include but
are not limited to: 1. how effective is sole price as an entry-deterring tool; in conjunction
with other barriers to entry; and as a post-entry retaliation mechanism? 2. What are the
facilitating conditions for limit and predatory (aggressive) pricing? 3. What are the long
term consequences of these strategies? After a thorough multi-disciplinary literature
review, I develop a process model and test my hypotheses using multiple methods. Logit
analysis enables me to examine different pricing strategies, market entry and exit as
dependent variables. Event-history analysis enables me to consider the conditions that
lead to or delay market exits. Finally, multi-level mixed coefficients modeling
(hierarchical linear modeling) allows me to go beyond the framework, take the nested
nature of the data into account, and illuminate the role that different types of barriers play
for entry deterrence. Building upon the advantages of these methods a la triangulation, I
attempt to contribute to the debate between Chicago and Post-Chicago Schools of
thought, and provide empirical ammunition towards this objective.
4
The remainder of this dissertation is as follows: In Chapter 2, I undertake a
literature review of the relevant theoretical base and develop a conceptual understanding.
In Chapter 3, I review and synthesize the literature on predatory pricing from a multi-
disciplinary perspective. The in-depth discussion of predatory pricing here leads to the
development of a network price competition framework i.e., Chapter 4 where I introduce
network industries and develop and present my hypotheses. In Chapter 5, I discuss the
context for my network data --the airline industry, in detail. Chapter 6 presents a review
of the methods and measures I intend to use. In Chapter 7, I analyze the data, present,
and discuss the results. In Chapter 8, I conclude with limitations and future research
avenues. The dissertation also includes ten Appendices in support of the eight chapters.
5
CHAPTER 2
THEORETICAL BASES AND CONCEPTUAL UNDERSTANDING
The relevance of competitive signaling, resource advantage theory, game theory,
and marketing/strategy literature for building a market defense framework is reviewed in
this chapter.
2.1 Competitive Signaling
Signaling theory serves as an important foundation to understand limit and
predatory pricing because both represent (potentially) costly signals to potential and
actual entrants. Limit pricing may signal that the cost structure of the incumbent is low
and/or is willing to protect the market through sacrifice, whereas predatory pricing may
indicate both a low cost structure for the incumbent and/or that the incumbent has deep
pockets and is willing to fend off an entrant at all costs (LeBlanc 1992). Similarly, all
price modifications can convey competitive signals, depending on the amount, timing,
and the context. Distinguishing among temporary, evolving and structural changes in
prices is important because they are likely to produce different results (Srinivasan et al.
2000). Facing a price cut, an incumbent firm may choose to accommodate, to defend
market share or to counter-attack with deeper discounts. The assumption on the degree of
(imperfect) information is key in distinguishing Post-Chicago from the Chicago School
of thought (Lande 1993).
A market signal is an action that conveys indications of intentions, motives, goals
or abilities (Porter 1980). Signals can be directed at customers, channel members,
6
competitors or other stakeholders (Prabhu and Stewart 2000). Firms usually infer
meaning from the signals and decide on their competitive responses. The same signal can
gain different meanings depending on the interpretation of the message. Characteristics of
the signal (i.e., clarity, consistency, and aggressiveness) are considered by the receivers
(Heil and Robertson 1991). For example, a price increase can be interpreted as a response
to market demand or a weakness on the sender’s side (Heil and Walters 1993; Prabhu and
Stewart 2000). Moore (1992) found that managers’ beliefs about the nature of the signal
affected the nature of their responses. If the particular signal was perceived to be
cooperative, the managers were more likely to cooperate and not to retaliate.
Signaling can render predation rational when there is imperfect information
(Hilke and Nelson 1987; Kreps and Wilson 1982; Milgrom and Roberts 1982a; Milgrom
and Roberts 1982b). In many markets, the incumbent is more informed about the
characteristics and conditions (e.g., demand, technology, production) of the market than
the new entrant. The utilization of the asymmetry of information can be predatory if the
incumbent influences the belief and expectations of a rival in a way to alter its decision
regarding entry, exit, price and output levels. This may especially be the case, if the prey
is more efficient than the predator but still decides not to enter a market as a result of
signaling, signal jamming, or reputation (Gundlach 1995). It was also suggested that
when competitive intelligence is not reliable, managers have incentives to overreact when
they face competitive actions (Leeflang and Wittink 1996).
Signal Jamming: Signal jamming refers to the unobservability of a sender’s actions
regarding key variables such as cost. A dominant firm might send signals to current and
7
prospective rivals that its costs are low and the chances for profitable entry are dim
(Milgrom and Roberts 1990). For example, a competitor may alternate its marketing
variables (e.g., sales promotions) to steal the short term demand for a potential entrant’s
product when they are conducting market tests (Gundlach 1995). The potential entrant
may not realize the ploy and decide not to launch the product. False/pre-mature product
announcements (i.e., vaporware) can mislead competitors, potential entrants, and buyers
in their decisions (Bayus et al. 2001; Eliashberg and Robertson 1988; Robertson et al.
1995). Thus, signal jamming may result in an exit, a decision to not enter, or have other
anti-competitive effects (Fudenberg and Tirole 1986; Grout 2000).
Reputation: As will be detailed in a later section, repeated interaction between firms
facilitates the formation of perceptions and beliefs of competitors. A hostile reputation
can be an asset for the firm in the long run (Weigelt and Camerer 1988). A reputation for
predation can deter entrants as well as potential investors of competitors. Burns (1986)
showed that alleged predation depressed the acquisition cost of the victims and others
through reputation.
Case in point: The Department of Justice (DOJ) filed a suit against the Airline Tariff
Publishing Company (ATP) in 1992. DOJ argued that the carriers were able to negotiate
and agree on prices by signaling them through the ATP computer reservation system for a
future date. Another charge was that carriers were able to negotiate the elimination of
discounted tickets via signaling within the system. The case was settled in 1994 after the
airlines agreed to restrictions in using the system (1994b).
8
2.2 Resource Advantage (R-A) Theory
Nonetheless, as an evolutionary, process view of competition, R-A theory
provides a viable starting point from which to debate antitrust issues. It
also provides public policy researchers with many potentially fruitful
avenues for empirical investigation (Hunt and Arnett 2001 p. 23-24)
R-A theory is a dynamic, process theory of competition developed by Shelby D.
Hunt and Robert M. Morgan. It has been developed through more than a dozen articles
over time (e.g., Hunt and Morgan 1995; Hunt and Morgan 1996) and is detailed in two
books (Hunt 2000; Hunt 2002). It draws from evolutionary economics, Austrian
economics, heterogeneous demand theory, differential advantage theory, historical
tradition, industrial-organization economies, resource-based tradition, competence based
tradition, institutional economics, transaction cost economies, and economic sociology
(Hunt 2000). Table 2.1 provides a summary of characteristics that R-A synthesis shares
with other theory.
Table 2.1: The Pedigree of Resource Advantage Theory
Research Tradition Representative Works Affinities with R-A Theory
Evolutionary Economics (Marshall 1890)
(Schumpeter 1934; 1950)
(Alchian 1950)
(Nelson and Winter 1982)
(Langlois 1986) (Dosi et al. 1988)
(Witt 1992) (Foss 1993)
(Hodgson 1993)
Competition is an evolutionary,
disequilibrating process. Firms
have heterogeneous competences.
Path Dependencies can occur.
Austrian Economics (Mises 1920; 1949)
(Hayek 1935; 1948)
(Rothbard 1962)
(Kirzner 1979; 1982)
(Lachmann 1986)
Competition is a knowledge-
discovery process. Markets are in
disequilibrium. Entrepreneurship
is important. Value is subjective.
Intangibles can be resources.
Heterogeneous Demand Theory (Chamberlin 1933)
(Smith 1956)
(Alderson 1957; 1965)
(McCarthy 1960) (Myers 1996)
Intra-industry demand is
substantially heterogeneous.
Heterogeneous supply is natural.
“Product” should be defined
broadly.
Differential Advantage Theory (Clark 1954; 1961)
(Alderson 1957; 1965)
Competition (a) is dynamic, (b) is
both initiatory and defensive, and
9
(c) involves a struggle for
advantages. General equilibrium
is an inappropriate welfare ideal.
Historical Tradition (North 1981; 1990)
(Chandler 1990)
(Landes 1998)
History “counts.” Firms are
entities that are historically
situated in space and time.
Institutions influence economic
performance.
Industrial-organization
Economies
(Mason 1939)
(Bain 1954; 1956)
(Porter 1980; 1985)
Firm’s objective is superior
financial performance. Market
place positions determine relative
performance. Competitors,
suppliers, and customers
influence performance.
Resource-based Tradition (Penrose 1959)
(Lippman and Rumelt 1982)
(Rumelt 1984)
(Wernerfelt 1984a)
(Dierickx and Cool 1989)
(Barney 1986; 1992)
(Conner 1991) (Grant 1991)
(Pfeffer and Salancik 1978)
Resources may be tangible or
intangible. Firms are historically
situated combiners of
heterogeneous, imperfectly
mobile resources. Firms are
constrained by a network of
resource inter-dependencies with
other entities.
Competence-based Tradition (Selznick 1957) (Andrews 1971)
(Hofer and Schendel 1978)
(Hamel and Prahalad 1989;
1994a; 1994b)
(Prahalad and Hamel 1990; 1993)
(Teece and Pisano 1994)
(Day and Nedungadi 1994)
(Aaker 1995) (Sanchez et al.
1996) (Heene and Sanchez 1996)
(Sanchez and Heene 1997)
(Christensen 1997; Christensen
and Bower 1996)
Competition is disequilibrating.
Competences are resources.
Renewal competences prompt
proactive innovation. Firms learn
from competing. Firms are
embedded.
Institutional Economics (Veblen 1899; 1904)
(Commons 1924; 1934)
(Hamilton 1932) (Kapp 1976)
(Neale 1987) (Mayhew 1987)
(DeGregori 1987)
(Ranson 1987) (Hodgson 1994)
Competition is disequilibriating.
“Capital” is more than just
physical resources. Resources
have “capabilities.”
Transaction Cost Economies (Coase 1937)
(Williamson 1975; 1985; 1996)
Opportunism occurs. Many
resources are firm specific. Firm-
specific resources are important.
Economic Sociology (Parsons and Smelser 1956)
(Granovetter 1985; 1994)
(Etzioni 1988) (Coleman 1990)
(Zukin and DiMaggio 1990)
(Powell and Smith-Doerr 1994)
(Smelster and Richard 1994)
(Scott 1995) (Uzzi 1996)
(Fligstein 1996)
Institutions can be independent
variables. Social relations may be
resources. Economic systems are
embedded.
Table 2.1 (continued)
10
Drawing from the above theory, Hunt and Morgan (1997) present the
foundational premises of R-A Theory as follows:
1. Demand is heterogeneous across industries, heterogeneous within industries
and dynamic.
2. Consumer information is imperfect and costly.
3. Human motivation is constrained self-interest seeking.
4. The firm’s objective is superior financial performance.
5. The firm’s information is imperfect and costly.
6. The firm’s resources are financial, physical, legal, human, organizational,
informational, and relational.
7. Resource characteristics are heterogeneous and imperfectly mobile.
8. The role of management is to recognize, understand, create, select, implement,
and modify strategies.
9. Competitive dynamics are disequilibrium provoking; innovation is endogenous.
Of particular interest for my purposes, Hunt and Arnett (2001) argue that R-A is
the remedy for the current antitrust stagnation. For example, they argue that a market
based advantage that stems from a contractual agreement that exclusively ties a
distributor is anticompetitive if the distributor is coerced into the agreement through the
bundling of a complementary product. In this case, the market advantage does not
necessarily come from a relational resource but through market power. They argue that
the focus on neo-classical theory and equilibrium economics has limited the exposure to
outside ideas in antitrust. The focus of Chicago School of antitrust has been solely
economic efficiency (as opposed to social welfare), which cannot be sustained through
the analysis of the static supply and demand curves which actually do not exist, and thus
cannot be calculated (Hunt and Arnett 2001).
R-A is a dynamic theory that also focuses on market segments, comparative and
competitive advantage/disadvantages and enables useful insight for framework
11
development (Hunt and Arnett 2001). It encompasses a wide body of theories (e.g.,
resource-based view (Barney 1986; Barney 1992)) and extant research (e.g., the role of
disruptive technologies (Christensen 1997; Christensen 2001)). Finally, it stresses and
avoids many of the unrealistic assumptions of the neo-classical price theory, which is also
a goal of the current research.
2.3 Game Theory
Isaac and Smith (1985) were among the first to study predation possibilities
through game-theoretic experiments involving decision making under competitive
contexts. They run different versions of a single market design experiment in search of
predatory pricing (including one with sunk costs), yet they did not detect any. However,
Jung et al. (1994) commonly observed predatory pricing in their experiment with
repetitions of a simple signaling game. This game was set up so that the potential entrant
would choose whether to enter or not, and the monopolist would choose whether to fight
or accommodate. Prospective entrants were permitted to observe monopolist’s decisions.
The entrants were better-off if they remained out of the market, unless the monopolist
chose to accommodate. Moreover, Harrison (1988) modified and implemented Isaac and
Smith’s (1985) work to a multiple market setting and found evidence for predatory
pricing. Harrison’s settings were further replicated with modifications and consistent
patterns of predatory pricing were detected in most of the markets. It was shown that
predatory pricing can be reliably simulated “both in stylized signaling games and in rich
market settings”(Gomez et al. 1999) and that it is “alive and well” (Jung et al. 1994,
p.73).
12
Kreps and Wilson (1982) analyzed the incentives of an incumbent to fight an
entrant in order to influence the beliefs of future entrants in a case of multi-market
predation. This work led to the classic work of (Milgrom and Roberts 1982b). According
to the Milgrom and Roberts model, an incumbent firm manipulates its price in order to
signal to a potential entrant, hence exerts influence on the entry decision (Milgrom and
Roberts 1982a; 1982b). In a dynamic signaling game with two-sided uncertainty, it was
shown that when the incumbent expected the entrant to be weak, predatory pricing was
chosen, and when the incumbent expected the rival to be strong, limit pricing was chosen
(LeBlanc 1992). “Predatory pricing is a rational strategy on game-theoretic models of
oligopoly, based on informational asymmetries, that take into account signaling, signal
jamming, and reputation effects” (Burns 1989, p.327). Thus, the entry deterring effects of
a reputation for predation have also been established with game theoretical experiments.
Guiltinan and Gundlach (1996a) suggested that the courts would benefit from the
competitive interaction and strategic decision-making insights. The advancement of game
theory has enabled the study of these complex issues. In that context, acting strategically
meant to consider the expected move of the opponent to come up with the best possible
move. This can be staged as a multistage game in which players intend to make
preemptive moves. Purpose and intent are also considered, and the information needed
for decision making of existing and potential players are described in strategic game
theoretic models. A simple game-theoretic framework is presented in Figure 2.1.
13
Figure 2.1: A Simple Game-Theoretic Construct in the Airline Context
Note: The expected values represent the average fare and the load factors of the incumbent and
the entrant respectively.
2.4 Marketing Strategy
Marketing strategy research (e.g., research focusing on marketing-strategy
content, formulation process, and implementation related issues) and strategy research in
marketing (research focusing on the role of marketing in the formulation of corporate and
business level strategy and knowledge management) are two broad research streams that
are getting intertwined (Bharadwaj and Varadarajan 2004). The determinants of business
performance that the academics have considered to be predominant have evolved over
time. The current paradigm appears to be the resource-based view of the firm also
adopted the by general R-A Theory. Bharadwaj and Varadarajan (2004, pp.222-23)
provide an overview of the evolution of thinking on this topic over time and complement
R-A theory.
Incumbent
Potential
Entrant
Potential
Entrant
Enter
Enter
Do not enter
Do not enter
EV: ($100, 90%;
$100, 85%)
EV: ($150, 80%; 0)
EV: ($180, 65%,
$120; 90%)
Predate
Do not
predate
EV: ($200, 65%; 0)
After reviewing these major theories, I focus on two relevant conceptualizations
from marketing strategy. The first of these works is by Gruca and Sudharshan (1995)
which focused mainly on pre-entry conditions. Their generic entry deterrence strategy
framework consisted of feedback loops that started and evolved around the competitive
environment (cost conditions, demand conditions, history, and legal climate). The loop
consisted of alternative entry deterrence strategies (at functional, business unit, and
corporate levels) leading to the anticipated entry decision, leading to anticipated
consequences for incumbent, leading to choice of entry deterrence strategy, leading to
entry decision, leading to realized consequences for the incumbent. Much of the attention
was devoted to the competitive environment which formed the core of their model. The
current research is focused on a single element (i.e., price) of the functional level entry
deterrence strategies that Gruca and Sudharshan (1995) describe. Therefore, the overall
marketing mix strategy (integrated product differentiation), building switching costs,
brand proliferation, and pre-announcements of new products are beyond the scope of this
research.
The second effort focused on post-entry competitive response options as opposed
to pre-entry deterrence. Kuester et al. (1999) examined and summarized the empirical
contributions to competitive market entry reactions to date and reported five dimensions
for it.
Instrumental: refers to the elements of the marketing mix used for reaction. If retaliation
occurs using the same instrument (e.g., counter product introduction or counter price cut)
reciprocal retaliation is said to occur.
Intensity: the weight of reaction (i.e., the funds allocated for counter promotional
activities.)
Breadth: the number (variety) of marketing instruments used,
15
Time: the speed (or time lag) of reaction.
Domain: refers to the choice of market for counterattack.
It should be noted that Kuester et al. (1999) focused on price and product
retaliation and ignored domain retaliation in the empirical examination.
The understanding of competitive responses was furthered by the “Defender”
model (Hauser and Shugan 1983). Using a zero-sum approach, empirical results have
supported that the optimal response to entry is to reduce price, advertising, and
distribution spending for non-dominant brands; to reduce price but increase marketing
budget for dominant brands (Gruca et al. 1992). This suggests that price reduction is the
de facto optimal response against competitive entries. Relative power theory suggested
that strong incumbents (i.e., those with market power) are expected to retaliate more
intensively and more often than weak incumbents (Kumar et al. 1998). The focus in
recent competitive interaction literature has been on competitive reactions at the retail
level and especially on sales (price) promotions, advertising expenditures, and store
brand sales due to the availability of scanner data. Most advertising and price promotions
do not attract retaliation by incumbents (Nijs et al. 2001). However, when there is
retaliation, it tends to focus on a single variable and often uses the same instrument of
aggression (Steenkamp et al. 2005). Despite the observation that relative levels of price
among competitors explain significant variance in retail strategy (Shankar and Bolton
2004), the clash of the relative prices in inter-type competition (rather than short-term
price reductions) remains a gap in literature.
16
2.4.1 Reputation effects from a marketing perspective
A CEO is ultimately responsible for the growth of a company as evidenced
by its financial performance, its capacity for self-renewal, and its
character. The only way you can measure character is by reputation.
Roberto Goizueta (Goizueta 1995).
If the firm can convince its rivals that it is committed to a strategic move it
is making or plans to make, it increases the chances that rivals will resign
themselves to the new position and not to expend the resources to retaliate
or try to cause the firm to back down. Thus, commitment can deter
retaliation (Porter 1980, p.101).
A vastly unexplored area of research is how to measure the reputation of a firm
and its effects on different constituents in the marketplace. A favorable reputation has
been linked to survival in crisis (Yoon et al. 1993), positive customer attitudes toward the
company’s products and salespeople (Brown 1995), enhanced buying intentions (Yoon et
al. 1993), and choice (cf. Traynor 1983; cf. Weiss et al. 1999).
Corporate Reputation has been defined as the “overall estimation in which a
company is held by its constituents. A corporate reputation represents the “net” affective
or emotional reaction –good or bad, weak or strong– of customers, investors, employees,
and the general public to the company’s name” (Fombrun 1996).
In the marketing domain, Reputation Management has been increasingly attached
to the public relations function. A good reputation may be considered the most important
asset of a company in the long run and can help it survive and even thrive during the
tough times. Reputation is relied on for many aspects of organizational decision-making
at different levels (e.g., choosing a supplier/distributor, promoting an employee).
Similarly, many organizational decisions impact the reputation of the firm, thus attention
17
to this largely ignored topic is necessary. The corporate need for a Chief Reputation
Officer (CRO) position has been proposed to manage the Reputational Capital of firms
(Young 1996). “Reputation is becoming central in the language of strategy and
competition, rather than in the old language of public relations” (cf. Garone 1998). Very
few empirical studies that study reputation effects exist (Landon and Smith 1997; Landon
and Smith 1998).
Often used interchangeably in the marketing domain, both corporate reputation
and corporate image reflect perceptions of an entity. However, they are conceptually
distinct in two main ways. Image summarizes a brand or firm’s identity (Park et al. 1986)
whereas “reputation reflects an overall judgment regarding the extent to which a firm is
held in high esteem or regard. Thus, whereas image reflects what a firm stands for,
reputation reflects how well it has done in the eyes of the marketplace. Image and
reputation are distinct concepts as each can vary independent of the other. A firm can
change its image through positioning, though its reputation remains intact” (Weiss et al.
1999). Moreover, the attractiveness of images is segment specific (e.g., Rolex for luxury),
but a favorable reputation is desirable by all customer segments (Weiss et al. 1999). For
example, Virgin Group’s businesses range from book publishing, radio and television
broadcasting, hotel management to entertainment retail, trading investment, and airlines.
“Virgin is considered the consummate specialist in all things for youth fashion and
fashionability” (Garone 1998) despite the bad publicity about the poor quality and
services of their railroad services (Bower 2000).
Sheth (in personal communication 2004) suggested two typologies to illustrate the
dynamics of reputation. First, he argued that a reputation can be represented in two
18
dimensions: strong/weak and good/bad. Therefore, similar to conjoint analysis (Malhotra
1999), the strength of the attribute (weight of attribute utility) can be multiplied with
reputation score to observe overall reputation (i.e.,
1 1
( )
i
k m
ij ij
i j
U x x ?
= =
=
??
(Jain et al. 1979)).
The same typology can be applied to the concept of image as well. Sheth (2004) also
suggested a 2X2 matrix in which firm image and form reputation form the two
dimensions. This typology is represented on the sample spatial map below (Figure 2.2):
Image +
Reputation +
Apple
Rolex
Enron
Jaguar
Wal-Mart
Nike
Harley-Davidson
Microsoft
Virgin
RJR Nabisco
Campbell
Figure 2.2: A Typology of Reputation and Image
It should be noted that reputation has a somewhat different meaning when studied
from an economics perspective. In this context, reputation has to do with the consistency
to keep promises and sticking to a particular strategy (e.g., a government with a
reputation for committing to a path for money supply) (Rogoff 1989). Expected actions
19
are anticipated through reputation (Evans and Thomas 1997). The rationality of
reputation building and its effects on decision making, bargaining and technology
adoption has been shown in game theoretical agency settings (DeJong et al. 1985;
Dobson 1993; Evans and Thomas 1997; Hendricks 1992; Park 1999). The significantly
positive reputation effects for not expropriating minority shareholders on stock prices and
IPOs have been shown in Finance (Gomes 2000).
Finally, reputation is important to examine from an antitrust perspective where it
gains a different interpretation. Limited work on reputation so far has focused on the
customer’s perspective. The classic definitions of reputation have not considered
competitors as direct constituents of reputation. However, it may be possible for a
monopolist to charge supra-competitive prices due to high barriers to entry, reputation for
predation and other signaling effects. In this context, a reputation for predation is
established by constant signaling of future intention to predate in the face of new entry.
This is best demonstrated by actions in a market. Potentially more efficient competitors in
(other) markets observe the reactions of the incumbent to entry, and often the quick
demise of the previous entrant. As a result, they may decide not to commit the high level
of resources needed to compete against the incumbent. Ceteris paribus, they would rather
enter a market where the incumbent does not have a reputation for predation. Thus, a
valid purpose of predation may be to develop a reputation as a tough competitor
(Comanor and Frech 1993; Kreps and Wilson 1982). Reputation for predation is neither
the only, nor the most effective factor affecting potential entry. However, a reputation for
predation implies strong and constant signaling to all potential competitors for all the
markets a company operates in, thus may be quite influential overall. Similarly, Weiss et
20
al. (1999) concluded that corporate reputation has a strategic influence beyond traditional
approaches. The DOJ stated that the existence of a reputation for predation could be
examined by means of industry surveys and that such a study had not been undertaken
(1999e, p.134).
Repeated interaction between firms facilitates the formation of perceptions and
beliefs of competitors. A hostile reputation can be an asset for the firm in the long run
(Weigelt and Camerer 1988). A reputation for predation can deter entrants as well as
potential investors of competitors. Burns (1986) showed that alleged predation depressed
the acquisition cost of the victims and others through reputation. Areeda and Turner
(1975) admitted that “a demonstrated willingness to indulge in predatory pricing might
itself deter some smaller potential entrants…” The reputation can develop from previous
experience with the incumbent in the market in question for potential entry or in other
markets. Similarly, the reputation may be derived from observing other firms’
competitive interactions with the incumbent in the market in question for potential entry
or in other markets. The anticipated retaliation for specific markets may differ because
the incumbent may have revealed its intent to protect a specific market(s) (e.g., fortress
hubs (Allvine 1996b)) at all costs and not react so sharply to other market entries. I
develop this notion into a framework in Figure 2.3 below:
21
Strategic
Entry
Deterrence
Aggressive Competition Experience
with Incumbent
in Specific Market
Aggressive Competition Experience
with Incumbent
in Other Markets
Observed Aggressive Competition
of Incumbent with Other Firms
in Other Markets
Observed Aggressive Competition
of Incumbent with Other Firms
in Specific Market
Framework for Assessing the Incumbent’s Reputation for Predation
and its Impact on Market Entry
Incumbent’s
Reputation
for Predation
+
+
+
+
+
Profitability
Supra-competitive
Prices
+
+
Rivalry
Familiarity
Size
Figure 2.3: Framework for Assessing the Incumbent’s Reputation for Predation and its
Impact on Market Entry
For example, in the airline context, a carrier “defending its turf” against
encroachment by a start-up carrier in a few markets can create a “reputation for
predation” that deters start-up carriers from entering its many other hub markets; this can
significantly alter the “cost-benefit” predation calculation for a hub carrier in a way
uncharacteristic of most other industries” (Nannes 1999). Frederick Reid, a former Pan
Am and American employee, who recruited to be the President and Chief Operating
Officer for Lufthansa admitted to this phenomenon: “Pan Am was effectively destroyed
by it and American was a winner. American has this reputation as a tough customer…”
(McCormick and Field 1997). “If the prey believes that the threat or promise will be
22
carried out, there is no need for actual predation. Thus, like collusion, the most successful
use of predatory threats or promises is difficult for outsiders to observe”(Comanor and
Frech 1993). Reputations are impressions of actual behavior by the constituents. Yet, the
courts continue to treat reputation effects as “industry folklore” because of a lack of
empirical evidence. The bottom-line implication of the reputation effects literature is that
it is important to consider the reputation effects as a barrier to entry when studying pre-
and post-entry market defense.
2.4.2 Inter- vs. intra- type competition:
Intra-type competition takes place between businesses of similar
cost/organization/service structure (e.g., Goldman Sachs/Merrill Lynch). By contrast,
inter-type competition is defined as competition between businesses with different
structure (e.g., Merrill Lynch/Charles Schwab) (Allvine 1996b). Examples in business-to-
business settings are commonplace as well: In IT outsourcing, IBM, EDS, and Accenture
are intra-type competitors whereas Infosys, TCS, and Wipro (i.e., Indian offshore
providers) would be considered their inter-type competitors. The competition between
large distributors (e.g., IKON for copier and printers) and the local dealers can be
characterized as inter-type. It is typical for intra-type competition to focus on non-price
factors (e.g., facilities, sales assistance, and extended warranties) and for inter-type
competition to focus on discounts. Such structural differences between inter- and intra-
type competition have long been observed (Allvine 1996; Miller, Reardon, and McCorkle
1999). Similarly, competition between Delta and American Airlines is considered intra-
type, while competition between Vanguard and American Airlines would be considered
inter-type competition. It is typical for intra-type airlines to focus on non-price factors
23
(e.g., promotion, frequent flier miles) and for inter-type airlines to compete on price.
Table 2.2 contrasts the two types of competition.
Table 2.2: Characteristics of Intra-type versus Inter type Competition
Intra-type Inter-type
Occurrence
Change
Risks
Established Methods
Innovations
Continuous
Small
Low
High
Minor
Discontinuous/random
Large
High
Low
Major
Changes in efficiency
Demand Effect
Low
Secondary
High
Primary
Leadership Style Managerial Entrepreneurial
Outcome Reduction of Uncertainty Novelty
Source: Dixit (2000)
A major take-away from the retailing literature is that inter-type businesses (e.g.,
general merchandiser (Sears), broad-line specialist (Home Depot), limited-line specialist
(Ace Hardware)) can co-exist and prosper (i.e., number of larger stores are positively
related to the size and number of smaller stores (Miller et al. 1999)). Similar formations
in many business-to-business contexts exist: in machine tools there are industry giants
such as Illinois Tool Works that supply a wide variety of equipment, but there are also
operations that focus on segments of the market (e.g., CNC tool rooms), and finally
smaller shops that only custom-build. In heavy construction equipment, Caterpillar and
Komatsu are the market leaders, but Linkbelt’s specialization in cranes pays-off.
At the absence of inter-type competition price competition typically suffers. For
example, Table 2.3 illustrates the lack of price competition between intra-type
competitors in the airline industry.
24
Table 2.3: Non-competitive Pricing among Intra-type Competitors in the Airline Industry
(Oster and Strong 2001, p.31)
The notion of inter-type versus intra-type competition is also supported by the
notion of mutual forbearance.
2.4.2.1 Mutual forbearance
Mutual forbearance theory implies that the higher the multi-market contact
between the same firms, the lesser the intensity of competition due to increased
familiarity between firms and their ability for deterrence (Jayachandran et al. 1999).
Therefore, gaining competitive intelligence becomes advantageous. Korn and Baum
(1999) observed that empirical evidence has robustly demonstrated that multimarket
contact leads to mutual forbearance.
Mutual forbearance moderates the effect of rivalry through tacit collusion and
leads to improved performance for firms. Increased deterrence (Porter 1980) and
increased familiarity among firms (Baum and Korn 1999) were observed as potential
reasons for this effect (Jayachandran et al. 1999). While direct collusion is illegal and has
25
antitrust consequences, tacit collusion is observed when firms understand the motives and
strategies of each party and implicitly coordinate so that intense competition is avoided
(Jayachandran et al. 1999). Since most discounters that provide inter-type competition are
smaller than their major counterparts they would have less market contact than intra-type
competitors. Therefore, mutual forbearance would not only explain the general intensity
aspect of inter-type competition but also help explain the exception for large inter-type
competitors (e.g., intra-type major competitors do not retaliate against large inter-type
discounters such as Wal-Mart and Southwest Airlines because of the high multi-market
contact with them).
Korn and Baum (1999) found out that firms (in the airline industry) did not
actively increase multi-market contact to achieve mutual forbearance. Perhaps the
potential for increased market contact is limited by resource constraints or regulation.
Moreover, previous studies have generally not differentiated market contacts between
inter- and intra-type competing firms. Jayachandran et al. (1999) argued that the positive
relationship between multimarket contact and the intensity of competition was moderated
by the organizational structure of competing firms, seller concentration, spheres of
influence (i.e., focal market distribution), and resource similarity (i.e., parity).
Mutual forbearance may have antitrust implications through collusion, and multimarket
reactions may have antitrust implications through predation (Jayachandran et al. 1999).
As previously discussed, an incumbent's competitive reputation can also deter market
entry in the context of multimarket competition. Clark and Montgomery (1998) have
experimentally shown that an incumbent's reputation for aggressiveness, but not
intelligence, makes a market less attractive and more risky to a potential entrant. They
26
found that reputation has a stronger effect when the degree of multimarket contact is
high.
The theoretical basis for the use of pricing strategies as a defense mechanism in
competitive interaction was inquired in this Chapter. The theories are linked to support of
specific hypotheses in Chapter 4 (also see Figure 4.4). In the next Chapter, an inquiry of
the actual use and evolution of aggressive pricing is undertaken from a multi-disciplinary
perspective and the contemporary thinking is synthesized with insights from marketing.
27
CHAPTER 3
AN INQUIRY OF THE USE OF PREDATORY PRICING IN MARKET
DEFENSE: EVOLUTION, REVIEW, AND MARKETING SYNTHESIS
3.1 Introduction to the Review
The monopolists, by keeping the market constantly understocked, by never
fully supplying the effectual demand, sell their commodities much above
the natural price, and raise their emoluments, whether they consist in
wages or profit, greatly above their natural rate.
Adam Smith, The Wealth of Nations (1776)
Aggressive competitive conduct by a monopolist, which is beneficial to
consumers, and aggressive exclusionary conduct by a monopolist, which
is deleterious to consumers, look alike.
(Blair and Esquibel 1995)
Pricing has become an increasingly complex and sophisticated marketing activity
over the last few decades. Interestingly, among all marketing mix elements, it is the one
that is the least under the control of the business under perfectly competitive market
conditions. However, when a market is not competitive, the reverse becomes true and one
or few businesses may enjoy the benefits of controlling the going price through the use of
their market power. The abuse of market power in terms of price manipulation is
considered illegal. This chapter focuses on a widely recognized and debated form of price
manipulation –predatory pricing, and describes how the marketing discipline can
contribute to its assessment.
The current stage of the evolution predatory pricing is quite important. The very
mechanism, drastic price cuts, that is associated with predatory pricing also happens to be
at the heart of healthy competition (1993b). The primary objective of public policy
28
making regarding competitive interaction is to distinguish anti-competitive conduct from
those that are pro-competitive (Scherer 1976). Despite the growing economic literature
on predatory pricing, the case law in Europe and the U.S. remains limited (Grout 2000).
This chapter reviews the extensive literature on the history of predatory pricing
and presents insights to a serious debate on whether or not a new set of ground rules for
evaluating predatory pricing antitrust cases should be adopted. The notion that the
Supreme Court’s Brooke decision has established an imperfect standard against the
plaintiffs/new entrants is also examined. The main goals of this chapter are to provide a
historical perspective on the topic, to summarize the new evaluation alternatives available
to the courts, and to present a synthesis that also integrates marketing insights.
Courts generally do not find predatory pricing to be rational and assume that
predatory pricing practices are rare. This emphasis comes from the concern for antitrust
litigation to not disturb the beneficial, competitive price-cutting behavior. At the absence
of absolute market power, the courts simply view the cases as competitive (Sheffet and
Petty 1994). The concept of predatory pricing represents a double-edged sword for the
policy makers in that if not prevented, the price competition that enhances consumer
welfare in the short-run can turn out to be disadvantageous for the consumers with return
of the supra-competitive prices in the long run (Gundlach 1995).
The positive impact of lower prices on consumer welfare is generally accepted.
However, predatory pricing is detrimental to consumer welfare in the long run because
once the competitors exit the market, the predator raises prices with the intention of
collecting supra-normal profits. Moreover, the problem with predatory pricing is not
limited to harm to consumers through the increase of prices back to monopoly levels.
29
Predatory pricing, successful or not, can potentially reduce incentives for investment and
innovation, and prevent new entry or expansion by more efficient firms. There are special
implications for network industries such as telecommunications and software where the
value of the product/service increases along with the number of users. Innovation can be
stifled when predatory prices induce consumers to continue to use an old technology as
opposed to a superior alternative offered by a new entrant (Guiltinan and Gundlach
1996).
Price competition is the core element of free markets. Lower prices and
competition increase the welfare of the consumers and the society in general (Grewal and
Compeau 1999). Yet, if consumers feel that the price they have to pay is unfair, then
social harm may occur (Guiltinan and Gundlach 1996). The line between competition and
anti-competitive conduct needs to be carefully drawn.
It appears that there is an increasing gap between the insights from the modern
economic theory and the enforcement of current judicial policy. Government
enforcement concern is high as evidenced by the DOJ lawsuits. The new economy
requires new rules for the assessment of predation because of the growing importance of
intellectual property (e.g., Microsoft litigation). Increasing market concentration in many
industries and number of mergers, are also of concern (Bolton et al. 2000). There is an
ongoing debate between two camps on how enforcement on predatory pricing should be
exercised.
The Chicago School of thought, with its more established neo-classical theory
economists and free enterprise institutes is supported by corporations and practiced by the
Supreme Court. Since 1993, the Supreme Court requires proof of below-cost pricing and
30
of recoupment of losses suffered during predation to concur a predation case, and no
predatory pricing plaintiff has been able to prevail in the courts (Bolton et al. 2000). The
Chicago School concurs that Predatory Pricing cannot be a logical business practice for a
company. They claim that it is not rational if they price below (short-run) AVC, and
perfectly legal if they are above AVC, and thus predatory pricing claims should be
ignored (DiLorenzo 1992).
Many companies are suffering and trying to get attention to what Post-Chicago
School of academics call blatant use of predatory pricing and harm to the competitive
process. The following scenario is observed in many monopolistic markets: The
incumbent signals the intention to predate once a new player announces entry. If the
entrant is bold enough to actually enter the market, the predator matches the entrant’s
(lower) price in the market and usually increases output. The predator declares an all-
front war against the much smaller (but usually more innovative and efficient) entrant
and does not budge until it is driven out of the market or out of business. As soon as the
entrant has been forced out, the monopolist ignores the newly stimulated demand,
reduces the capacity and increases its prices to levels (sometimes higher than) before the
entry. This pattern has been observed for many cases in the airline industry (1996a;
Allvine 1996b). There are many independent experts and academics, who concur that
predatory pricing can indeed be a viable business strategy for the major player in a
monopolistic market. This view validates the need for attention to this important antitrust
topic. The Post-Chicago view of predation has been on the rise in the nineties during the
Clinton Administration as demonstrated by the U.S. versus Microsoft, and U.S. versus
American Airlines cases. However, it would not be surprising if most pending cases were
31
settled during the administration of President George W. Bush, who stated during his
election campaign that he would only pursue price fixing antitrust cases (Financial
Times, 2000c).
Grewal and Compeau (1999) suggested that marketing researchers have not
engaged in public policy implications of pricing until recently and that a focus on this
issue is long overdue. They argued that developments such as the Internet, global
markets, mega-corporations, and cooperative marketing arrangements created the
necessity of taking a closer look at the pricing and public policy interaction with
consumer welfare in mind. After all, courts consider economic harm to consumers as the
best way of assessing harm to society (Baer 1996).
Guiltinan and Gundlach (1996) argued that marketing was in a unique position to
help form public policy guidelines with comprehensive measurement and modeling
procedures, and that predation and predatory pricing have not been addressed by
marketers until recently. Gundlach (1995) suggested that the marketing discipline had the
potential to further the understanding needed for the development of a more suitable
antitrust policy.
The literature on predatory pricing is overwhelming and dominated by the
disciplines of law and economics. This chapter represents a literature review of the
existing knowledge on this topic. It is an effort to thoroughly understand, summarize, and
then synthesize and present a big picture of the subject from the perspective of cases and
law, and the assumptions and development of the thinking in economics.
32
3.2 Pricing
Price can assume many meanings depending on the specific context. It can mean
rent, tuition, fee, fare, rate, interest, toll, premium, honorarium, dues, assessment,
retainer, salary, commission, wage, even bribe and income taxes (Schwartz 1981).
Merriam-Webster Dictionary defines price as “the amount of money given or set as
consideration for the sale of a specified thing.” It is also defined as the quantity of one
thing that is exchanged or demanded in barter or sale for another or the cost at which
something is obtained. More elaborate definitions have involved the concepts of value
and worth. Price can mean a fair return or equivalent in goods, services, or money for
something exchanged, or the monetary worth or value of something (Mish 1995). Simply,
it is the amount of money the customers have to pay for a product or service (Grewal et
al. 1998).
The marketing mix is defined as the set of controllable marketing variables that
marketers employ to obtain the desired responses from their target markets (Kotler and
Armstrong 1991). Price is one of the key components of the classic “four Ps: product,
price, place, and promotion” grouping of the marketing mix (McCarthy 1960). It has been
cited as the most important component of the mix by marketing executives (1983). Price
has special importance for the marketers due to its inherent flexibility and close
association to profitability. General pricing approaches include cost-based pricing (cost-
plus (mark-up), break-even, and target profit), buyer based pricing (perceived value), and
competition based pricing (going rate, sealed bid, competitive response) (Allvine 1999;
Kotler and Armstrong 1991).
33
The pricing strategy is dependent on the nature of the product in question (e.g.,
innovative versus imitative), and the product mix of the firm. For example, market
skimming and market-penetration are two strategies that can be employed for an
innovative product or service. Products have to be positioned in the marketplace to be
profitable and the following are some of the options that can be employed: product-line
pricing, optional-product pricing, captive-product pricing, by-product pricing, and
product-bundle pricing (Kotler and Armstrong 1991, p.351).
Prices can also be adjusted through the use of discount pricing (quantity,
functional, seasonal discounts, payment terms) and allowances (trade-in, promotional);
discriminatory pricing (customer-segment, product-form, location, time); psychological
pricing; promotional pricing (loss leaders, special-event pricing, cash rebates, low-
interest financing, longer warranties, free maintenance, discounts); and geographical
pricing (FOB-Origin, uniform delivered, zone, basing point, freight absorption).
However, both buyers’ and competitors’ reactions need to be considered before changing
prices (Allvine 1999; Assael 1990).
Price theory in economics defines how the firms should set prices under certain
assumptions to maximize their profits (Pass and Lowes 1994). However, the static nature
of the price theory and its rigid assumptions for cost, price and quantity (product) make it
inapplicable to marketing practitioners. When these assumptions are violated, it becomes
hard to measure the nature of the demand (demand curve) and a profit-maximizing price
cannot be determined. Assumptions that marketers cannot afford to have include, an
unchanging environment, single product firms, and all customers paying the same price
(Allvine 1999).
34
Due to the absence of a single profit maximizing guideline in practice, the
marketers determine their strategy for the product or service before they set the price for
it. There are internal (e.g., marketing objectives, marketing mix strategy, costs,
organization for pricing) and external factors (e.g., nature of the market and demand,
competition, economy, resellers, government) that impact pricing decisions. Marketing
objectives of the firm can include survival, current profit maximization, market-share
leadership, product-quality leadership among others. The marketing mix for a particular
product is closely tied to the marketing mix strategy of the related items that the firm
offers. As most of the dot-com start-ups have painfully discovered, prices must start to
exceed their costs (as effected by their economies of scale and learning curve) at some
point if a company is to survive in the long run. The procedures for pricing and the
flexibility for changing the price also have impact on the pricing decision as an internal
factor (Kotler and Armstrong 1991). It should be noted that the nature of the market and
demand characteristics can be industry specific (e.g., seasonal). Economists have
identified four general types of markets (pure competition, monopolistic competition,
oligopolistic competition, and pure monopoly). The price elasticity of demand is also of
concern. The state of the economy impacts the purchasing power of consumers and has to
be considered. The bargaining power of middlemen and retailers can become important
in making pricing decisions. Finally, laws regarding pricing are also important and
marketers need to make sure that their pricing policies do not violate them. These issues
include price fixing, resale price maintenance, price discrimination, minimum (predatory)
pricing, price increases (ceilings), and deceptive pricing (Kotler and Armstrong 1991).
35
Price fixing refers to price collusion among competitors. It is considered illegal except
when supervised by a government agency (e.g., local milk industry agreements, fruit and
vegetable cooperatives). Resale Price Maintenance problem implies that manufacturers
cannot require that their dealers sell at pre-specified prices. They can only propose
suggested retail prices. They cannot refuse to sell to a dealer or punish the dealer
otherwise because of pricing issues. Regulated price increases refer to certain industries
(i.e., utilities) in a free market economy. Government may also use its influence to
discourage major industry price spikes during shortages or in times of inflation.
Deceptive pricing problem means that the price reduction should not be advertised unless
it is a saving from the usual retail price, not advertise inaccurate factory or wholesale
prices, and not advertise comparable prices for different goods. FTC Guidelines against
deceptive pricing were issued in 1958 (United States 1958). Price discrimination problem
implies that sellers must offer the same price terms for a given type of transaction.
Robinson-Patman Act forbids price discrimination unless the seller can prove that its
costs are different in selling to a particular customer than others (1936). This can usually
be the case when order quantities vary in size. However, the seller needs to prove that
these differences are proportional to the differences of quantities. It can also be justified
if the seller can prove that it is trying to meet competitor’s prices in good faith. Even
then, price discrimination should be temporary, localized, and defensive rather than
offensive (Dalrymple and Parsons 1990). For effective discriminatory pricing, segments
in the market should indicate varying levels of demand. Discrimination does not work if
the low price segment customers can resell the product to the higher price segments.
Similarly, competitors who sell to the higher price segments at lower prices can be very
36
disturbing. This may lead some firms to illegally employ predatory pricing practices to
drive the competition out of the market (Kotler and Armstrong 1991). Thus, last but
certainly not least, minimum (predatory) pricing is an important issue that marketers need
to be aware of.
3.2.1 Predatory Pricing
The best known form of predation is predatory pricing, yet it has currently no
statutory definition (Guiltinan and Gundlach 1996). Areeda and Turner (1975) argued
that predatory pricing occurs when a firm lowers its prices in order to eliminate a current
competitor in the relevant market or to prevent new firms from entering the market:
[P]redation…cannot exist unless there is a temporary sacrifice of net
revenues in the expectation of greater future gains.…Thus, predatory
pricing would make little economic sense to a potential predator unless he
had (1) greater financial staying power than his rivals, and (2) a very
substantial prospect that the losses he incurs in the predatory campaign
will be exceeded by the profits to be earned after his rivals have been
destroyed.
Others such as Professor (then Judge) Posner simply defined predatory pricing as
pricing at a level calculated to exclude from the market an equally or more efficient
competitor (Posner 1976). Viscusi et al. (1995) later commented that, pricing at a level to
exclude a less efficient competitor is naturally what competition is supposed to do. Some
researchers focus on the exclusionary conduct (Grout 2000), and some focus on losses
imposed on others (Sullivan 1977) in their definitions. Professor Baumol’s 1996 article
refined the Areeda-Turner definition:
Indeed, one can, perhaps, define a price to be predatory if and only if it
meets all three of the following conditions. First, the choice of that price
37
must have no legitimate business purpose. Second, that price must
threaten the existence or the entry of rivals that are at least as efficient as
the firm (call it firm F) that has adopted the price at issue (price P). Third,
there must be a reasonable prospect of recoupment of at least whatever
initial costs to firm F were entailed in the company’s adoption of the price
in question, that recoupment taking the form of monopoly profits made
possible by reduction (as a result of price P) in the number of competitors
facing F. (Baumol 1996, p.52)
Hence, Professor Baumol would define a business act as legitimate if the expected
(long-run) net return is positive, and if that return does not depend on the exit of any of
equally or more efficient competitors or the prevention of entry of such firms.
Gregory T. Gundlach, who has published numerous articles on the topic of
predation (e.g., Guiltinan and Gundlach 1996a; Gundlach 1990; Gundlach 1995;
Gundlach and Guiltinan 1998) perceives predatory pricing as a reduction of prices
(usually below cost) with the intention of punishing a competitor or gaining higher profits
in the long run by driving competition out of business (Gundlach 1990). Predatory
pricing can improve a firm’s bottom-line through exclusionary conduct or other anti-
competitive effects in the market. Specific cases of pricing below cost, unlawful price
discrimination and price warring are considered predatory pricing (Grewal and Compeau
1999). The same actions regarding pricing can be interpreted differently, depending on
what the intent of the case is. Pricing below cost with the intent of dumping excess
inventory can be considered legal (Grewal and Compeau 1999).
When the cost concept is built into the descriptions above, a definition which
would generally be accepted by the courts in the U.S. is obtained (i.e., pricing below an
appropriate measure of unit cost (e.g., total or variable) with the intent of driving out
rivals (presumably with shallow pockets), and later raising price above unit cost to recoup
38
losses through supra-competitive pricing. The court definition was shaped through cases
such as Matsushita Electric Industrial Co. v. Zenith Radio Corp.; Cargill Inc. v. Monfort
of Colorado Inc.; and finally in Brooke Group Ltd. v. Brown & Williamson Tobacco
Corp. However, the measure of cost that would distinguish predatory from non-predatory
pricing has not yet been prescribed (McCareins 1996). Figure 3.1 summarizes the current
perception of predatory pricing enforcement in the courts:
Below Cost
Pricing
Recoupment
PREDATORY
PRICING
Intent
Market
Power
Figure 3.1: Perception of Predatory Pricing under Current Law
Barriers
to Entry
Whether a particular conduct is predatory may depend on the jurisdiction in which
a firm does business or whether the suit is brought under state or federal law (McCareins
1996). State and Federal courts use different standards to detect predatory pricing. State
courts are usually more receptive to predation cases and tend to protect the preys (i.e.,
smaller firms) more. On the other hand, the Federal courts have stated that predatory
pricing is “inherently uncertain” (1986b, pp.588-89) and that it is generally implausible
39
(1993b, p.226). Today, intent is at best a secondary concern for the courts in the United
States. Predatory pricing is found only when below-cost pricing exists along with (a
dangerous possibility of) recoupment (1993b). This puts a heavy burden of proof on the
plaintiffs.
Perhaps in one of the most widely quoted statements in predatory pricing history,
the Supreme Court in the Matsushita Case stated that predatory pricing is “rarely tried,
and even more rarely successful” (1986b, p.590). The reason for this sharp conclusion is
discussed later in this chapter under the evolution of predatory pricing section. Similarly,
it will be discussed later that there may be exclusionary and anti-competitive effects of
predatory pricing that are not as obvious as the elimination of a direct competitor. The
following definition however, successfully captures the essence of the diverse nature of
the topic, and hence has been adopted for the purposes of my dissertation:
Predatory Pricing is a price reduction that is profitable only because of
the added market power the predator gains from eliminating, disciplining
or otherwise inhibiting the competitive conduct of a rival or potential rival
(Bolton et al. 2000, p.3).
3.2.2 Predatory Pricing versus Predation
Even though predatory pricing is the most widely recognized form, it is only a
part of the larger conduct of predation. Ordover and Willig (1981) defined predation as
actions that are unprofitable but for their possible contribution to a rival’s exit. There is a
rather wide list of activities that can be considered to be predatory depending on their
impact on societal welfare.
Misleading advertising about a competitor’s product, specifically designing
interfaces for complementary products that are incompatible for a rival, initiating
40
excessive regulatory hearings and/or lawsuits have been noted as predation (Bork 1993).
Cutting supply of an essential output or refusal to provide access to essential resources
for competition (e.g., slots at airports) is also predatory according to the “essential
facility” doctrine (Gundlach and Bloom 1993). Prentice (1996) discussed the predation
liability for fraudulent product announcements that a manufacturer knows will not ever
be launched. In addition, raising rival’s costs through acquisition and sleeping on of
patents, product pre-announcements, useless product modifications, exclusionary market
channel arrangements, market share agreements, bundled discounts, discounts designed
to reduce rivals’ ability to obtain display space (Gundlach 1990), refusals to deal,
mandatory tying arrangements, and other uses of the power in one market to increase the
cost to competitors in other related markets (Meeks 1998) have been reported. Non-price
predation (e.g., reputation, raising rival’s costs) may be used to depress the acquisition
cost of rivals.
Predation is like warfare in that it can only be rational when less expensive
measures fail (Scherer 1980). Indeed, it has been argued that that non-price predation is
more effective than predatory pricing because it is less costly, less risky and more often
successful (Bork 1978). Snyder and Kauper (1991) summarized literature on raising
rivals’ costs, a category of predation. Salop and Scheffman (1987) showed that non-price
predation strategies that increase rival’s costs are more advantageous than predatory
pricing. Not every non-price predation is pre-branded illegal, non-price predation does
not have to force exit to be successful and it does not have market power as a
prerequisite. It may be more cost-efficient for large firms to employ contractual
provisions and other strategies to increase barriers to entry to maintain market power. In
41
the broadest sense, sixteen non-price predation strategies have been reported which
included outright sabotage such as burning down a rival’s plant but also vertical
integration, innovation, and product promotion (1991). The next section discusses the
evolution of the theory and practice of predatory pricing but also reports on non-price
predation where necessary.
3.3. Evolution of the Theory and Practice of Predatory Pricing
3.3.1 A Historical Perspective
Predatory pricing cases have a major role among the anti-competitive practice and
antitrust violation cases (Gundlach 1995). Since the turn of the last century, the courts’
approach to evaluate predatory pricing cases has swung back and forth between being in
favor of the plaintiffs and the defendants. Economic theory and scholarly articles seem to
have played an important role, for better or worse, in shaping the federal court policies.
Especially, the articles by McGee (1958), Koller (1971), and Areeda and Turner (1975)
seem to have influenced the courts to an extent that their impact is still observed today in
their reflection from the 1993 Brooke decision. The early influence of academics ended
the “populist era” of predation enforcement, where plaintiffs won most of the cases. The
adoption of Areeda-Turner rule by the courts as a standard shifted the balance in favor of
the defendants. Predatory Pricing was defined by the courts as irrational business
(1986b), and the neo-classical price theory school continuously presented it as a myth
(DiLorenzo 1992; Koller 1971). Moreover, the 1993 Brooke decision has made it even
harder for the plaintiffs to survive in courts. In this decision, the Supreme Court required
not only evidence of below cost pricing but also of recoupment. In none of predatory
42
pricing cases since the Brooke decision (at least 38 of them), did the plaintiffs prevail
(Carney and Zellner 2000). However, modern view holders such as Guiltinan (1996),
(Bolton et al. 2000), as well as distinguished economists such as Alfred Kahn (1998)
have pointed out that certain assumptions of price theory do not hold in the real world.
Currently, there seems to be a consensus among those who hold the modern Post-Chicago
view of predation that predatory pricing can be a viable and profitable business strategy
especially under monopolistic market conditions (Bolton et al. 2000). This indicates that
a reassessment of the standards that the courts use today may be well justified and
overdue. Milestones of the evolution of predatory pricing and an approximation of
plaintiffs’ success rate in the courts are presented next (Table 3.1; Figure 3.2), followed
by a discussion of the evolution.
Table 3.1: A Summary of the Evolution of Predatory Pricing in the U.S.
Timeline Description
1890
Sherman Act Prohibited contracts or conspiracies that restrain trade
including price fixing, and monopolization.
1910
Clayton Act Defined unlawful uncompetitive behavior and practices
other than monopolization.
1911
Standard Oil Co. v. United
States
Standard Oil found guilty in this classic case of
monopolization. A major component of the case involved
the allegations that Standard Oil had employed predatory
pricing to drive its competitors either out of business or to
force them to sell it to Standard Oil at distressed prices.
Standard Oil was broken up into 33 geographically distinct
companies (Gibb and Knowlton 1965).
1914
Federal Trade Commission
(FTC) Act
Extended Sherman Act in terms of restraints for trade. FTC
authorized to interpret antitrust statutes.
1936
Robinson-Patman Act Price discrimination that lessens competition or promotes
monopoly declared illegal. Protected small business from
price-cutting by large sellers.
1938
Wheeler-Lea Amendment FTC authorized to protect consumers as well as competitors.
1940’s
Strong FTC enforcement Cases are infrequent until 1940’s. More cases observed with
strong FTC enforcement (Bolton et al. 2000).
1936-70’s
The Populist Era Roughly 77% of predatory pricing plaintiffs win their case
(Koller 1971).
1967
In this classical example of the populist era, three producers
in California were accused of charging less for their pies in
43
Utah Pie Co. v.
Continental Baking Co.
Utah than in markets closer to their plants after Utah Pie’s
entry to the market. Court decided in favor of Utah Pie. The
Supreme Court reinstated the jury verdict though this
decision was widely criticized. Justice Stewart argued that
the consumers benefited from lower prices and increased
competition in Utah.
1975
Areeda-Turner AVC Rule From the seminal article proposing a per se AVC standard
for detecting predatory pricing.
1975-80
A Defendant’s Paradise Dramatic change in enforcement –no plaintiff prevailed
during the five years following Areeda-Turner (Bolton et al.
2000).
1982-92
Augmented AVC Era Intent and market structure also considered. Equilibrium was
claimed to be reached at 17% plaintiff success rate (Bolton
et al. 2000).
1986
Matsushita Electric
Industrial Company v.
Zenith Radio Corporation
American television manufacturers sued twenty-one
Japanese corporations that sold televisions in the United
States. Plaintiffs’ argument was that the defendants
conspired to drive them out of the U.S. market with
predatory pricing. The court’s assessment became one of the
most quoted in the following cases to come: “consensus…
that predatory pricing schemes are rarely tried, and even
more rarely successful. . .” (1986b, p.590) Thus, the court
did not find an economic motive for the defendants to
predate. Earlier verdict of the Court of Appeals was reversed
and remanded.
1989
A.A. Poultry Farms, Inc. v.
Rose Acre Farms, Inc.
The defendant’s prices were less than its average variable
costs for a period during (and below ATC throughout) the
price war. Predatory intent of defendant’s executives was
also documented: “We are going to run you out
of…business. Your days are numbered”(1989, p.1398). The
court focused on recoupment and concluded that intent by
itself did not help determine the probability of recoupment,
and in the absence of recoupment, even the most vicious
intent was considered harmless to the system.
1993
Brooke Group v. Brown &
Williamson Tobacco
Brooke Group (formerly known as Liggett) alleged that
Brown & Williamson Tobacco Corporation introduced its
own line of generic cigarettes (that the plaintiff had
pioneered with great success) and used predatory pricing to
stifle price competition in the economy segment of the
national cigarette market. Brooke argued that the defendant
used below cost pricing and offered discriminatory volume
rebates to wholesalers. Brown not only matched the
plaintiff’s retail price but also consistently undercut its
wholesale price. A harsh price and rebate promotion war
took place at the wholesale level that lasted eighteen months.
Ultimately, Brooke gave in and increased its prices. Generic
brand prices increased by 71% and branded cigarettes prices
increased by 39% whereas the costs where roughly constant
(Bolton et al. 2000). Even though below cost pricing and
predatory intent was documented, and the defendant was
found guilty by a jury, the Supreme Court required proof of
below cost pricing, and of recoupment, and subsequently
decided in favor of the defendant.
Table 3.1 (continued)
44
Defendants’Ultimate
Paradise Era Commences
Brooke decision led to excessive summary dismissal by
lower courts. Augmented AVC Era equilibrium destroyed.
Plaintiffs have not prevailed in a case since Brooke.
Summary dismissal is the norm. (Bolton et al. 2000)
1998
DOT Proposed Guidelines Department of Transportation (DOT) realized the strategic
problem and would allow proof of recoupment based on
reputation effects (and would not require proof of below
United States v. Microsoft Microsoft is accused of predatory pricing against Netscape
for bundling its Internet Explorer with its operating system
free of charge, and later settles the case.
1999
United States v. AMR
Corp.
Department of Justice (DOJ) filed complaint against
American Airlines based on strategic and reputation effects
in parallel with DOT guidelines.
2000
Presidential Elections George W. Bush stated during his election campaign that he
would only pursue price fixing antitrust cases (2000d).
DOT switches to case by
case approach
After evaluating thousands of responses from different
parties, DOT dropped proposed guidelines and decided to
take a case by case approach (January).
2001
American Case Dismissed Judge Morton dismissed the case basing on “time honored
rules” (April); DOJ decides to appeal the decision (July).
2002 -
Evolution continues… DOJ appeals the dismissal of American case but does not
pursue the objective aggressively in the aftermath of
September 11 events (the summary dismissal is later verified
by a panel of three judges in July 2003). American Antitrust
Institute observed that this marked “the death of predatory
pricing as a critical antitrust tool in this political climate” (cf.
Foer 2003, p.14). The low cost carriers start to make
headway into bleeding major carriers’ territories. Another
wave of predation against the discounters may be in order
after the major carriers get their act together, however the
discounters (e.g., Southwest, JetBlue, AirTran) now have
deeper pockets. As a preliminary step, the majors launch a
new wave of low-cost versions of themselves (e.g., Song for
Delta, and Ted for United Airlines)…
0
10
20
30
40
50
60
70
80
90
1
9
3
5
1
9
4
5
1
9
5
5
1
9
6
5
1
9
7
5
1
9
8
5
1
9
9
5
P
l
a
i
n
t
i
f
f
S
u
c
e
s
s
%
Figure 3.2: Approximated Predatory Pricing Plaintiff Court Success over Time
Table 3.1 (continued)
45
3.3.2 Regulatory Acts Regarding Predatory Pricing
Predatory pricing is a violation that is monitored by antitrust regulation bodies.
Antitrust regulation is constructed to prevent substandard industry performance
(efficiency) and promote competitive performance and equitable distribution of market
power. The typical targets of antitrust regulation are those industries with oligopolistic
structures with high entry barriers where market power stifles innovation. The following
are renowned Acts that were designed to regulate such behavior:
Sherman Act – 1890: Section 1 of this Act basically prohibited contracts or conspiracies
that restrain trade. The behavior considered illegal (e.g., price fixing, allocating territories
among competitors, tying purchases of one product to another) was defined by courts.
Section 1 makes concerted action illegal regardless of their market impact. If the
conspiracy is proven, the plaintiff does not have to prove actual or potential
monopolization (Hawker and Petty 1996).“Rule of Reason" doctrine, which suggested
analyzing the context of behavior to see if the case displays unreasonable restraint of
trade or legitimate business practice, was used for cases of ambiguous nature (Sullivan
1991). Section 2 declared monopolization and conspiracy to monopolize, as a felony.
Intent was defined further with subsequent legislation (Kovaleff 1994).
Clayton Act – 1914: Defined unlawful anticompetitive behavior and practices other than
monopolization (e.g., price discrimination (Section 2), exclusive dealing and tying
contracts, mergers (Section 7), and inter-locking directorates.) Section 7 was an important
46
antitrust statute in terms of its impact on litigation and structure. Cellar-Kefauver (1950)
made mergers and acquisitions that decrease competition or promote monopoly unlawful
(with respect to Section 7 of Clayton). This Act was intended to prevent unfavorable
market structures. However, labor unions and agricultural organizations (cooperatives)
were exempt (Martin 1959; States 1984).
Federal Trade Commission (FTC) Act – 1914: Extended and overlapped with the
Sherman Act in terms of restraints for trade. It defined certain conduct modes businesses
must compete in (e.g., Unfair Methods of Competition (Section 5)). The FTC was
authorized to interpret antitrust statutes. With the extension of Wheeler-Lea Act of 1938,
FTC was authorized to protect consumers as well as competitors (American Bar
Association. FTC Act Editorial Committee. 1981).
Robinson-Patman Act –1936: Amended Section 2 so that price discrimination that lessens
competition or promotes monopoly was considered illegal. The original Clayton Act had
exempted price discrimination in the form of quantity or volume discounts (1936).
Two main types of price discrimination were recognized. Primary discrimination
included injury to competing sellers, whereas secondary discrimination included injury to
competing buyers. A key case for primary discrimination was the Utah Pie Case (1967).
Pie producers in California, were accused of charging less for their pies in Utah than in
markets closer to their California plants, after Utah pie’s entry to the market. Court
decided in favor of Utah Pie, though Justice Stewart argued that the consumers benefited
from lower prices and increased competition in Utah (Van Cise and McCord 1969). A
47
key case for secondary discrimination was Morton Salt Case (1948). Morton gave
volume discounts to selected buyers. Large chain grocery stores were given the lowest
wholesale prices. Court found against Morton based on Robinson-Patman Act (1948).
It appears as if the Robinson-Patman Act was designed to protect small businesses
against the “chain-store revolution.” Secondary type of price discrimination was enforced
in the courts more, typically in the form of “mom and pop” grocery stores versus the
major food chain. However, today Robinson-Patman Act is interpreted rather strictly, and
is not considered a major factor in food retailing anymore (Dickinson 2003). The
standard for primary-line Robinson-Patman violations has become basically the same as
it is for ordinary predatory pricing under Section 2 of the Sherman Act (Kintner 1979).
Most predatory pricing cases fall under either an attempt or actual monopolization charge
under Section 2 of the Sherman Act or Section 2(a) of the Robinson-Patman Act
(McCareins 1996).
Antitrust legislation is enforced by the Antitrust Division of the Department of
Justice, and by FTC with some overlap. However, most lawsuits are filed by private
parties rather than the government agencies. The cases can be criminal or civil. Most
forms of price discrimination are not illegal themselves, though the market power
exercised to employ price discrimination may be illegal under Section 1 of the Sherman
Act. The Robinson-Patman Act extended antitrust for price discrimination to credit terms,
delivery times, quality, and volume discounts. It was designed to protect small
competitors rather than the competitive process. Robinson-Patman Act is easier to
pursue in courts as it requires proof of a reasonable possibility of substantial injury,
whereas Sherman requires a dangerous probability of actual monopolization. Still,
48
Robinson-Patman Act is not favored by the FTC and Department of Justice (DOJ) which
do not currently enforce the Robinson-Patman Act (Dickinson 2003; MacAvoy 2000).
There was also one Act that specifically regulated marketing behavior within a particular
industry --Agricultural Marketing Agreement Act –1937 (United States 1937).
3.3.3 Early Years and the Populist Era (1890 - 1975)
Very few cases were observed before the 1940’s. This inactive period continued
until the 1936 Robinson-Patman Act. Plaintiffs gained substantial power after Robinson-
Patman which protected smaller firms from price cutting by large ones. FTC initiated a
strong enforcement effort in the early 1940s. More lawsuits started to emerge. Oral and
written statements were used as evidence of intent. It was relatively easy to establish
predatory intent (Koller 1978). Koller (1971) reported that the Federal courts had
identified that predation had occurred in 95 out of the 123 cases. This equals to a 77 per
cent litigated case success rate for plaintiffs. This high rate may be due to the fact that
plaintiffs have won some cases “they probably should have lost. It seems no exaggeration
to call this the populist era of predatory pricing enforcement” (Bolton et al. 2000, p.14). It
was in this context when Justice Sullivan proposed that predatory behavior can be
identified by two non-cost criteria – whether it looks “jarring or unnatural”, and whether
it is aimed toward a particular target rather than an abstraction such as market share
(Sullivan 1977, p.112). Alan Greenspan commented that, “the entire structure of antitrust
statutes in this country is a jumble of economic irrationality and ignorance” (Greenspan
1962).
Despite the cases in which firms were found to employ predatory pricing, there
49
was a lack of economic theory supporting the presence and rationale of it. On the other
hand, Koller’s dissertation titled “The Myth of Predatory Pricing” (1971), and the
literature stemming from this work, was relied upon by the Chicago School and cited by
the courts and influential academics such as Areeda and Turner (1975).
3.3.4 Areeda-Turner Era (1975- 1982)
A new era started to emerge in the 1970’s, in which law’s condemnation of many
forms of predation was criticized. This notion gained significant support among the
Federal judiciary and antitrust scholars (Bernstein 2001), and the legal environment that
favored the plaintiffs came to a halt with the publication of the seminal 1975 Areeda-
Turner article. The Areeda-Turner rule basically stated that a firm should be found guilty
of predatory pricing whenever it sets price less than its marginal cost (i.e., the cost of
material and labor in making the last unit, excluding the startup/fixed costs). It followed
that any price set equal to or above the firm’s marginal cost is non-predatory. According
to Areeda-Turner, competition in an industry would naturally drive prices toward the
marginal costs. Pricing below marginal cost, meant operating at a loss, and was
considered irrational, except for the intention to drive out competitors, which was
considered predatory. After competitors were driven out of business, a predatory
monopolist could potentially recoup its losses by charging supra-competitive prices,
which hurt the welfare of the consumers. Areeda and Turner also recognized that
marginal cost data (i.e., how costs vary with each additional unit of output) were not easy
to compute. In practice, marginal cost was more of a conceptual tool for economists.
Thus, Areeda and Turner suggested the use of a per se standard of Average Variable Cost
50
(AVC) as a substitute for marginal cost. AVC is calculated by identifying those costs that
vary with output, adding them up, and dividing the result by the total number of units
produced (Pass and Lowes 1994). Even though there are differences between MC and
AVC, a firm with prices below AVC is not even covering the variable costs, not to
mention its fixed costs. Deductive reasoning follows that, in the absence of a justification,
a firm with below AVC pricing must be conducting predatory pricing since the only
profits (rationale) would come from rewards through the outcome of predation.
MC
AVC
AC
Price
Output
0
Po
Perf.
Comp .
Demand for
Perfect
Comp etit ion
Q o
Demand for
a monop olist
Po
Mon.
Figure 3.3: Cost Relationships
With the lack of a major alternative theoretical model of predatory pricing to
consider, the courts embraced Areeda-Turner AVC rule and replaced the uncertain
factors adopted previously. As a result, the legal trend was totally reversed against the
plaintiffs. Plaintiffs’ success rate immediately fell down to only eight per cent as opposed
to seventy-seven per cent during the populist era (Hurwitz and Kovacic 1982). Areeda
and Turner also spread the view that predatory pricing is a rare event in practice (Brodley
51
and Hay 1981).
3.3.4.1 Counter Cost and Non-cost Proposals
Despite its increasing popularity in the courts, Areeda-Turner rule was criticized
by economists because it did not capture the strategic factors and long run welfare
effects. Sharp price reductions can also be viewed as market signaling that communicate
threats and sanctions. Alternative cost and non-cost standards were proposed to overcome
the practical drawbacks of Areeda-Turner rule, but none of them could replace the
precedent in court practice (Bolton et al. 2000, p.15):
Alternative Non-cost Standards: Some economists argued that predatory pricing is a
matter of intent, not costs. The price could even be set higher than ATC for a case to be
considered predatory (Shepherd 1986).
Williamson Output Increase Rule. Williamson (1977) suggested that predatory pricing be
evaluated as a long-run strategy, particularly from the perspective of the incumbent firm's
response to entry. Temporary price cuts have negligible benefits and long-term welfare
problems occur when the predator raises prices after competitors exit. Williamson
concluded that his rule would have superior welfare consequences to the Areeda-Turner
cost rule where strategic responses were not taken into consideration. Williamson output
increase rule basically stated that it would be considered predatory conduct if a firm
raised its output significantly within the twelve to eighteen months following a rival’s
entry.
52
Baumol Price Reversal Rule. Baumol (1979) proposed that it is predatory conduct if the
incumbent first decreases the price sharply, but then increases it again after the
competitor exits, and this is not accounted for by a rise in cost or demand.
Alternative Cost Standards: In order to replace Areeda-Turner, other counter cost rules
have also been proposed.
Joskow and Klevorick (Two-stage rule). Under the Areeda-Turner rule, a firm covering
its AVC is considered lawful yet the fixed costs remain unaccounted for in the equation.
Average Total Cost (ATC) includes both fixed and variable costs and it was argued that
pricing below ATC for a considerable period could also be predatory. Thus, even though
the firm may be pricing above AVC, it is still informative to compare the price to ATC.
In one of the more comprehensive alternative proposals to Areeda-Turner, Joskow and
Klevorick (1979) basically argued that a price below AVC is always predatory, and a
price greater than AVC but less than ATC is predatory unless the defendant shows that it
has or had a reasonable justification for the price:
Therefore, the adoption of a strategy of pricing below average variable
cost by a dominant firm confronted with entry is sufficient to demonstrate
predation. A price below average variable cost, and for that matter, a
price below average total cost, could not possibly be sustained in the long
run since, to survive, firms must cover total costs in the long run. A firm
with market power-the ability to control price-would only have an
incentive to impose losses on itself when faced with an entrant if the
promise of future monopoly gains made such a tactic profitable from a
long-run perspective (Joskow and Klevorick 1979, p.252).
In their two-stage test Joskow and Klevorick (1979) would first focus on the
market structure to determine if predation is likely to be successful. This meant a careful
examination of the entry and exit costs. They argued that screening out markets where
53
predatory pricing was unlikely would ensure not discouraging truly competitive price
competition. The assumption was that only firms with market power would have
incentives to exclude rivals through predation. On the second stage, they proposed to test
if price was below cost. They would also determine if Baumol price reversal happened
(price cut, exit of competitor, price rise), and see if corporate predatory intent was
documented.
Average Total Incremental Cost. Incremental costs are defined as the costs that will be
incurred as the result of a decision (e.g., capacity increase). In other words, it is the
difference between the firm’s total cost at the two output levels divided by the change in
output. Average Incremental Cost (AIC) (also known as Average Avoidable Cost
(AAC)), is the unit cost of the added output, or in other words, the cost that would have
been avoided had the additional amount of output not been produced (Baumol and Sidak
1994). As the time frame of the cost measure gets longer, more costs become avoidable,
because there is a greater opportunity for redeploying assets that are fixed in the short
run. Therefore, average long run incremental cost (LAIC) is the incremental cost when all
costs are considered variable, and serves as a natural upper threshold.
Baumol (1996) argued that combinations of the firm’s products had to be
considered for a proper Areeda-Turner test. Accordingly, the price of each product by
itself had to equal or exceed that item’s average avoidable cost. Moreover, any
combination of the firm’s products had to be priced so that the incremental revenue had
to exceed the avoidable cost incurred by that combination of products.
Still, there are a number of instances in which the odor of predation is
strong, as when an entrant airline with its six-plane fleet, operating on
54
almost as many routes, proposes to fly a route coveted by a large
incumbent airline, whereon the latter announces that it will open for
business (for the first time) along each of the most promising of the
entrant’s routes. Analogous examples in which predatory pricing is the
issue are also easily imagined. There is reason to provide the entrant in
such a scenario effective recourse against overaggressive acts by the large
incumbent. Accordingly, the rules for proper execution of an average
variable cost test that are described in this article are designed not to
offer undue protection to the firm suspected of predatory pricing (Baumol
1996, p.52).
Other proposals included that of Posner (1976) (required AVC test, high market
concentration and proof of intent), Scherer (1976) (full rule of reason inquiry with focus
on intent and market structure), and Ordover and Willig (1981) (predatory conduct if
benefits depend on added market power through forced exit).
The following statement summarizes the position of most courts regarding cost
before the decisive Brooke Case (1992b):
If the defendant’s prices were below average total cost but above average
variable cost, the plaintiff bears the burden of showing defendant’s
pricing was predatory. If, however, the plaintiff proves that the
defendant’s prices were below average variable cost, the plaintiff has
established a prima facie case of predatory pricing and the burden shifts
to the defendant to prove that the prices were justified without regard to
any anticipated destructive effect they might have on competitor.
As summarized in the above statement, AVC and ATC were commonly used as
lower and upper thresholds for proving predatory pricing by the courts. However, some
researchers observed that this standard was not appropriate for certain industries (e.g.,
telecommunications) (Grout 2000). Joskow and Klevorick (1979) were the first to
recommend the use of long run average incremental cost as an upper threshold. Grout
(2000) evaluated alternative thresholds and concurred with Joskow and Klevorick in that
LAIC was the most appropriate upper threshold for high fixed cost industries with
55
multiple product catalogs.
3.3.5 Augmented AVC rule era (1982 – 1992)
The lower courts faced with difficulties when they tried to implement the AVC
standard as proposed by Areeda-Turner. Cost was very hard to determine, and there were
many criticisms toward the use of a per se short term cost test. The fact that no single
plaintiff won a case during the five years following Areeda-Turner made the courts “a
defendant’s paradise” (cf. Bolton et al. 2000, p.18).
In the absence of a binding Supreme Court example, lower courts adopted what
was called an augmented AVC rule, which also took intent and market structure into
consideration. High concentration in markets and barriers to entry were recognized as
enablers of recoupment. During the ten year augmented AVC rule period until the Brooke
Decision (1983-1993), the plaintiff success rate increased to seventeen per cent and
probably would have been much higher if out of court settlements were included in that
figure. Bolton et al. (2000) suggested that “a more or less satisfactory equilibrium” was
reached during the augmented AVC era. Dismissal of cases by summary judgment was
common, and cases remained tough to win, yet the occasional excessive jury awards may
have prevented many cases of predatory pricing from taking place.
3.3.6 The Recoupment requirement Era (1993 - )
Recoupment element started to gain importance in the courts during late eighties
as evidenced by three cases leading to the Brooke Decision --Matshushita (1986), Cargill
(1986a), and Rose Acre (1989). The reasoning used in the Matsushita Case required that
56
plaintiffs prove not only below cost pricing, but also that the alleged predator either
recouped its losses or at least had a very high probability of recoupment. Recoupment
would be through succeeding in driving competitors out of the market, restricting output,
and raising prices (Sheffet and Petty 1994). During the same year, the verdict of the
Cargill Case, in which the plaintiff had alleged that a proposed merger had antitrust
implications by enabling the defendants to implement a price-cost squeeze strategy
against the plaintiff, was reversed. The Supreme Court held that the proposed merger did
not constitute antitrust harm and that the Court of Appeals had erred (1986a). In the Rose
Acre Case, the court decided to analyze the possibility of recoupment before analyzing
below cost pricing. The rationale was that the determination of likelihood of recoupment
was easier than undertaking the cost analyses. It would not be necessary to analyze below
cost pricing if recoupment was not considered to be possible (Calvani 1999). The
recoupment element evolved into today’s final and strict form with the critical Brooke
case verdict.
The Brooke decision established a framework in which the plaintiffs were
required not only to prove below-cost pricing, but also recoupment of losses suffered
during predation. No plaintiff has been able prevail in the courts since the Brooke
Decision (Bolton et al. 2000). However, if a company can ever prove that an incumbent
used below-cost pricing and also intended to recoup its losses later through supra-pricing,
than the courts will rule that this was illegal conduct because the incumbent has harmed
the welfare of the society by diminishing the efficiency of the allocation of resources
across the society (Guiltinan 1996). Yet, the Supreme Court did not define what an
appropriate measure of cost is, or elaborate on the issue of recoupment sufficiency.
57
3.3.6.1 Brooke Case
A brief summary of Brooke is provided here and the case is discussed in a later
section. The market in the Brooke Case consisted of six main players with little price
competition, resulting in high profitability. The plaintiff was the first to introduce a
generic and less expensive product. As a result of the introduction of this new brand
category, smokers of both the plaintiff’s and the competition’s brand switched to the less
expensive product, increasing the “generic” sales. The defendant (third largest player
with 12% market share) introduced counter generic cigarettes with wholesale prices even
lower than that of the plaintiff’s. A price war took place during which the plaintiff
claimed that the defendant sold its generic product at a loss. It was claimed that the
defendant was trying to force the plaintiff to increase its generic product prices in an
effort to minimize the cannibalization of the defendant’s premium brand and maintain its
profitability. The jury found the defendant guilty, however that verdict was overruled,
reasoning that recoupment was not probable (1993).
First a plaintiff … must prove that the prices complained of are below an
appropriate measure of its rival’s costs … second … is a demonstration
that the competitor has a reasonable prospect [under Section 2a of the
Robinson-Patman Act], or under Section 2 of the Sherman Act, a
dangerous probability, of recouping its investment in below-cost prices
(1993, pp.4702-703).
Starting with Brooke, recoupment was perceived to be the ultimate objective of
predatory pricing. Unless recoupment was probable, the general view held that price
wars (and predatory pricing attempts that failed) were beneficial for the consumers.
The court admitted that below-cost tests might not always be effective in determining
whether predatory pricing was employed. Nevertheless, it chose to dismiss the
58
possibility in its administration, arguing that it would be beyond the practical ability
of any jurisdiction to control:
As a general rule, the exclusionary effect of prices above a relevant
measure of cost either reflects the lower cost structure of the alleged
predator, and so represents competition on the merits, or is beyond the
practical ability of a judicial tribunal to control without courting
intolerable risks of chilling legitimate price-cutting. (1993 p. 223)
With Brooke, the Supreme Court also welcomed the idea of dismissals by
summary judgment in the absence of proof of below-cost pricing and recoupment. The
lower courts embraced the invitation and have dismissed a vast majority of the predatory
pricing cases since Brooke.
3.3.7 The Emerging School of Thought
It has been noted that Brooke decision was the most important predatory pricing
decision of the modern times. It set a milestone in jurisdiction, and was particularly
crucial, as it seemed to destroy the satisfactory equilibrium that was achieved with the
augmented AVC rule. It is important to note that almost all predatory pricing cases have
been dismissed by summary judgment and that no plaintiff has prevailed in the courts
since Brooke. The reason may be that the Supreme Court encouraged lower courts to
dismiss predatory pricing cases by summary means (at the lack of market concentration,
entry barriers, and capacity to absorb additional market share). Another explanation could
be the Supreme Court’s skepticism that predation can be a rational business strategy and
its disregard of modern theories of predatory pricing. Today, the survival chances of a
predatory pricing case beyond the summary judgment phase are slim (Bolton et al. 2000).
However, the fact that reputation effects were considered plausible in at least one case
59
(1995) was a positive development. The courts may start to integrate the thinking from
modern theories into the jurisdiction in the near future. DOJ’s civil complaint against
American Airlines was also based partially on reputation effects.
The proof requirement for recoupment of losses as well as below-cost pricing put
a heavy burden on plaintiffs to prove their case. The Supreme Court "sent a chilling
message to predatory pricing plaintiffs" with the Brooke decision (Denger and Herfort
1994). Following Brooke, it became very difficult to successfully prove a predatory
pricing case in most markets either with below-cost or recoupment elements (Meeks
1998). Proof of recoupment was a legal element additionally required by the Brooke
Case.
The critical Brooke Decision was not without its critics. Bolton et al. (2000)
argued that the notion that predatory pricing was “rare and implausible” stems from early
and old economic theory and that the new approaches have to be used to assess predatory
pricing cases. They argued that the Supreme Court failed to take a strategic approach to
develop an understanding of the case. The defendant successfully deterred aggressive
pricing for an extended period in the future. Aggressive pricing was also effectively
prevented in other markets (e.g., branded cigarettes).
The limits of the existing enforcement stemming from earlier decisions including
that of the Brooke Case were increasingly criticized. For example, Guiltinan and
Gundlach (1996b, p.88) explained that predatory pricing litigation relied on the narrow
assumptions of neoclassical price theory. They questioned the underlying assumptions of
the theory, which are exclusive motivation for profit maximization, possession of perfect
information, and calculated rationality in decision making. They also questioned whether
60
or not the welfare of the consumers should be judged solely on the basis of allocative
efficiency. Ordover and Willig (1981) suggested that opportunity cost of capacity
decisions should also be considered. The lack of strategic theory approach and lack of
consideration of game theoretic models were also criticized (Bolton et al. 2000).
Predatory conduct in many industries and the inability of neo-classical theory to detect
and deal with it became increasingly apparent. Growing literature on the inability of the
courts to deal with the problem (Bolton et al. 2000), and the observation of anti-
competitive affects of predation on the airline industry (Oster and Strong 2001), led DOT
to draft and propose guidelines for governing the airline industry.
DOT Guidelines: The traditional assumption of the Chicago School does not hold well in
industries where there is a high fixed cost structure. That was the reason why DOT felt
the need to propose a set of guidelines to govern anti-competitive conduct in the airline
industry in 1998. Interestingly, DOT guidelines allowed for reputation effects (expected
gains from deterring future entry by competition) as evidence for predatory pricing, and
did not require proof of below-cost pricing (1998d). In essence, DOT was trying to
update the rules that defined anticompetitive conduct for the airline industry. These
guidelines are detailed in Chapter 6.
DOJ Efforts: DOJ Antitrust Division Section Chief Roger Fones announced that they
would not be using Areeda-Turner (AVC) test, but would be focusing on a measure of
costs that the alleged predator “could have avoided had it not embarked upon the
pricing/capacity strategy under review” (Fones 1997). It was stressed that the division
61
would not look into the very short-term average avoidable cost, but would have a short to
medium run perspective. He made many remarks regarding the airline industry including
“a popular misconception that predation cannot occur in the airline industry because the
avoidable cost of filling an empty seat is low” (Fones 1997). Around the same time as
this announcement, DOJ also took a proactive stance in detecting and preventing antitrust
violations by suppliers. They communicated with general counsels of corporations
offering to make presentations on bid rigging, price fixing, and market allocation, and
distributed a booklet named "What You Should Know About Detecting and Preventing
Antitrust Violations, Antitrust Primer for Procurement Officials" (1997b).
DOJ filed a civil complaint against American Airlines based on reputation effects
in parallel with the DOT guidelines in 1999. American Airlines case was the first DOJ
brought in more than 20 years. It was seen as a tough case because of the very high
standards of proof imposed on the plaintiffs since the Brooke Decision (Carney and
Zellner 2000). The case is detailed in Chapter 6.
3.3.8 Differing Practices among States:
The Federal courts generally embrace the notion arising from the early work of
McGee (1958), Koller (1971) and Areeda and Turner (1975) that predatory pricing is
irrational. Since the Brooke (1993) decision, the Federal courts have required proof of not
only below-cost pricing but also recoupment of losses suffered through such pricing
(1993).
State courts have generally been generally less demanding and more hospitable to
predatory pricing plaintiffs than the federal courts (Sheffet and Petty 1994). For example,
62
Wal-Mart was found guilty in an Arkansas district court even after the Brooke decision,
in a verdict that was later reversed by higher courts (Hawker and Petty 1996). Many
states have laws to treat predatory pricing in form of sales below cost statutes or
minimum mark-up laws (Haynes 1988). These laws can be generic in nature or may
address concerns for competition in specific markets such as gasoline sales. Some states
require proof of below cost sales, and some others still adopt a focus from the populist
era such as of harm to competitors, rather than harm to competition (Calvani 1999). Cost
criterion used by the states is usually Average Total Cost (ATC) or a form of AVC or
ATC. State courts infer intent of predatory pricing action either from the below-cost
behavior or from other means. The concern for bankruptcy of small competitors is higher
at the state level than that at the federal level (Guiltinan and Gundlach 1996). Some
states use a sliding scale requiring differing levels of proof depending on how much the
price was below ATC (Bruckmann et al. 1995). Appendix A includes an unexhaustive
survey of the states regarding their antitrust laws with respect to predatory pricing.
3.3.9 Insights from European Union Regulation:
Even though the outcomes of the U.S. litigation affect the rest of the world and
the U.S. appears seemingly unaffected from other country influences, it is still
worthwhile to take a brief look to how a continent with a long industrial history is dealing
with the predatory pricing issue.
Europe appears to be more aggressive in its antitrust enforcement. Whether above
AVC pricing can be predatory is debated. The European Commission comfortably
employed AIC cost but one with a longer frame in order to not set the floor too low for
63
predatory claims. Proof of recoupment is not required, hence a defendant may be found
guilty even when it had no reasonable prospect of recouping losses made through
predatory pricing (1994c).
The European Commission also recognizes the uniqueness of network industries
due to the high overhead involved. For example, there is a general understanding that the
average variable cost rule does not normally apply to the telecommunications industry
since “the variable costs of providing access to an already existing network are almost
zero” (1998b). Similarly, the 1998 Competition Act acknowledges that AVC tests may
not be relevant for regulated industries since their variable costs may be close to zero.
There is a general agreement among the policy-makers that alternative approaches may
be necessary for such industries. Long run incremental cost (LRIC) is considered to be an
acceptable test for such industries as it includes both capital and operating costs of the
increments. As long as the price is above LRIC, the decision will be profitable, hence
rational and non-predatory. In particular, LRIC is the accepted test for the
telecommunications industry in particular. As for the time frame, neither the very short,
nor the very long time frame is considered appropriate. However, the Commission has
admitted that it may have to examine AIC of longer than a year (1998b).
Moreover, the Commission has employed and is supporting the use of
combinatorial cost tests for the telecommunications sector. Combinatorial tests can be
defined as a sequence of tests where the revenue earned from each service or combination
of services must cover the total incremental cost of adding that collection of services to
the remaining services (2000a). Unfortunately, combinatorial tests can be very hard to
calculate since it requires all combinations to be considered (2000a).
64
3.4 Critical Cases for the Evolution of Case Law
The evolution of predatory pricing can probably be best observed through a study
of important cases through history. The following cases were selected since they
represent a summary of the evolution of jurisdiction over time.
3.4.1 Standard Oil Co. v. United States, 221 U.S. (1911)
Standard Oil of the Rockefellers was found guilty in this classic case of
monopolization. A major component of the case involved the allegations that Standard
Oil had employed predatory pricing to drive its competitors either out of business or to
force them to sell it to Standard Oil at distressed prices (1911). It had aggressively
purchased 223 independent companies prior to 1907 (McGee 1958). As a result of the
case, it was broken up into thirty-three geographically distinct companies (Gibb and
Knowlton 1965).
Standard Oil was accused of engaging in predatory pricing (pricing below ATC)
to drive competitors out of business and then raising price above ATC to recoup losses.
Standard was also accused of buying key input suppliers (i.e., pipeline companies) to
control the market, of using its market power position for negotiating discriminatory rail
freight rates, and of engaging in business espionage. Even though McGee (1958) later
disputed the verdict of the case and argued that no predation actually took place, other
researchers have shown that subtle predatory tactics were employed: United Rail charged
cartel prices to the whole oil industry except for Standard Oil which got rebates. Standard
Oil and the railroads shared the cartel freight profits. By effectively raising the cost of its
65
rivals, Standard Oil was then able acquire them at distressed prices (Granitz and Klein
1996). Predatory pricing coupled with acquisition is more viable because the period
during which the predator incurs losses is shortened and market power is further
enhanced. Celler-Kefauver Amendment (1950) to the Clayton Act section 7 prohibited
the creation of monopoly through horizontal merger.
3.4.2 Utah Pie Co. v. Continental Baking Co, 386 U.S. (1967)
In this classical example of the populist era, Utah Pie, the leading vendor of
frozen pies in its market, brought a suit under Section 2(a) of the Robinson-Patman Act,
against three national bakeries alleging that they had engaged in a geographically focused
predatory pricing campaign to increase their local market share. The court concluded that
the defendants’ charged less for their pies in Utah than elsewhere. The plaintiff’s market
share had diminished from 66.5% to 45% during the forty-four month price war. The
Supreme Court reinstated the jury verdict for the plaintiff (1967; Calvani 1999).
Interestingly, the sales volume of the plaintiff had increased and the firm had
remained profitable throughout the price war. The court suggested that ATC could be
considered an appropriate standard for below-cost. It also “left open the possibility that
prices above cost may be predatory if coupled with evidence of anticompetitive intent
and a deteriorating price structure. The Court did not comment on whether the evidence
in that case would support a predatory pricing claim under the Sherman Act” (McCareins
1996).
The Utah Pie decision was widely criticized. Nevertheless, it marked the high
time of early era predation litigation. The winning predation cases of the time were
66
characterized by the large predators, geographic price discrimination, sales below average
total costs, and predatory intent. Dissenting Justice Stewart then observed about the Utah
Pie Case that “if we assume that the price discrimination proven against the respondents
had any effect on competition, that effect must have been beneficent. The Court has
fallen into the error of reading the [statute] as protecting competitors, instead of
competition” (cf. Calvani 1999, p.5).
3.4.3 Matsushita Electric Industrial Company v. Zenith Radio Corporation 475 U.S
(1986)
In 1986, American television manufacturers sued twenty-one Japanese
corporations that sell televisions in the United States. Plaintiffs’ argument was that the
defendants conspired to drive them out of the U.S. market with predatory pricing. The
court’s logic, and assessment of the Matsushita case became one of the most quoted in
the following cases to come:
Any agreement to price below the competitive level requires the
conspirators to forgo profits that free competition would offer them. The
forgone profits may be considered an investment in the future. For the
investment to be rational, the conspirators must have a reasonable
expectation of recovering, in the form of later monopoly profits, more than
the losses suffered. . . .[T]he success of such schemes is inherently
uncertain: the short-run loss is definite, but the long-run gain depends on
successfully neutralizing the competition. Moreover, it is not simply to
achieve monopoly power, as monopoly pricing may breed quick entry by
new competitors eager to share in the excess profits. The success of any
predatory scheme depends on maintaining monopoly power for long
enough both to recoup the predator's losses and to harvest some
additional gain. Absent some assurance that the hoped-for monopoly will
materialize, and that it can be sustained for a significant period of time,
"[t]he predator must make a substantial investment with no assurance that
it will pay off." . . . For this reason, there is a consensus among
commentators that predatory pricing schemes are rarely tried, and even
more rarely successful. . . . (underline added) These observations apply
67
even to predatory pricing by a single firm seeking monopoly power
(1986b).
Following the above argument, the Supreme Court concluded that the defendants
did not have a significant chance to achieve monopoly power and subsequently raise
prices in the U.S. market. The court also did not find that the industry had high barriers to
entry. Since the court did not find an economic motive for the defendants to predate, it
demanded the plaintiffs to provide more persuasive evidence. Earlier verdict of the Court
of Appeals was reversed and remanded. An appropriate measure of cost was not defined
(McCareins 1996).
3.4.4 A.A. Poultry Farms, Inc. v. Rose Acre Farms, Inc. 881 F.2d (1989)
The case was regarding a price war among egg producers. It was testified by the
plaintiff’s expert that the defendant’s prices were below its average total costs during the
war. The defendant’s prices had also been less than its average variable costs for a period
during the price war. Predatory intent of defendant’s executives was also documented:
“We are going to run you out of…business. Your days are numbered.” (1989, p.1398).
The court stated that it was much easier to conclude from the structure of the
market that recoupment was not probable, than it is to determine the appropriate measure
of cost and measure cost. Thus, the court decided to first analyze the probability of
recoupment before applying a cost test. If recoupment was not possible then it could be
inferred that the low price was not predatory even if it were below cost:
Predatory prices are an investment in a future monopoly, a sacrifice of
today’s profits for tomorrow’s. The investment must be recouped. If a
monopoly price later is impossible, then the sequence is unprofitable and
we may infer that the low price now is not predatory. More importantly, if
there can be no “later” in which recoupment could occur, then the
consumer is an unambiguous beneficiary even if the current price is less
68
than the cost of production. Price less than cost today, followed by the
competitive price tomorrow, bestows a gift on consumers. Because
antitrust laws are designed for the benefit of consumers, not
competitors…, a gift of this kind is not actionable. (1989, p.1401)
The court also concluded that intent by itself did not help determine the
probability of recoupment, and in the absence of recoupment, even the most vicious
intent was harmless to the system: “Entrepreneurs who work hardest to cut their prices
will do the most damage to their rivals….If courts use the vigorous, nasty pursuit of sales
as evidence of a forbidden “intent”, they run the risk of penalizing the motive forces of
competition” (Calvani 1999, p.8). Not surprisingly, the plaintiffs did not prevail in this
case.
3.4.5 Brooke Group v. Brown & Williamson Tobacco 509 U.S. (1993)
The Supreme Court decision regarding the Brooke case has been considered to be
“the most important predatory pricing decision in modern times.” No single plaintiff has
won a case of predatory pricing in the Federal courts since the Brooke Decision was
made (Bolton et al. 2000). The Supreme Court developed a two-stage proof framework
for analyzing predatory pricing claims during the Brooke Case that no plaintiff has been
able to survive since (Bolton et al. 2000).
Brooke Group (formerly known as Liggett) alleged that Brown & Williamson
Tobacco Corporation introduced its own line of cigarettes in the generic segment, which
Brooke Group had pioneered, and used predatory pricing to stifle price competition in the
economy segment of the national cigarette market. Liggett argued that the defendant used
below cost pricing and offered discriminatory volume rebates to wholesalers. The
69
allegation was that the defendant conspired to force Liggett to raise its own generic
cigarette prices so that it could introduce oligopoly pricing in the economy segment.
R.J. Reynolds had 28 per cent and Philip Morris 40 per cent market share at the
time of the trial. The defendant was a distant third with about 12 per cent. The plaintiff’s
share (once 20 per cent at its peak) was just around two per cent in 1980 and around five
per cent in 1984. On the verge of bankruptcy, the plaintiff introduced low-cost generic
(black and white) cigarettes in 1980. This new category of cigarettes was around 30 per
cent less expensive and it was an immediate success.
As the success of the generics became apparent, larger firms responded. The
defendant was the manufacturer that was hit the hardest. Even though they sold around
11% of the branded cigarettes, their customers were the most price-sensitive and 20% of
them had switched to plaintiff’s generic cigarettes. The defendant introduced its own
generic cigarettes in 1984. The defendant was not the first to respond to Brooke’s generic
cigarettes, indeed R.J. Reynolds had already repositioned one of its existing brands as
generic by 1984 (1993).
However, the defendant not only matched the plaintiff’s price at the retail level
but also consistently undercut it at the wholesale level. A harsh price and rebate
promotion war took place at the wholesale level. After the war, the plaintiff gave in and
increased its prices. Generic brand prices increased by 71% and branded cigarettes prices
increased by 39% (Bolton et al. 2000).
The Supreme Court held that Section 2 of the Sherman Act is not violated unless
plaintiff proves not only that defendant's prices were ‘below an appropriate measure of …
costs, but also that defendant had a dangerous probability, of recouping its investment in
70
below-cost prices” (1993). Unable to meet the burden of the increased level of proof, the
plaintiff lost the case despite strong evidence of predation. The defendant had cut its
prices below costs, had predatory intent, had decreased its output and increased prices
following the period of price war even though its costs were pretty much constant. The
parties in Brooke had both agreed that AVC was the appropriate standard therefore the
Court did not discuss the issue further (Watson 1998).
The Brooke case brought clarity to a couple of issues to the disadvantage of the
plaintiffs: below cost pricing would be a prerequisite to predatory pricing (it followed
that a firm can not be held liable of predatory pricing if its prices are above its costs), and
proof of recoupment (in form of a reasonable prospect or a dangerous probability) would
be required. Condemning above-cost price cuts would be “beyond the practical ability of
a judicial tribunal to control without courting intolerable risks of chilling legitimate price-
cutting” (1993). Unfortunately, this message enables the incumbents in network
industries (which have very low AVC structure) to employ predatory pricing legitimately
with above (variable) cost prices.
The logic for the recoupment aspect was as follows: the predator needs to recoup
its losses of predation to be profitable in the long run (i.e., rational), if predatory pricing
does not result in elimination of the rival, then recoupment cannot occur. If there is no
recoupment and the competitor is not eliminated, there is no harm to competition and
consumers have simply benefited from the low prices. If recoupment is not possible then
summary dismissal of a case should be appropriate. This view also shared by the Chicago
School of thought, has its flaws, discussed in the Post-Chicago School of thought section
of this chapter.
71
3.4.6 United States v. AMR Corp. et al. (American Airlines) (1999)
--dismissed (2001)
The 1999 DOJ suit against American Airlines was the first predatory pricing suit
brought forward in more than twenty years (Carney and Zellner 2000). Following the
DOT guidelines in spirit, the prosecutors wanted to set an example for the industry. DOJ
basically alleged that American Airlines used a temporary capacity expansion and fare
reductions to drive new entrants out of its Dallas/Ft. Worth (DFW) hub (1999c). DOJ
also had evidence of intent:
“If you are not going to get them [LCCs] out then no point to diminish
profit.” Don Carty, Chairman and CEO of American Airlines, 1996
DOJ’s complaint alleged that American Airlines dominated many of the routes
from its Dallas hub and charged monopoly prices. American controlled seventy percent
of the flight capacity from DFW. Low-cost carriers (LCCs) were proven to have positive
impact on consumer welfare through lower prices. DOJ complained that American
Airlines cut its prices to a level that would not make business sense except if it could
drive LCCs out of DFW before they could get a foothold in the market. Vanguard, Sun
Jet, and Western Pacific were victims of the alleged predatory strategy (1999c). All of
them were successfully driven out of the DFW hub and subsequently out of business. The
DOJ built its case incorporating elements of game theory and strategic economic theory.
These allegations, if proven, should have met the burden of predation (Piraino 2000).
The case was to go on trial in Wichita, Kansas on May 22, 2001. However, the
Federal court dismissed the case by summary judgment before the trial date. The
reasoning was derived from the Brooke Decision. The government had failed to show
72
that American’s pricing was below cost and that recoupment was highly probable. DOJ’s
modern approach to the case was perceived as an attempt to change antitrust law (Priest
2001), and time-honored rules (case memorandum and order 2001). With the new
administration in place, it is not likely that DOJ will be aggressive to bring a new case
forward. However, the decision to simply apply the Brooke philosophy to a network
industry case basically made parts of the new economy (e.g., airlines, software,
semiconductors, bio-technology) exempt from antitrust law (Carney 2001). Even so,
Judge Morten concluded:
The government's claims in the present case fail because American did not
price below an appropriate measure of cost, because it at most matched
the prices of its competitors, and because there is no dangerous
probability (even assuming below-cost pricing) of recoupment of
American's supposed profits by means of supra-competitive pricing. With
respect to costs, the evidence shows that American priced its fares
consistently above its average variable costs. Alternative, creative
measures of costs proposed by the government are inconsistent with
existing law, and inconsistent with an antitrust regime which seeks to
nurture rather than throttle vigorous price competition. With
respect to the question of recoupment, the government's claims suffer from
a pervasive failure of proof…Actual or likely recoupment by supra-
competitive pricing finds no basis in the evidence…The government's
theory of liability by reputation is not the law, and should not be. A
fundamental principle of antitrust law is that it be capable of effective and
accurate administration, and not chill the competition it seeks to foster.
The government's reputational liability approach would violate this
principle, permitting claims of predation based solely upon the subjective
and unverifiable complaints of a defendant's competitors. The low fare
carriers in question entered the core markets seeking to play a new sort of
ball game. The government's theory — that an established competitor
should not, and indeed, cannot deviate from its existing market strategy in
the face of aggressive price cutting by a new entrant — represents a whole
new mid-game spin on time-honored rules. Here American played by the
traditional rules. It competed with the low fare carriers on their own
terms. It did not price its fares below cost; it did not undercut the other
carriers' fares... Summary judgment is appropriate (U.S. v. AMR Corp et
al., Summary Judgment, 2001, pp.136-37).
73
3.5 Legal Elements of Proof
A look into the legal elements to prove for winning a case is insightful for an
understanding of the mechanics of the case trials. This section is based on McCareins
(1996).
Appropriate Measure of Cost: The courts have not agreed on a definition of an
appropriate measure for the cost tests. Appendix A: State Law Survey demonstrates that
the states employ many different standards. Similarly, Federal circuits do not have a
general standard. AVC and ATC are generally used as thresholds for burdens of proof.
The reason for the lack of a generally accepted measure is due to the understandable
hesitation of the Supreme Court to define it (McCareins 1996). However, twelve Federal
Appellate Courts have typically interpreted the appropriate measure as selling below
average variable cost.
Aggregation of Products: Whether cost calculations should be done with respect to a
single product, a product line, or a particular outlet or production facility has not been
defined. Some circuits have favored an assessment of predatory pricing based on the
costs and revenues associated with a full product line, as long as the competition was not
limited to a narrower scope (McCareins 1996).
Classification of Costs. There involve heavily debated issues such as determining what is
an appropriate measure of cost (average variable cost, average total cost, average
avoidable/incremental cost etc.), and classification of the incurred costs as variable, fixed
or avoidable depending on the definition of relevant market(s) and choice of time
horizon. Accounting and inventory systems vary; the distinction between costs and
investments is not clear cut, and it is hard to attribute costs to a specific product line or a
74
product in the often encountered case of multiple product lines. More recently, the
complexity of these issues has been further elevated where whether and how above-cost
pricing can be predatory is being examined (e.g., Edlin 2002; Meeks 1998).
Feasibility of Recoupment: This element refers to the proof that the predator will actually
profit from predation. This would require that once the rival is eliminated, the predator
would be able charge monopoly prices. The predator should be able to exert the
monopoly prices for a sufficient period of time so that he can recover his losses during
predation and also make a fair return on its investment in predation. Evidence of
recoupment was not a required element of predatory pricing in all courts before the
Brooke decision, which announced that the predator must be able to recoup losses
through supra-competitive prices for the case to be found predatory in nature. It was also
inferred that recoupment would not be possible if the markets are competitive, barriers to
entry are low, and the alleged predator either does not have the capacity or lacks the
resources to create capacity to capture the market share of an eliminated rival (Watson
1998).
There are two general ways that a plaintiff can attempt to prove recoupment. The
first path is to show that recoupment occurred through actual market data (supra-
competitive pricing for sufficient time to recoup). The second path would be trying to
prove that the predatory behavior was likely to bring tacit coordination and oligopoly
pricing. Tacit coordination would depend on similarities between the goals of the
competing firms, product variety and differentiation in the industry, and likelihood of
firms successfully signaling each other about price and output (1993).
75
Role of Defendant's Subjective Intent: Whether or not the intent of the defendant will
play a role in the case depends on the circuit. Predatory intent may be admissible to
reverse the assumption that above AVC is legal. Some courts have argued that proof of
intent leads to illegal scrutiny through corporate documents and hence is
counterproductive. Some have argued that intent is not important as long as the rival
could not be eliminated through predation. Intent to monopolize may also be derived
from attempts to create artificial barriers to entry, by entry deterring pricing, restrictive
marketing practices, acquisition of key inputs, or aggressive expansion of capacity
(McCareins 1996).
Role of Barriers to Entry: The Supreme Court view holds that high barriers to entry
enable the firm to recoup profits over a longer term. However, with low barriers,
predatory pricing may have limited effect on competition since as soon as the incumbent
increases prices new rivals will see the opportunity and enter.
Relevant Market: Last but not least, the relevant market refers to the product, service or
geographical areas that are involved in the case. Monopoly power of the predator applies
to the relevant market through its exclusion of rivals in the market and the subsequent
supra-competitive pricing. A product/service market should have its own elasticity of
demand, whereas geographic markets can be local, state, national, and even international.
“Evaluation of actual, probable, or presumed anticompetitive effects can be done only in
the context of a relevant market. Therefore, one of the first steps in an antitrust analysis is
determining the boundaries of the relevant market(s)…” (Enders 1986, p.24).
The determination of the relevant markets becomes harder when the multi-product
nature of the markets is taken into account. Most businesses have at least one product
76
line, and most large companies have multiple product lines. A static analysis of predatory
pricing may have indicated that it is irrational in a single market, however it can be
rational in a multiple market environment where a reputation effect of predation can deter
potential entrants in related markets (Trujillo 1994). From a legal standpoint, the relevant
market refers to the product, service, or geographical areas that are involved in the case.
Market Power: Monopoly power has been defined as “the power to control prices or
exclude competition” (1956, p.391; 1979, p.272). The existence of market power can be
simply inferred when a defendant has pre-dominant market share (1966, p.571).
The structure of a market may be derived from the study of the number and size
of firms, their cost and demand conditions, product differentiation, the nature of any entry
barriers, and degree of regulation. Industry Concentration Ratio (percentage of total sales
of the n (usually four) largest firms in an industry) and the Herfindahl-Hirschmann Index
((HHI) –sum of the squared market shares of all firms) are two commonly used market
structure indicators. However, the calculations for both measures need to be based on the
relevant market figures to be meaningful.
Application of performance based market power tests include practical or
conceptual difficulties, hence most studies use market structure tests. Previous studies
have generally shown a positive correlation between market concentration and
profitability (Shepherd 1970) (though the actual relationship may be more complex due
to omitted third variables (Bharadwaj and Varadarajan 2004)).
The U.S. Department of Justice's merger guidelines suggest that an HHI over
1,000 may raise antitrust concerns (1997a). According to the 1992 DOJ/FTC Horizontal
Merger Guidelines, markets with an HHI greater than 1800 are considered "highly
77
concentrated," and the DOJ/FTC will not approve mergers and acquisitions in highly
concentrated markets if there is even a slight increase in the HHI (1997a). Courts will not
predation credible unless the incumbent has market power.
3.6 The State of the Debate on Predatory Pricing
The purpose of the Antitrust Act and its application in the courts is a continuously
debated topic. As mentioned previously, there are two main schools of thought. The
Chicago School of thought based on the neo-classical theories argues that predatory
pricing is “rarely tried and even more rarely successful” (1986b). On the other side, the
emerging Post-Chicago School of thought argues that there is plenty of evidence that new
approaches such as strategic theory and game theoretic models provide that predatory
pricing can be rational and profitable for the predator (Bolton et al. 2000). The two
schools are compared next:
3.6.1 The Chicago School
In an early work that continues to influence the courts today, McGee (1958)
studied the 1911 Standard Oil Case, which was considered to be the classic case of
predation. He found no evidence in the case trial records that Standard Oil had indeed had
cut its prices below cost to drive out smaller competition and later intended to increase
prices. He argued that predatory pricing by Standard Oil would have been irrational
because of the relatively larger losses it would have had to suffer due to its higher market
share. He also argued that the prey would not be inclined to leave the market since it
knows that the predator cannot afford such large losses infinitely. Funding for the prey
78
would not be a problem either since capital markets would effectively step in as long as it
is an efficient producer. He claimed that the predator would not have gained anything
even if it could drive competition out of the marketplace because the prey, a purchaser of
prey’s assets or new competition could enter the market as soon as the predator increases
its prices.
With the lack of a rival theory, McGee’s (1958) analysis was considered to be the
only coherent economic theory of predatory pricing (Bolton et al. 2000). Other scholars
such as Koller (1971; 1978); Areeda and Turner (1975; 1996), Harold Demsetz (1973),
Demsetz and Weiss (1975); George Stigler (1987); and Wesley Liebeler (1986) expanded
upon this work and built the foundation of the Chicago School of thought. During the
nineties however, despite the late effort by John Lott (1999; 1996) the Chicago School
seems to have lost its dominance in academic journals.
2
Free enterprise institutes carry on the mission of the Chicago School. Institutes
such as the American Enterprise Institute, Heritage Foundation, and Cato Institute
(www.cato.org) promote free markets and limited government. They sponsor Chicago
School stream of research through their funding which comes from foundations,
corporations, and individuals. They maintain numerous web sites and publish journals,
magazines, and newsletters to keep the Chicago torch burning. Foer and Lande (1999)
reported that each of the above institutions spends around up to $30 million and argued
that their lobbying has resulted in considerable decline in Federal antitrust funding over
time.
The courts continue to empathize with the Chicago School and have been
2
I am joined by Bolton et al. (2000), and Sappington and Sidak (2000) in my critical view against
the validity of Lott’s findings.
79
exercising it as though there are no opposing theories. Judges who have openly sided
with the Chicago School include judges Bork (1978) and Easterbrook (1984; 1981).
There seems to be a prejudice that most cases are brought forward by the small
companies to create heavy litigation costs for large corporations. For example, Judge
Easterbrook estimated the average cost of a predation case for a major corporation to be
around $30M and even went as far to argue that the antitrust offense of predation should
be forgotten (Easterbrook 1981, p.337).
The general argument of the Chicago School has been summarized here based on
the work of DiLorenzo (1992). The counter argument of the emerging school of thought
is presented in the next section.
7.1.2 Chicago School Claim: Predatory pricing is irrational and very rarely (if ever)
occurs because:
1. Predatory practices are more costly for the large firm due to its larger market
share (it has to incur losses on a larger number of units) (McGee 1980).
2. It is not possible to continuously charge supra-competitive prices. Potential
entrants will be lured one after another. The incumbent cannot recoup losses
(Gomez et al. 1999).
3. Price wars are inherently uncertain, thus recoupment possibility cannot be
calculated accurately. A price war could also spread to surrounding markets,
making predation even more risky. McGee (1980) argued that even though
predatory pricing is usually irrational, it is generally rational for the victim to hold
on because predatory pricing strategies represent temporary cuts in prices.
Furthermore, Easterbrook (1981) argued that the victim has the same incentive as
80
the predator to outlast its rival and possibly collect eventual monopoly rents. In
expectation of supra-normal price levels and profits that will follow the price war,
capital markets can step in and help the prey. Thus, deep pockets argument will
not hold (Gattuso and Boudreaux 1999).
4. The prey firm can shut down and wait for supra-normal prices. In the meantime, if
they go out of business, someone else can take over.
5. There are opportunity costs associated with the funds allegedly used for predation
by the large firm.
6. Consumers could potentially stock up during predation; limiting quantity per
consumer also would not work as competition would step in and supply the
unserved demand.
7. Victims can arrange for long term contracts above predation prices with
customers if they also realize that monopoly prices will follow. Since the
customers will benefit from the prey’s continued existence as a supplier, they may
agree to such long term agreements to buy at a truly competitive price.
8. Anticipated monopoly profits have to be discounted to present value, diminishing
any value predation may have had. Indeed, the future recoupments must be
discounted by the probability that monopoly power will not be achieved, and then
discounted again to present value (Easterbrook 1981, p.272). Acquisition is a
much more profitable way of eliminating a competitor (Gomez et al. 1999).
According to the Chicago School, economic efficiency of the market should be
the goal of antitrust (Bork 1978). This interpretation evolves around maximizing
81
productive and allocative efficiency. The productive efficiency is simply defined as the
ratio of the outputs in relation to the inputs the companies employed. Allocative
efficiency refers to the general efficiency of the markets and involves how the limited
societal resources should be allocated across industries. Allocative efficiency in general is
referred to as the consumer welfare (Gundlach 1995).
The stance regarding merger and acquisitions is also a very important antitrust
issue. It is also related to predation since mergers and acquisitions typically lead to
dominant market power for the firms involved, the prerequisite for predatory pricing. The
Chicago School view of mergers can be summarized as follows: as long as benefits
(productive efficiency benefits to merging firms, and allocative efficiency benefits to
consumers) of a proposed merger is not lower than the potential loss to the competitors, it
would not be seen as a violation of antitrust. Similarly, the Chicago School claims that in
an industry with two levels where the production is monopolized and distribution is
competitive (i.e., gasoline), the monopolist cannot increase its profits by acquiring the
distributors. Increasing the retail markup will mean decreasing the producer markup by
the same amount. The monopolist cannot maximize its profits beyond the monopoly
profits (Posner 1976).
The intent and drastic price cuts that the populist era emphasized are not
important for Chicago School: “intent plays no useful role… Firms “intend” to do all the
business they can, to crush the rivals if they can…[A] desire to extinguish one’s rival’s is
entirely consistent with, [and] often is the motive behind, competition…[P]rice
reductions are carried out in a pursuit of sales, at others’ expense. Entrepreneurs who
work hardest to cut their prices will do the most damage to their rivals…” (1989,
82
pp.1401-1402). It follows that the intention to compete is not different than the intention
to exclude a rival.
Some Chicago School extremists take the position even further by claiming that
below cost pricing should also be legal. It has been argued that meeting price cuts,
discounting for introducing new products and excess capacity of perishable products are
already acceptable reasons for below cost pricing (Boudreaux and Kleit 1996). It was
proposed that all government monopolies including the postal service and public schools
be deregulated (Boudreaux and Kleit 1996b) and that all direct competitor suits be
prohibited (Boudreaux and Kleit 1996a).
The Federal Courts attempt to detect cases of predatory pricing through the
assumptions and the logic of Chicago School which is derived from the neo-classical
price theory view (Guiltinan and Gundlach 1996). Figure 3.4 illustrates the decision-
making criteria that the Federal courts use today.
83
Is the defendant sacrificing current profits?
1. Cost-based Tests
Are the Relevant Competitors Affected?
1. Exit?
2. Reduce Output?
3. Decide not to Enter?
Will There be Injury to Consumer Welfare?
1. Allocative Efficiency
2. Productive Efficiency
Is Recoupment Probable?
1. Market Power
2. Barriers to Entry
3. Industry Characteristics
Case Dismissal
by Summary Judgment
Verdict:
Defendant Prevails…
Verdict:
Plaintiff prevails…
Harm to consumer and
competitive process
Adopted from Guiltinan and Gundlach (1996, p.89)
chasm
Yes
No
No
No
Yes
Yes
Yes chasm
No
Figure 3.4: Decision-Making Framework in the U.S. Federal Courts
3.6.2 Post-Chicago School of Thought
Even though Koller (1971), Areeda and Turner (1975) and other earlier literature
provided seemingly counter evidence for a rationale of the existence of predatory pricing,
the claim for a myth of predatory pricing remained unjustified (Bolton et al. 2000). Koller
had reported that out of 23 cases he studied, actual predation was attempted in seven
(30%) and achieved in four cases (17%). Zerbe and Mumford (1996) studied the same
cases since 1940 and updated them until 1982. They detected predatory pricing in 27 out
of 40 (68%) of the cases.
Both of these studies probably under-reported the cases of predation because they
did not include settlements, predatory disciplining where no suit is filed, forced
acquisitions, and cases that were not filed because supporting theory was not yet
84
discovered or known (Bolton et al. 2000). Granitz and Klein (1996) conducted a re-
assessment of the Standard Oil case, and on contrary to McGee's (1958) work found
evidence that predation had occurred. Perhaps, actual predatory pricing cases are not as
rare as the courts have concluded (Adams and Brock 1996). “Exclusionary strategies are
frequent, not exceptional business practices”(Brodley and Hay 1981, p.1045). Bolton et
al. (2000) argued that there is now a consensus in modern economics that predatory
pricing can be a rational and successful business practice. They indeed made the claim
that no major article has had a counter statement during the last three decades. Many
weaknesses of the Areeda-Turner rule and the Chicago School arguments have been
reported (McCall 1987).
Many of the early critics of Areeda-Turner focused on its short run focus. For
example, Joskow and Klevorick (1979) argued that “...to dismiss entirely an assessment
of long-run effects, as for example Areeda and Turner seem to do, is to dismiss the
essence of the predatory pricing problem.” Williamson (1977) argued that temporary
price cuts have negligible benefits and resulted in long-term social welfare problems.
The solely cost based approach of Areeda-Turner was also criticized for being
impractical. Accounting and inventory systems vary; the distinction between costs and
investments is not clear cut; and it is hard to attribute costs to a specific product line or
product in the often seen case of multiple product lines. Areeda-Turner cost-based tests
are difficult to apply due to the multi-product nature of most businesses (Gomez et al.
1999). The emerging view’s counter arguments of the Chicago School are summarized
below, and additional points follow.
85
7.2.2 Post-Chicago School Response: Predatory pricing can be rational and may
frequently occur because:
1. Predatory practices are not necessarily costly for the large firm due to its larger
market share because predation does not occur in every segment of the market. It is
usually geographically or otherwise localized to a segment. Instead, the predator
may subsidize its local losses by its profits in other segments and wage war much
longer.
2. It may be possible to charge supra-competitive prices due to high barriers to entry,
reputation and other signaling effects. Potentially more efficient entrants see what
happened to the previous entrant, and may decide not to commit the high level of
resources needed to fight the incumbent. A valid purpose of predation may be to
develop a reputation as a tough competitor (Comanor and Frech 1993), (Kreps and
Wilson 1982). Bad lock-ins in network industries may result in loss of welfare.
3. Price wars often result in the destruction of the small competitors due to the deeper
pockets of the predators. Capital markets will not necessarily step in and help the
prey because of reputation effects. “Each time a start-up is driven out of a market,
the difficulty of obtaining funding ratches up a notch.” The fact that there is no
readily available financing for a start-up that has faced predatory practices is a direct
contradiction of Chicago theory. Lenders normally require substantial security, a
demonstrated ability to repay. If the loan is perceived to be risky, higher interest
rates will result. It is not likely for prudent lenders to give credit at all to a small
prey facing predation by a large and experienced firm (Atwood 1998).
86
4. It is not realistic to argue that the prey firm can shut down and wait for supra-normal
prices. Brand equity is hard to build up and customers value continuity of products
and services.
5. There are opportunity costs associated with the funds allegedly used for predation by
the large firm, however there are opportunity costs associated to the capital of
potentially more effective entrants as well. They will not enter the market if they
have other options where they are not likely to meet with predation. Moreover, the
opportunity cost argument presents a rationale for predatory pricing. Foregone
profits, that would have been earned if the monopolist had employed its assets in
alternative options rather than using them to predate in the subject market, also
represent costs (Ordover and Willig 1981). Comanor and Frech (1993) argued that
the predator incurs losses from an economic sense, but not necessarily from an
accounting sense. The loss is the lower profits than otherwise could have been
earned.
6. The argument that consumers could potentially stock up during predation, simply
does not hold for network industries (e.g., transportation, software,
telecommunications) and services and is not practical for many other industries.
Customers would not engage in long-term contracts above predation prices unless
they realize the full implications of the ongoing predatory pricing. They normally
would try to maximize their short run returns as long as they believe they have
alternate suppliers. Moreover, long term contract argument does not hold for
consumer markets (e.g., airlines).
87
7. Anticipated monopoly profits may have to be discounted, however the strategic
long-term gains also need to be built into the equation. What would the incumbent’s
losses be if it had remained inactive? How many potential entries were discouraged
by the predatory behavior? These are hard to estimate but the long term impact on
business is not. Recouping the investment may be much easier than thought. Case
studies have suggested that predators have occasionally succeeded in recouping their
losses (Adams and Brock 1996).
8. Contrary to what Easterbrook (1981) argued, litigation can be much more costly for
smaller companies which throughout the case have to incur legal fees they cannot
afford. Typically, large firms departmentalize the case as they have the personnel
and the resources to do so; however, the prey’s management can easily get drawn
into this side effort and drown in the increasing details of the case.
The arguments above and the emerging Post-Chicago School of thought could be
observed in the works of Klevorick (1993), Ordover and Saloner (1989), Craswell and
Ratrik (1985). The Post-Chicago School raised numerous issues that the Chicago School
theory had not captured before (1998a). Guiltinan and Gundlach (1996) criticized the old
school of thought because it relies on the assumptions of neo-classical theory such as
singular motivation of managers for profit maximization, and firms’ possession of
complete information.
Sole Profit Maximization: There may be practical objectives other than sole profit
maximization. Managers may settle for satisfactory rather than optimal levels (Baumol
1967; Simon 1979). Predatory pricing generally leads to a drastic increase in market
88
share, which if aligned to performance evaluations, may be a motivational factor for
managers. Maximization of career opportunities has also been offered as an alternative
for sole profit maximization (Stelzer 1987). Urbany and Peter (1994) have shown
empirical evidence for a preference for volume rather than profit orientation. Their
experiment indicated that manufacturing firms held a long-term perspective for customer
acquisition and in estimating their worth.
Rational Decision Making: Decision making under uncertainty is not rational. Guiltinan
(1996) argued that managers do not estimate probabilities accurately when they are
dealing with risky situations. Simon’s (1957) highly regarded concept of bounded
rationality implies that: a) buyers/managers are not aware of their comprehensive set of
alternatives, b) they do not know the outcome of taking a specific alternative with
certainty, and c) they do not have the mental capacity to rationally process all perceived
alternatives. Individuals tend to develop noncompensatory preferences (e.g., elimination
by aspects) when faced with complex decision scenarios (Bettman et al. 1998).
Kahneman and Tversky (1979) have long demonstrated that individuals do not
necessarily maximize their economic utility in the way predicted and assumed by neo-
classical economics. Their Nobel prize winning prospect theory essentially claims that
individuals have differing risk preferences and display risk-averse characteristics for
gains and risk seeking characteristics for losses (i.e., a sure $50 bonus would be preferred
over a coin toss for $100 or nothing, whereas a coin flip for a $40 parking ticket or
nothing would be preferred over a sure $20 parking ticket). That is, the utility curve for
gains is concave whereas the utility curve for losses is convex and also with a steeper
slope. Thus, managers are risk-averse for gains but risk-affinitive for losses.
89
Post-Chicago Proposals:
Meeks (1998) criticized the static nature of the traditional analysis and argued that
it may lead to unsatisfactory results if cost is used as the main criterion. He argued that
for anti-competitive conduct to occur, the price need not even be below cost, as long as it
was low enough to deter entry, and supra-competitive prices would occur once the
potential competition is eliminated. He proposed a non-cost based approach for detecting
predatory pricing in transition markets (e.g., telecommunications). It was important to
consider if the market was recently deregulated or associated with one that is regulated.
He suggested that if the lower price was offered to a large segment of the market, it was
probably not a static move aimed at preventing new competition. The predating firm had
to possess market power (at least 50%). Strategic harm to potential competition had to be
the most likely explanation for the predatory action and the barriers to entry had to be
high (Meeks 1998).
Bolton et al. (2000) in their award winning effort argued that modern economics
principles should be employed to submit proof of predatory pricing. In particular, they
suggested that the pro-competitive dynamic gains such as reputation effects should be
considered, and short and long run incremental costs should be employed for proof of
below-cost pricing. Essentially, their proposal required the following elements to be
incorporated:
1. Facilitating market structure: “Short-run pricing power” had to be present in the
market. This would be typically observed by the existence of one or more
dominant firms and high entry and re-entry barriers.
90
2. Scheme of predation and supporting evidence: A persuasive evidence of
recoupment would have to be presented at the absence of a plausible scheme of
predatory pricing.
3. Probable recoupment: Probable recoupment as opposed to actual should be
sufficient. Moreover, intangible benefits from injury to competition and
exclusionary effects would be acceptable.
4. Price below cost: They proposed that AVC should be substituted with Average
Avoidable Cost
3
, and ATC should be substituted with long run average
incremental cost (LRAIC).
5. Absence of an efficiencies or business defense: A plausible efficiencies gain, no
less restrictive alternative, and efficiency-enhancing recoupment would have to be
demonstrated.
This proposal was a novel attempt to incorporate the dynamic strategic
perspective into the current antitrust policy. Guiltinan and Gundlach (1996) argued that a
dynamic strategic approach underlines the insufficiency of the current law to cope with
aggressive and predatory pricing.
In summary, the Post-Chicago School contends that it may be rational and
plausible for firms to employ predatory pricing (Hazlett 1995). It examines the relevant
markets with a strategic perspective, considers impact of opportunity costs, and imperfect
information (Bolton Brodley, and Riordan 2000; Brodley and Hay 1981; see Bloom and
3
Average Avoidable cost: Average per unit cost that predator would have avoided during the
period of below cost pricing had it not produced the predatory increment of sales.
91
Gundlach (2001a) for a discussion of differences between Chicago and Post-Chicago
Schools of thought). Simply put, the Post-Chicago view is that markets are generally not
perfect, and they are not necessarily self-correcting (Gundlach et al. 2002). Some of the
prominent names that have supported this stream include Lawrence Sullivan, Paul
Joskow, Alvine Klevorick, Januzs Ordover, Robert Willig and Joseph Stiglitz (Elhauge
2003).
Marketing perspective is well aligned with the Post-Chicago School of thought in
its basic position. Additional assumptions of marketing to that of the Post-Chicago
School are alternatives for assessing consumer welfare (e.g., variety, innovation,
satisfaction), and the possibility of irrational decision making and non-profit maximizing
goals (Gundlach 2001). Marketing could well assume a complimentary role due to its
cross-disciplinary nature which has not yet been realized (Gundlach et al. 2002).
3.7 Synthesis and a Marketing Perspective
3.7.1 The Need for a Marketing Perspective:
The unique position of marketing to offer public policy insights has been
observed by several scholars working on the subject (Grewal and Compeau 1999;
Guiltinan and Gundlach 1996b). It was suggested that a marketing focus on predatory
pricing is overdue (Grewal and Compeau 1999). The many aspects of consumer welfare,
which is of key concern to public policy, can be captured with the comprehensive
marketing measurement and modeling tools that enable the study of the benefits of
quality, service, variety, and innovation to consumers (Guiltinan and Gundlach 1996).
The marketing discipline has the potential to further the understanding needed for the
92
development of a more suitable antitrust policy (Gundlach 1995). This potential has not
been utilized by the courts so far. A review of the marketing mix variables could reveal a
better understanding of predation than a focus on pricing alone.
The issue facing legislators of predation is indeed similar to one researchers face
when designing a scientific study. Defendant’s conduct can be presumed legal (i.e., non-
predatory) for the null hypothesis (stemming from the notion of being innocent until
proven guilty). Alternate hypothesis would be that the defendant’s conduct is illegal
(predatory). The risk that an incorrect conclusion may be reached always exists. A
balance needs to be sought through the manipulation of Type I (rejecting the null
hypothesis when it is true), and Type II (failing to reject when null hypothesis is false)
errors, and the respective power (probability of successfully rejecting the null hypothesis
when it should be rejected) of the test. Legislators set the evaluation criteria (e.g., Areeda
–Turner AVC rule), imposing rigid parameters that apply to everyone in the name of
being just. The set criteria, hence the parameters can drastically shift over time as
differing success rates of plaintiffs for similar cases show. In some cases, however, the
judges may have set unfortunate examples when they interpreted these criteria too
strictly. The influence of these (Supreme Court) decisions may have resulted in the self-
fulfilling prophecy that predatory pricing is a “myth.” Yet, the courts seem to be content
with this situation due to their comforting assumption that predatory pricing “rarely”
succeeds and that consumers benefit from failed predation attempts. A minimal standard
(i.e., higher alpha level) would have been costly for the competition, because a number of
wrong inferences (i.e., rejections through Type I error) would have a negative impact on
desirable price cutting behavior (1993, p.226).
93
3.7.2 A Focus on Deregulated Markets:
It was argued that price discrimination has to accompany strategic predatory
pricing since predation is usually targeted by geography or other segmentation (Meeks
1998). Dempsey (1989) discussed how deregulation of an industry facilitates
discrimination. Brennan (1995) studied deregulation in the telecommunications industry
and detected potential for predatory pricing. The same hazard was detected for electric
power, natural gas (Meeks 1998) and airline industries (Kahn 1987). Meeks (1998)
argued that recently deregulated industries are much more likely to have examples of
predation than those that are not. This interesting notion is reflected in the following
framework (Figure 3.5).
Figure 3.5: Framework for Deregulation and Predatory Pricing Relationship
Naturally, industry characteristics influence all factors in this framework
including deregulation. The number and size distribution of firms, cost and demand
conditions, and barriers to entry impact firm strategies. However, it is also true that
deregulation often transforms a monopolistic industry to an oligopoly with a number of
Deregulation Price
Discrimination
Predatory
Pricing
New / Potential
Inter Entry
Non-price Predation
94
firms with high market power. Since no precedent has been set, it can be easy to employ
price discrimination. It is particularly easy to discriminate in previously regulated service
industries because there are few arbitrage opportunities for the consumers (e.g., leisure
traveler cannot transfer plane pre-purchased plane tickets to businessmen on the run).
Robinson-Patman Act limits price discrimination for commodities and does not generally
apply to service or lease industries (1936). This enables such industries to set supra-
competitive prices. Kotler and Armstrong (1991, p.343-44) described an airline industry
that has taken full advantage from the possibilities of price discrimination:
The passengers on a plane bound from Raleigh to Los Angeles may pay as
many as ten different round-trip fares for the same flight—first class; first
class-night; first class-night, child; first class-youth; coach; coach-night;
coach-night, child; Super-Saver, nonrefundable fare; Super-Saver, 25
percent cancellation penalty; and military personnel. These fares vary
from $238 to $1512!
The existence of discrimination could provide rationale for predation because
discounters with their low cost structure can be very damaging to the incumbent’s
established price discrimination strategy based on market segmentation. For example,
revenue management is essential to the success of the hospitality and airline industries,
and is closely related to marketing through price. Major carriers’ predatory behavior
against the low cost carriers has been observed in the airline industry (1998c; 1999b;
1999c; 1999d). Incumbent firms may also employ non-predatory tactics to prevent new
entry to their bases of market power.
95
3.7.3 Epilogue
By all accounts, the 1975 Harvard Law Review article “Predatory Pricing and
Related Practices under Section 2 of the Sherman Act” by Phillip Areeda and Donald F.
Turner was the seminal article in predatory pricing. All U.S. courts have been influenced
by this article in one way or another. It has been noted that the widespread acceptance of
Areeda-Turner and the conservatism in the U.S. courts, especially since the eighties, may
be explained by judiciary appointments by conservative Presidents, the intellectual
influence of the Chicago School and the extensive public relations campaigns to
communicate its aspects to the judges, and the contrasting lack of an organized antitrust
effort (Foer and Lande 1999). Following Areeda and Turner (1975), that predatory
pricing is rare and AVC is the measure to test the claims of predatory pricing have been
generally presumed by the courts with some exceptions mostly on the State level.
However, Areeda and Turner and the Chicago School have had their share of criticisms:
Scherer has demonstrated that the Areeda-Turner rule would not promote
long-run economic welfare, would not ensure an efficient allocation of
resources, and would encourage firms to maintain excess capacity. Greer
has shown that the Areeda-Turner rule relying on either an average
variable or marginal cost floor would be overly lenient in that it would
allow the destruction of equally (or more) efficient rivals. Dirlam has also
rejected cost-based rules, noting that they are too rigid and would require
difficult and ambiguous short-run cost measurements. Finally,
Beckenstein and Gabel have argued that succinct per se rules are unable
to deal with some anticompetitive practices and other subtle forms of
business behavior like predatory investment, economies of scope, vertical
integration, and experience curve learning (McCall 1987).
The criticisms of the Post-Chicago School are summarized in Table 3.2.
96
Table 3.2: Criticisms regarding the Areeda-Turner rule and the Chicago School
Theoretical
• Uses unrealistic assumptions of neo-classical price theory (e.g., Lande 1993;
Guiltinan and Gundlach 1996)
• Describes the manufacturing era of the past versus the current knowledge/network
society (e.g., Bolton, Brodley, and Riordan 2000)
• Previous empirical findings were flawed (e.g., Granitz and Klein 1996; Zerbe and
Mumford 1996).
• Predatory pricing can be rational (e.g., Adams and Brock 1996; Sullivan and Grimes
(2000)
• Lacks a strategic perspective (e.g., Williamson 1977; Joskow and Klevorick 1979)
• Ignores reputation, signaling effects and information asymmetries (e.g., Comanor and
Frech 1993; Jung, Kagel, and Levin 1994; Milgrom and Roberts 1982)
• Ignores the nature of network/service industries (e.g., European Commission 1998)
• Ignores opportunity costs of actions (e.g., Ordover and Willig 1981)
• May exclude (more) efficient rivals (e.g., Greer 1979)
• Promotes inefficient allocation of resources and excess capacity (e.g., Sherer 1976)
Legal
• Cost rules are hard to calculate, impractical (e.g., Sievers and Albery 1991)
• Short-run cost focus is irrelevant (e.g., Dirlam 1981; Meeks 1998)
• Per se rule is inapplicable to all cases (e.g., Beckenstein and Gabel 1986)
• Litigation can be costly for the prey (e.g., Atwood 1998)
Managerial • Promotes selective and geographic predation (e.g., Allvine 1996)
• Narrow view of market power (e.g., Trujillo 1994)
• Ignores deep pockets of predators (e.g., Atwood 1998)
Buyer • Does not promote long run economic welfare (e.g., Scherer 1976)
Despite these criticisms, the Chicago School has keenly used Areeda-Turner as
the basis for their static, neo-classical price theory arguments for more than two decades.
However, not only the contexts of predation and the nature of the problem have changed
over the years, but also the opposing Post-Chicago views have been growing stronger
(1998b). Hence, it is worthwhile to revisit the Areeda and Turner (1975) article and
comment on its key limitations in today’s economy.
The context in which Areeda and Turner published their article was radically
different than that of today. In the Areeda and Turner world, the economy was still driven
by production of oil, steel, and commodities. Sheer output was what mattered most. In the
manufacturing era, it was presumed that the bigger was better, higher capacity would lead
to better economies of scale and a better cost structure. Being large enabled them to gain
97
dominant market share and hence market power. Areeda and Turner projected of a
dominantly manufacturing world where there is either perfect competition (i.e., many
firms take going price) or monopoly (i.e., one firm sets the price). “The view that
predatory pricing is rare or even fanciful is based on theoretical constructs that ignore the
realities of markets in which oligopolistic structure, incomplete information and strategic
behavior are commonplace” (Sullivan and Grimes 2000, p. 145).
The “real world” is almost always somewhere in between a monopoly and perfect
competition, with differentiated products of oligopolies instead of commodity outputs
and identical firms. In the real world, the firms do not compete only on quantity and
price. It is a service-based economy that is pre-dominant in the U.S. society today. The
small firm can be more efficient and welfare enhancing than the corporate giant. Small
service companies from catering to transportation, and entertainment to management
consulting can be as or more efficient than the big players. Southwest in the airline
industry is one clear example of an efficient firm. The nature of the airline industry
enables clear illustrations of the new issues facing the policy makers.
Areeda and Turner perceive two pre-requisites for predatory pricing: a greater
financial staying power by the predator and a very substantial prospect of recoupment.
The use of Chicago School doctrine, made it really easy for Judge Marten to decide that
American had indeed not violated the law and to dismiss the case by summary judgment.
American had not priced below AVC, and the recoupment of losses incurred during the
alleged predation was not probable anyway (memorandum and order) (1999c). Yet,
looking at the same facts, DOJ and the Post-Chicago School concur otherwise. Modern
thinking in antitrust, dubbed here as the Post-Chicago School deals with reputation
98
effects, game theory, conspiracies to raise rival’s costs or reduce rival’s revenues and
even scanner data input analyses where applicable. The Post Chicago School studies the
relevant markets with a strategic perspective and considers the firm’s opportunity costs of
its actions.
The current antitrust enforcement does not consider the differences in competitive
reaction to entry by intra- and inter-type companies. This purposeful ignorance extends
from the Pre-Areeda-Turner enforcement era heavily criticized by the Chicago School.
The Robinson Patman Act (1936) was passed with the general intent of protecting the
mom-and-pop stores from the chain-store revolution. Many small stores won cases of
predatory pricing against the large (and generally more efficient) chains until Areeda-
Turner set the tone for the rest of the century, and the objective of antitrust became
guarding the competitive process and not the (small) competitors. For example, entrant
Wall-Mart grew as a discounter by undercutting the high-service/high cost department
stores and established itself as one of the largest retailers in the world. The incumbent
local mom-and-pop stores did not have much other than their older but loyal clientele to
defend themselves against the chains. The stores that were less efficient simply faded
away.
Areeda and Turner (1975) utilized a number of purely theoretical examples to
illustrate their points, and they readily admitted that some of these scenarios posed threat
to long-term competition. Areeda and Turner did not believe that a long-term dynamic
test of predation could be practically developed. They reasoned that their short-term static
cost focus would be practical. However, the very notion of recoupment represents a
strategic assessment of a long-term return (supra-normal profits) on a short-term
99
investment (predatory pricing). Thus, the nature of the problem imposes the need to use a
strategic-dynamic test (rather than a short-run static rule) to assess predatory pricing.
Today, technology enables market segmentation at the micro-level. Companies
offer product/service lines with different levels of utility at varying prices (e.g., premium,
standard, discount). Specialization and focusing on different market segments or niches is
also common. Customer Relationship Management and other sophisticated marketing
tools permit customizing the product/service offer even to the individual consumer’s
needs. However, this targeting capability becomes destructive if predatory pricing is
pinpointed at a local rival. For example, American Airlines did not cut prices on all or
even most of its routes on its DFW hub when it faced entry by discounters. It cut its
prices only on those routes in that it competed head to head with them. This geographic
segmentation enabled them to subsidize losses incurred during the time of predation on
targeted markets. The revenue and losses at stake were insignificant for American but
vital for the survival of Vanguard (and SunJet and Western Pacific which are now out of
business). As the American Airlines case justified “devices other than a general price-cut
may, however, be the subject of suits for predation” (Areeda and Turner 1975). Such
selective or geographic price cuts provide the major airline next to infinite staying power
through cross-subsidization.
Moreover, American increased capacity on these routes by shifting its aircraft
from more profitable routes. Any business expanding capacity entails opportunity costs.
“[A]ny cost calculation that totally ignores the opportunity cost component is likely to be
illegitimate…it is essential to include all opportunity costs of ownership inputs…”
(Baumol 1996). Foregone profits that would have been earned if the monopolist had
100
employed its assets in alternative options rather than using them for predation represent
costs (Ordover and Willig 1981). Ironically, Areeda and Turner rule does not take the
opportunity costs into account. Predatory pricing analyses appear to be plagued by cost
issues.
Areeda and Turner have advocated the per se AVC rule, however, in their seminal
work they also asserted that “virtually all costs are variable when a firm, operating at
capacity, plans to double its output by constructing new plants and purchasing new
equipment” (1975, p.701). Deep discounts coupled with drastic volume increase can
indeed be the typical response to market entry depending on industry characteristics (e.g.,
air and truck transportation). Thus, the use of an ATC test (i.e., a less rigorous test than
that of AVC), may be practically justified under special circumstances.
Price discrimination today is a typical practice in service industries and is not
regulated by the Robinson-Patman Act. It has been shown that economies of scale can
motivate a price-discriminating monopolist to engage in unprofitable conduct and sell
even when the average cost is higher than the price. The incurred losses can be
effectively subsidized with profits from other markets. This can have harmful effects on
social welfare (Park 2000).
Areeda and Turner (1975) argue that “a demonstrated willingness to indulge in
predatory pricing might itself deter some smaller potential entrants… Repeated predation
in the same market, moreover, is not only costly but is likely to be easily detectable and
thus the occasion for severe antitrust sanctions.” Indeed, this argument is the essence of
the theory of reputation effects. Areeda-Turner is practically the standard rule in courts
today. Yet, it does not equip us to detect predatory pricing in many cases in a service-
101
based economy and makes it almost impossible to detect it in network industries (Carney
2001). Even the most deliberate cases of repeated predation may not be detected today
because of over-reliance on the Areeda-Turner rule.
According to the Federal courts, (predatory) intent is at best a secondary criterion
to consider (A.A. Poultry Farms, Inc. v. Rose Acre Farms 1989). This may have to do
with the fact that both the Chicago School and the Post-Chicago School agree that the
firm’s objective is profit maximization. Interestingly, a decision-making perspective
adopted by marketing, and one that goes beyond both schools is an understanding that
alternative and multiple managerial objectives such as survival, satisficing (e.g., simply
meeting the Wall Street numbers), and enhancing sales, personal welfare (i.e., self-
compensation, career advancement), reputation effects, and social welfare do exist
(Gundlach 2001). Sales management researchers have long studied the optimal employee
compensation mix for different business objectives such as building, holding, harvesting,
or divesting market share (Strahle and Spiro 1986). Similarly, top executive
compensation seems to correlate positively with sales but not necessarily with
shareholder return (McKnight and Tomkins 2004). The use of a sales/market share
growth objective typically enhances not only personal welfare but also corporate
reputation effects (i.e., warranted aggressive/predatory response by incumbent).
Therefore, the courts should recognize that alternative objectives can be legitimate goals.
For example, maximizing social welfare can be prevalent in the case of non-profit
organizations. Often an overlooked research dimension, the nature of competition
between organizations with contrasting objectives presents challenging issues for policy
makers. Manufacturers may aggressively cut prices (i.e., market share growth objective)
102
to secure contracts in their main business market (e.g., copiers/high speed printers) to
make premiums on complementary products (e.g., ink cartridges/services). Not-for-profit
hospitals have been accused of abusing market power in concentrated markets (Simpson
and Shin 1998). Competitive bidding/pricing by public universities for government
contracts may also be considered as unfair competition by the private industry
members/universities. This interesting phenomenon underlines the necessity of
undertaking a broader examination of the consumer welfare construct which is discussed
at a later section.
Moreover, the strategic marketing orientation of the aggressor business should be
examined in the courts. The orientation of the business can directly impact its
profitability and selection of performance objectives (Gatignon and Xuereb 1997). For
example, competitor-oriented businesses would likely retaliate more aggressively than
customer-oriented businesses in the face of competitive entry (Narver and Slater 1990).
Predatory sacrifice perceived by managers in competitor-oriented businesses is likely to
be much smaller than that perceived by their customer-oriented counterparts. The notion
of perception of sacrifice requires a discussion of rationality and risk tolerance as well.
The state of marketing and consumer research has progressed beyond the basic
risk-averse/prone notion and explored the role of multiple and conflicting personas and
non-financial (e.g., social) risks and moderating roles of risk perceptions on attitude
formation and decision making (Campbell and Goodstein 2001). Choices can also be
heavily influenced by the framing/perception of the proposition: a teenager might drive
through town to save $5 on a single DVD that would have cost $20 but would not do the
same to save $5 on a pair of athletic shoes that cost $150 (Thaler 1985). Similarly, a
103
purchasing agent/jobber may fly across the world to resolve a $50K dispute on an office-
supply deal worth $1 million, but may be reluctant to do so to save $50K in a $15 million
worth of automation equipment. Admittedly, the reluctance of the agent/jobber in the
latter example can also be explained by other factors (e.g., complexity of the deal,
purchase category etc.) but the point is that there are implications not only for consumer
and small business contexts but also for million dollar decisions. Organizational buying
committees consist of individuals with different backgrounds, departments, levels of
expertise, interest, motivations, power, and so on. The study of the consistencies in
irrational buyer behavior is currently a hot topic for behavioral economists,
psychologists, and marketers. It would be fruitful to bring new decision-making insights
to the courts (Korobkin and Ulen 2000). It was shown that decision-makers may “over-
compete” on price to maximize the profit difference and perform significantly inferior
than rational price models that maximize the profits (Griffith and Rust 1997). Irrational
competitive responses such as incumbents’ herding to new markets have also been
observed (Debruyne and Reibstein 2005). Effects such as information availability (also
information primacy and recency), biases in representativeness, categorization, and
optimistic overconfidence can change inferences drastically (Tor 2002). As such,
temporal, internal and external strategic reference points (historical precedents/cognitive
constructs) complement prospect theory to explain managerial decision-making and
predict competitive reactions (Shoham and Fiegenbaum 1999). Accordingly, managers of
underperforming businesses would be expected to be more risk-prone (increasing
likelihood of predatory conduct) under pressure than those of businesses that are doing
104
well. Therefore, predatory pricing (e.g., dumping by distressed steel producers in Russia
and Brazil) may be attributed to high debt and deteriorating performance.
3.7.4 A Reaction Framework for Competitive Price Reductions
In Figure 3.6, I integrate the learning from the literature review and present a
reaction framework for competitive price reductions from a cost-based perspective. The
framework illustrates how the current law might be interpreted by executives, and has
implications for business marketers, managers, and policy makers.
Resource base obviously plays a role but I do not want to convey the message that
the business with the deeper pockets has the final word in predatory pricing engagements.
Many small suppliers/manufacturers are able to defend their markets effectively. An
important reason is that marketers typically develop ongoing relationships with their
customers. Obviously, developments in the CRM arena have enhanced the effectiveness
of this key phenomenon. The stronger the relationship, the lower is the price elasticity for
the product, hence the lower the threat from predatory pricing. A very strong form of pre-
emptive market defense against predatory pricing occurs when these relationships are
converted into binding commitments in the form of legal contracts (i.e., business cannot
lose the buyer for a pre-specified time period regardless of predator’s pricing).
Moreover, many bids are not granted to the lowest cost provider but to the value
package solution provider. Therefore, bundling becomes a strategic tool and is especially
effective when the business has differentiated itself. However, careful consideration is
advised to marketers when employing strategic bundling since the same tool can be used
for aggression. For example, 3M recently got in trouble for bundling rebates for its office
105
supplies anti-competitively (2003b).
In essence, the Post-Chicago School does not yearn for a return to the Pre-Areeda-
Turner era, it rather incorporates the learning from deregulation, modern strategic
thinking, advanced technology, the new economy and globalization. It concludes that
Areeda –Turner is inadequate to cope with today’s dynamic and complex environment.
It is my contention that the nature of the network industries is such that, when
considered in an integrated paradigm, the use of pricing strategies for pre- and post-entry
defense is not only feasible but also rational. I discuss the nature of the network
industries, develop a price competition framework for network industries, and present my
hypotheses next.
Yes
No
No Yes
Yes No
Yes
No
Price above ATC
Price below ATC
above AVC
No Yes
Price
below AVC
Figure 3.6: Incumbent’s Reaction Framework for Competitive Price Reductions
Read:
The Case of Accommodation: The incumbent businesses tend not to compete on price when challenged by
a company of the same type. Rather, they tend to emphasize quality, design and differentiate their
products/services. Alternative reactions include temporary price cuts to distract buyer attention on
competitor or try to cross-sell existing customers.
The Cases of React and Defend and Predation: These scenarios typically take place between a major
company and a discounter. Discounter cuts the price to a level that is disturbing for the incumbent. If the
incumbent chooses to accommodate, the discounter can establish a foothold of the market which it can use
to drive the prices further down. The incumbents typically choose the other route of aggressive
competition.
Has the entrant
cut the price?
Accommodate
(Intra-type competition)
-Non-price competition
-Differentiate product
-Run temporary price cuts
-Cross sell existing
customers
React and
Defend/Counter-attack
(Inter-type competition)
-Intensive competition
-Non-price predation
-Match price
- Selective discounting
-Increase capacity
How much has the
price been cut?
Recoup losses and
resume pre-entry
pricing strategy
Did the rival fold?
-Increase price?
-Exit?
-Reduce activity to non-
significant levels?
Retreat and
Restructure
-Restructure and
decrease AVC/ATC
-Cut product variety
and focus on
competency
-Acknowledge
defeat and accept
new market share
Predate
-Predatory pricing
-Non-price predation
-Flood the market with
capacity
-Cannibalize rival at
all fronts
-Subsidize losses
-Sue
Maintain price;
monitor (potential)
competitors
Will the price cut
considerably deter the
bottom-line?
Does the entrant have
deep pockets?
Is the price cut
introductory (non-
permanent)?
106
CHAPTER 4
NETWORK PRICE COMPETITION FRAMEWORK
The Network Pricing Strategy framework presented is in Figure 4.1. The
framework has three features that should be noted. First, it represents a dynamic process a
la Resource Advantage Theory. That is, two-way knowledge flows (competitive signals)
and competitive responses are embedded in the framework although not explicitly drawn.
It follows the logic from empirical industrial organization literature and the PIMS
paradigm in that the structure of the market, the competitive position, and competitive
strategies determine a business unit’s performance (Bharadwaj and Varadarajan 2004).
Incumbents perceive market entry to strategic markets with discounts as acts of
aggression and will consider retaliation. Similarly, the entrant choice for market entry
depends on the choice of pricing strategy of the incumbent and strategic assessment of
the market. The state equilibrium of the model relies on the incumbent’s and the potential
entrant’s behavior. For example, the optimal pricing strategy for the incumbent may
change even without new entry, it is only necessary to increase the incremental potential
likelihood of entry (i.e., aggregate strategic assessment) sufficiently. Equilibrium points
of dynamic models have been shown to be more realistic than those of static models that
assume perfect information (Coughlan and Mantrala 1992).
Second, the framework is focused on a special but important type of market entry,
inter-type entry (see section 2.4.2 for a discussion of inter-type competition). Inter-type
competition is based on price, whereas firms with similar cost structures compete on non-
price factors such as promotion and service (Allvine 1999; Palamountain 1955). The
107
108
Department of Transportation and independent based analyses have observed this
phenomenon (Oster and Strong 2001; DOT reports 1997-2000). “Entry by major carrier
on a point-to-point basis into another carrier’s hub has become very much the exception”
(Nannes 1999, p.4). The rules of competitive conduct between major (i.e., intra-type)
firms has been the focus of research in marketing strategy literature. It has been laid out
that conduct in these markets can be explained by leader-follower pricing system (Roy et
al. 1994).
The framework is domain specific and focuses on price competition among firms
with inter-type cost structures. In particular, it is assumed that the incumbents are major
firms with deep pockets and the new entrants are discounters (that are more efficient than
and have cost advantages over the incumbents but have shallow pockets). Therefore,
intra-type competitors and inter-type competitors form two strategic groups of
competition (Hunt 1972). There are “mobility barriers” that prevent changing group
membership as suggested by the Strategic Groups literature (Bharadwaj and Varadarajan
2004).
Third, the model is designed for network industries (or industries with network
characteristics –discussed next) which are typically characterized by high barriers to
entry. Generic (industrial organization school) and sustainable capabilities (R-A theory)
are also incorporated in the form of market and firm specific barriers. I posit that
developing a framework by integrating these theories reveal a scenario in which
predatory strategies can be rational.
Three main factors, market power of the incumbent, barriers to entry and strategic
assessment of the incumbent, act as antecedents for choice of the pricing strategy of the
109
incumbent. This is consistent with the classical industrial organization literature. It is also
suggested that firm specific factors should be more influential than market and industry
specific factors, which is consistent with the Resource Advantage theory.
The price levels in markets and strategic assessment of the incumbents lead to the
decision on whether or not to enter any given market. This is consistent with classical
economics and signaling theory literature.
Incumbents are expected to react sharply in markets that they consider to be
strategic with the intention to drive entrants out of their markets. This is consistent with
findings from industrial organization and strategy literature. If the inter-entrant is driven
out of the market as a result of retaliation and or lower strategic prospects, it is expected
that consumer welfare will be negatively affected. This is consistent with views from
marketing and Post-Chicago school of thought (see Figure 4.1).
Next, I discuss the network industries and then develop my hypotheses following
the five phases in the model: pre-emptive defense, entry decision, post-entry defense, exit
decision, and policy consequences.
Pricing
Strategy:
•Supra-competitive
•Competitive
•Limit
Incumbent’s
Response:
•Predate
•React
•Accommodate
Consumer
Welfare
BEFORE DURING AFTER
NETWORK PRICE COMPETITION FRAMEWORK
Incumbent’s
Market
Power
Incumbent’s
Strategic Assessment (t
0
)
-
Barriers to Entry
-
Inter-
Entry
Inter-
Exit
Entrant’s
Strategic Assessment (t
exit
)
-
Potential Entrants’
Strategic Assessment (t
0
)
Figure 4.1: Pre-entry and Post Entry Network Price Competition Framework
1
1
0
4.1 Network Industries
Network industries have a unique and dynamic nature of competition. There are
four main characteristics that differentiate network industries from others:
complementarity, compatibility, and standards; consumption externalities; switching
costs and lock-in; and significant economies of scale in production (Shy 2001).
Complementary products (e.g., PC and mouse) require compatibility for industry-wide
appeal and this leads to the development and adoption of standards. This notion is
especially emphasized by firms in information markets. Consumers’ perceived value of a
network increases as others use or adopt complementary or compatible products or
services which leads to positive consumption externalities (Katz and Shapiro 1985;
Lemley and McGowan 1998). High switching costs in three categories (e.g., procedural
(economic risk, evaluation costs, learning costs, and set-up costs), financial (benefit loss
costs, monetary loss costs), and relational (personal relationship loss costs, brand
relationship loss costs)) cause many consumers to continue with their existing providers
(Burnham et al. 2003). Network industries are also characterized by high start-up (fixed)
costs, and low unit production (marginal) costs (Watson 1998). Heil and Robertson
(1991) argued that incumbents’ reaction propensity increases with high fixed costs and
economies of scale. Network learning dynamics (within firm, between firm and end-user
based dynamics), surveillance, and shared resources also enhance the capabilities of
network firms (Dickson et al. 2001). Users in network markets may derive benefits from
the user network (e.g., e-mail), the complements network (i.e., Windows-based software),
and the producer network (level of competition) (Frels et al. 2003).
111
112
Industries such as airlines, telecommunications, and software have been at the core
of the growth of this society. Network industries are often associated with capital-
intensive industries serving mass markets. Historically, these industries have been heavily
regulated due to the neo-classical economics argument that the social welfare is better-off
when these industries are preserved as natural monopolies. However, the positive
consumer experience after deregulation of several of them (e.g., telecommunications) has
demonstrated otherwise. Other industries that possess network industry characteristics
include: broadcasting, cable television, electricity, water, pipelines, sewage systems, oil
pipelines, natural gas pipelines, road and highway systems, bus transport, truck transport,
inland water transport, ocean shipping, postal service, package delivery systems, refuse
pickup systems, airline computer reservation systems, bank automated teller machine
systems, bank and non-bank credit card systems, bank debit card systems, bank check
and payment clearance systems, local real estate broker multiple listing services, and the
Internet (White 1999).
An important feature of network externalities is the Positive Feedback Cycle (or
Snowball effect) –the greater the potential for growth, the higher the number of future
participants. Watson (1998) argued that as in a natural monopoly, a network industry
would usually converge towards a single product (e.g., Windows computer operating
system). A lock-in happens when the costs of switching from a network become higher
than gains due to externalities. A bad lock-in occurs if inferior products win over others.
Positive feedback cycles can cause bad lock-ins. That is why the Microsoft Case has been
paid high attention and linked to the future of antitrust (2000b).
Positive feedback cycles can enable a firm to achieve monopolization very
113
quickly. Once achieved it may be even harder to resolve the monopolization due to
network effects (Rajiv 1998). Indeed winner-take-all (or most) competition is commonly
observed in network industries with snowballing --the more the customers, the more
valuable the product/service, the more the attractiveness of the product/service to even
more customers (Valente 1995). Customer switching costs become higher over time.
Also, due to the high-fixed costs of network firms, increasing sales decrease average
costs substantially. These factors can easily motivate an Internet company to give its
products for free (e.g., Adobe Acrobat Reader) to penetrate the market and set the
standard (Evans and Schmalensee 2001). The antitrust standards become irrelevant
because of the unique dynamic competition in the network industries to capture the
dominant position in the markets.
Network industries are very visible and include some that are key to the welfare
of the society such as utilities, software, telecommunications, cable services, credit cards,
and transportation (e.g., airlines). Thus, they are of uppermost concern for antitrust
enforcement bodies. The current antitrust approach is characterized by heavy use of cost
analysis and is not appropriate for all cases in dynamic markets, especially where high
level of fixed costs (e.g., networks) are involved (Sievers and Albery 1991).
Evans and Schmalensee (2001) argued that there is no cost-based test to
distinguish predatory innovation from non-predatory innovation in a winner-take-all
setting. The marginal cost of a network firm (with excess capacity) is often negligible.
Average Variable Cost (AVC) test provides a safe harbor for network firms enabling
them great flexibility to drop prices for predatory reasons. Network effects can enhance
the effects of market power particularly in dynamic industries. Stifled innovation may
114
cause bad lock-ins to obsolete technologies, resulting in additional loss of welfare. At
the same time, it is common for such companies to price their products/services below
AVC for non-predatory reasons (e.g., introductory low prices to penetrate the market).
The interconnections in network markets and the high tendency of market power in one
market to spill-over to other connected markets make the defense of strategic markets
even more critical.
4.2 Hypotheses
4.2.1.1 Market Power
Typically, the existence of market power has been simply inferred when an
incumbent has pre-dominant market share (1966, p.571). Firm(s) may use a monopolistic
market structure or conduct to achieve market power, which it then uses to achieve better
performance at the expense of its competitors and the competitive process. I define
market power as the ability to influence (pricing) conduct in the marketplace for the
purposes of the current research.
Structure-conduct-performance paradigm implies that higher the concentration in
a marketplace the further it is from the case of perfect competition, the higher the prices,
and the lower the societal welfare (Lopez 2001). Previous studies have generally shown a
positive correlation between market concentration and industry profitability (Shepherd
1970) among other factors.
Establishing the relationship between market power and pricing strategy is also
important from a public policy perspective because, in order to win a case, one of the
elements that the plaintiff must show is that the incumbent had market power to act upon.
It must establish that the incumbent used or pursued market power by anti-competitive
115
means.
Demonstration of market power is typically sought through high market share.
The structure of a market may be derived from the study of the number and size of firms,
their cost and demand conditions, product differentiation, the nature of any entry barriers,
and degree of regulation. Industry Concentration Ratio (percentage of total sales of the n
(usually four) largest firms in an industry) and the Herfindahl-Hirschmann Index ((HHI)
–sum of the squared market shares of all firms) are two commonly used market structure
indicators.
Pre-entry:
The incumbent’s pricing strategy for a specific market is heavily influenced by its
market power for the market in question. Market price premiums are conceptualized at
the three levels of strategic pricing options: supra-competitive, competitive, and limit
pricing, which is intuitive and consistent with the microeconomics view of market
performance.
1. Supra-competitive pricing: This refers to pricing at levels that yield premium margins
for the incumbent. This type of pricing is typical of a monopolist. The incumbent is able
to reap profits at levels that would not otherwise be possible because of a lack of
competition in the market place. Neo-classical economic theory suggests that supra-
competitive prices cannot exist in the long run because of the threat of potential entrants
(a.k.a. theory of market contestability). However, industry observations have shown that
firms can and do differentiate themselves and as a result, they are able exert consistent
116
price premiums in the marketplace especially when barriers to entry are high (Baker and
Pratt 1989).
2. Competitive Pricing: This refers to the going price levels in the marketplace in which
the incumbent makes reasonable profits (i.e., industry average for comparable markets).
Neo-classical theory would predict that competitive profits would converge towards zero
in the case of perfect competition. However, since markets are not perfectly competitive,
we observe modest levels of profits in competitively priced markets.
3. Limit Pricing: As discussed in the introduction, this strategy is also called entry
deterring pricing (Porter 1980, p.14). With limit pricing, the incumbent prices its services
low so as to prevent competition. Thus, the price is intentionally set low so that entry is
discouraged and (lower) profits are secured for a longer time. The apparent signal is low
current and future profits (Bain 1956). It should be noted that an entry-deterring price
does not necessarily have to be below (variable) cost to be anticompetitive. It could
negatively influence the entry decision of a more efficient firm at above cost levels.
Potential small scale entrants may assume that the incumbent enjoys economies of scale,
and potential large scale entrants may assume that the total demand is inelastic (i.e.,
increased supply will lead to even lower prices) (Gruca and Sudharshan 1995).
Limit pricing can also be justified by interests in protecting market leadership or
to prevent competitor growth in strategic territory. In doing so, the incumbent may utilize
signal-jamming to influence to potential entrants’ to believe that their costs are lower
than they actually are. The result would be little or no profits in markets where limit
pricing is employed. Similar to predatory pricing, the evidence so far is that the use of
limit pricing is rare (Smiley 1988). Low prices do not deter entry (especially against
117
innovative entrants, e.g., the cases of disruptive technologies or leaps in production
efficiencies (Christensen 1997; Han et al. 2001)), but they do diminish the chances of the
entrant’s survival in a given market (Lieberman 1989). Furthermore, limit pricing has
been documented to be a rational strategy in game-theoretic experiments (LeBlanc 1992;
Milgrom and Roberts 1982a). Thus, network industries may prove to be the ideal setting
to show that there may be more limit pricing in practice than meets the eye. If limit
pricing can stimulate and absorb the demand it stimulates, the residual demand upon new
entry may be too low to justify the entry effort (Eliashberg and Jeuland 1986).
Furthermore, both industrial organization literature and PIMS based studies
suggest that it pays off to be the dominant firm in an industry (Buzzel and Gale 1987;
Sudharshan and Kumar 1988). “The superior financial performance of businesses with
large market shares is attributable to their ability to obtain inputs at lower costs, extract
concessions from channel members, and set prices rather than be price takers” (cf.
Bharadwaj and Varadarajan 2004, p.223). Aggregate market concentration and
performance relationship has been found to be positive and significant in two meta-
analyses (Capon et al. 1990; Dutta and Narayan 1989). There is a well established link
between market power and price levels in economics and strategy literature (Abunassar
1994; Borenstein 1989) and it is expected that this will hold true for the network
industries as well.
4
Therefore,
H1: Market power of the incumbent and the market price premium will be
positively associated.
4
It should be noted that some scholars have argued that the market share-performance
relationship does not exist and may be attributed to third factors (e.g., Symnanski et al. 1993).
118
4.2.1.2 Barriers to Entry
“Entry conditions are central to antitrust analysis”
John M. Nannes, Deputy Assistant Attorney General, Antitrust Division (1999).
Barriers to entry can be defined as factors that prevent an entrant from competing
on an equal footing with the incumbent(s) (Dolan 1986, p.602). For a particular industry,
if the barriers to entry are high and retaliation from the incumbents is expected, potential
entrants are not likely to be enthusiastic about entry (Mintzberg and Quinn 1996).
Karakaya and Stahl (1989) used a broad categorization of the concept and identified
nineteen barriers to entry. Their list included cost advantage of incumbents, product
differentiation of incumbents, capital requirements, customer switching costs, access to
distribution channels, government policy, advertising, number of competitors, research
and development, price, technology and technological change, market concentration,
seller concentration, divisionalization, brand name or trademark, sunk cost, selling
expenses, incumbent’s expected reaction to entry, and possession of strategic raw
materials (essential facility). They later advanced this list to 25 items for consumer goods
markets (Karakaya and Stahl 1992).
Porter (1980) identified six major categories for barriers to entry. These six
barriers were cost advantages of incumbents, product advantages of incumbents, capital
requirements, customer switching costs, access to distribution channels, and government
policy. Barriers to entry may potentially enable the firm to recoup profits over a longer
term and make predatory pricing rational. Multiple industry data also indicate that firms
use entry deterrence less frequently when other barriers exist (Bunch and Smiley 1992).
The model posits that the market power of the firm is a primary determinant of its
119
pricing strategy. However, this link should hold true and probably be even stronger with
the inclusion of the comprehensive barrier to entry measures that include firm specific
(i.e., brand equity) and those determined by outside forces (e.g., regulation --tobacco
settlement that requires new entrants to pay additional State taxes along with the industry
giants that originally caused the case (Forbes 2005)). The industrial organization
literature argues for the positive effect of barriers to entry whereas the efficiency school
suggests that barriers are not necessary for superior performance (Bharadwaj and
Varadarajan 2004). This leads to the following hypothesis:
H2: The positive relationship between the market power of the incumbent and the
market price premiums will be positively moderated by barriers to entry.
It is also possible that barriers to entry could have a direct effect on the pricing
strategy of the incumbent, thus forming a quasi-moderation effect. This should be
investigated as an alternate specification. Therefore,
H3: Barriers to entry and the market price premium will be positively associated.
Of particular interest here are the firm specific barriers that can be enhanced as
opposed to market specific barriers (e.g., regulation) that are generally beyond the control
of the incumbent. Neoclassical price theory traditionally emphasized the choice for the
industry as the strategic decision whereas the Resource Advantage theory predicts that
firm specific barriers would be more important than market specific barriers (Viscusi et
al. 1995). Varadarajan and Jayachandran (1999) observed that the earlier focus in
industrial organization literature (Bain 1956) on “why some firms are more profitable”
later shifted to “why some firms are more profitable” (Demsetz 1973), and ultimately, the
resource-based view of the firm (Barney 1992) where profitability is determined by
120
competitive advantage. Hunt (2002, p.287) observed the state of the theoretical debate on
the firm specific versus industry (i.e., market) specific factors:
By the time of McGahan and Porter (1997), the entire nature of the debate
over firm performance had changed dramatically. Originally, advocates of
industry-based strategy (e.g., Montgomery and Porter 1991) were citing
Schmalensee (1985) to justify their focusing on “choosing industry” as the
key strategic decision. After Rumelt’s (1991) replication and extension of
Schmalensee found firm factors to account for almost six times the
variance of industry factors (46 percent vs. 8 percent), the debate shifted
toward whether industry choice at all, Thus, McGahan and Porter’s
(1997) study, which finds that firm effects dominate industry effects by
only 36 percent to 19 percent, is interpreted by its authors as confronting
the challenge from Rumelt and others that industry, far from being key,
doesn’t seem to matter at all. The point to be emphasized here is that no
one now claims empirical support for the neoclassical position. That is,
after Rumelt’s (1991) and other studies, no one argues seriously the
neoclassical position that either industry is everything or industry effects
dominate firm effects.
McGahan and Porter (2002) found even lower influence of industry factors on
performance than their earlier findings (48 percent business-unit factors and 10 percent
industry factors). Still, market-specific barriers (e.g., institutional environment) can be
important from a public policy perspective since all types of barriers can be antecedents
to monopoly conditions in a market (Yip 1982). Karakaya (2002) reported four major
underlying dimensions of barriers (i.e., firm specific, product differentiation, cost of
market entry (financial requirements), and profit expectation of entering firms) with firm
specific barriers being the most important. His study was in an industrial setting. I inquire
if this finding will hold true in network industry setting as well. The relative effectiveness
of these two categories is also interesting to verify good, and potentially bad and ugly
effects that barriers to entry may have (Han et al. 2001). Therefore,
H4: The positive moderating effect of firm specific barriers on the incumbent’s
pricing strategy will be higher than that of market specific barriers.
121
Firm specific barrier influence would also be expected as direct effects.
Therefore,
H5: The positive effect of firm specific barriers on the incumbent’s
pricing strategy will be higher than that of market specific barriers.
These hypotheses, if supported would also have managerial implications since
building firm specific barriers is a prerogative that managers typically do not have with
market specific barriers.
4.2.1.3 Strategic Assessment
The model also alludes to the role of potential entrants’ strategic assessment for
entry for the market in question. Game theoretical analyses have indicated that
competitors’ moves are calculated by investigating its resources and entry patterns (e.g.,
Milgrom and Roberts 1982b). The resource-advantage theory advanced by Hunt (2000)
suggests that resources play a critical role in the long term prosperity of the firm. Firm
growth is constrained by internal management resources (Penrose 1959). Available
resources (e.g., human resources, capital) constrain the choice of markets for entry
(Wernerfelt 1984b). For example, employees (relations), type and capabilities of the
aircraft, and management leadership that potential entrants have can also be potential
resources in the airline industry context. The fit of the market with potential entrants’
existing portfolio would also be important. It is not likely that an existing airline would
start flying a route that is not connected to its existing routes on either end (Dixit 2000).
The attractiveness of the market is also an important part of the strategic assessment
(Baldwin 1995) as growing markets are more likely to be under the threat of new entrants
122
both in the form of existing or start-up competitors (Gatignon et al. 1989). The successful
analysis and perception of these issues have been linked to higher performance (Clark
and Montgomery 1996). The incumbent is less likely to employ supra-competitive
pricing and make the market more attractive for entry if it deems that the potential for
entry for a given market by inter-competition is high after strategic assessment.
Therefore,
H6: The positive relationship between the market power of the incumbent
and the market price premiums will be negatively moderated by the
incumbent’s strategic assessment (i.e., resources, strategic fit,
market growth) of the potential entrants.
H7: Incumbent’s strategic assessment and the market price premium will be
negatively associated.
4.2.2 Entry Decision
The questions that an incumbent should normally answer before reacting to a
competitor’s price cut are: “is the price cut likely to have significant impact on our
sales?” and “is it likely to be a permanent price cut?” If the answer to any of these two is
negative, then there actually is no need to react to a price cut (Kotler and Armstrong
1991) (also see Figure 3.8). However, in practice (i.e., with information asymmetry), it is
not very easy to answer these questions accurately, especially with a strategic perspective
and that encompasses uncertainty. A price cut insignificant today could reshape the
industry landscape tomorrow (e.g., the case of Dell Computers). Similarly, price levels
that are below cost and non-sustainable today could become sustainable with economies
of scale tomorrow (e.g., eBay.com, Amazon.com). Incumbent firms in network
industries, especially when challenged by low-cost start-ups with drastically lower cost
structures, tend to fear the worst (e.g., Microsoft versus Linux; American versus
123
Southwest Airlines). In high–tech industries, this fear is partially transformed into
defensive acquisitions by leading firms (e.g., acquisitions by Ticketmaster; Microsoft;
Cisco). However, in recently deregulated industries where merger and acquisitions have
to be approved after a detailed antitrust scrutiny, it can be quite rational to employ non-
price and predatory pricing schemes. Thus, managers in these firms may answer
affirmatively when they strategically analyze the two important questions. Recoupment
may not be feasible in the short run and not even in the foreseeable future. But perhaps
these firms are sacrificing gains and incurring losses in exchange for lower future losses
or even survival.
The entrant can fairly accurately observe the pricing strategy of the incumbent.
Upon analysis of its resources, strategic fit, the market growth, and the current pricing
strategy of the incumbent, the entrant decides whether or not to pursue entry to the
market. The framework suggests that the response of the incumbent will be heavily
influenced by its pricing strategy for that market. If the incumbent were using supra-
competitive pricing, it is anticipated that it will predate (sharp price and capacity
reactions) to protect its market power, if the incumbent were using competitive pricing, it
is anticipated that it will react competitively (reasonable price and/or capacity reactions),
and finally if the incumbent was using limit pricing, it is anticipated it will accommodate
(insignificant price and/or capacity reactions). This structure has not been previously
empirically tested, however, content analyses of reports from the popular press support
these assertions. In the predation case, the odds are against the entrant to establish itself
and succeed in the marketplace (exceptions such as Southwest Airlines can occur
depending on the strategy, efficiency and the funds available). In the competitive reaction
124
and accommodation cases the entrant is increasingly more likely to survive given that it is
more efficient than the incumbent and managed well.
The following set of hypotheses are also supported by the signaling literature
which suggests that price levels are indicators of market potential and the cost structure
of the incumbent (Heil and Walters 1993, Prabhu and Stewart 2000). Heil and Robertson
(1991) argued that that the major benefits from signaling are preemption and
development of competitive norms of conduct, and proposed that market power and
antitrust action due to price signaling would be positively related. This verifies the
before-during-after price modifications pattern previously mentioned.
Adam Smith (hence, the classical industrial organization literature) would also
predict that capital flows into markets with above average returns on investment. This
capital flow could be in the form of new market entries. Still, this relationship should not
be taken for granted. Dixit (2000) hypothesized that the higher prices would lead to
higher probability of entry to markets (in the airline industry) but was perplexed by a
negative relationship. Therefore,
H8: Incumbents’ pre-entry market price premiums for markets with inter-category
entry will be higher than those of markets without inter-category entry.
The rationale for the consideration of strategic assessment for Hypothesis
7, and the Resource Advantage theory also applies to the entry decision phase.
Therefore,
H9: Potential inter-category entrants’ strategic assessment of the markets will be
higher for the markets that they enter.
125
Pre-entry defense summary: The model posits that the main positive effect of market
power on pricing strategy is moderated by the incumbent’s strategic assessment of
potential entrants’ and the barriers to entry. The moderating effect of barriers to entry is
positive. However, the moderating effect of strategic assessment is expected to be
negative. Direct linkages from both factors to pricing strategy are also plausible. Ceteris
paribus, the trade-off between these countervailing forces determines the pricing strategy
of the firm. If the positive effects are dominant, the firm is likely to employ a supra-
competitive pricing strategy resulting in large fare premiums (and loss of consumer
welfare). If there is a balance between these effects, the firm is likely to employ
competitive pricing. Finally, if the negative effects are dominant, then the firm is likely to
engage in limit (entry-deterring) pricing which results in lower prices in the short run but
has negative welfare effects in the long run due to lessened competition (and decreased
efficiency due to stifled competition and potential innovation).
Potential entrants’ own strategic assessment and the pricing strategy of the
incumbent lead to entry analysis. The framework flow is interrupted, and equilibrium is
reached if there is no entry. However, if the challenger decides to enter the market, a
response by the incumbent is triggered as measured by a price change. This response may
be categorized at three levels, which are discussed next. The incumbent may choose to
predate, to react or to accommodate depending on its pre-entry strategy. If the entrant is
able to endure the retaliatory response of the incumbent, a new market equilibrium will
be observed.
126
4.2.3 Post-entry Defense
Trend analysis indicates that the firms react sharply when an entry by a discounter
(i.e., inter-type entry) occurs in their markets (Oster and Strong 2001). Previous work on
competitive interaction also suggests that retaliatory actions to entry depend on the
perception of threats (Kuester et al. 1999). Namely, it is expected that the incumbent
would react most sharply in cases where its supra-competitive profits are threatened. As
discussed in Chapter 3, the current tests of predatory pricing are based on average
variable cost, the courts are not effective in distinguishing predatory action from vigorous
competition in network industries. Therefore, the extent of paired price and capacity
reaction could easily reach the scale of predation without being detected by the courts.
The evidence for that would be consistent with observations of market exits by the inter-
category entrants (presumably more efficient than the incumbents) which were likely to
survive had predatory tactics not been utilized. Supra-competitively priced markets
improve the profit margins and are more valuable and strategic from the perspective of
the incumbents. Hence, the expected incumbent retaliation in the case of inter-entry to
supra-competitively priced markets is predation. The reaction of the incumbent will likely
not be as drastic in the case of entry to competitively priced markets. Finally, the
incumbent is expected to react little or not at all in the case of entry to markets where
limit pricing was employed (i.e., accommodation).
The logic of accommodation is supported by the widely established Defender
model (Hauser and Shugan 1983) which predicts that an opposite reaction such as cutting
back on advertising or increasing price can be the optimal strategy against market entry.
Moreover, there are several PIMS based studies that reported that no or limited reaction
127
to market entry is the norm (Biggadike 1979; Robinson 1988; Yip 1982). Retaliation was
associated simply with high-growth markets.
On the other hand, price reduction was found to be the optimal response against
competitive entries (Gruca et al. 1992). It is well documented that hostile acts trigger
stronger competitive actions (Heil and Walters 1993). Due to the interconnected nature of
the networks, the number of markets that are considered strategic could be even higher.
MacMillan and colleagues (1985) have reported that such strategic challenges accelerated
competitive response. Swift and more aggressive responses are expected if the focal
market of entry is viewed important by the incumbent (Chen and MacMillan 1992; Chen
et al. 1991). Assuming that firms derive their market power from strategic emphasis to a
given market, and that market power and price are positively correlated, it is only natural
that the entries to markets that are supra-competitively priced would attract more
intensive retaliation to entry than others. Bowman and Gatignon (1995) also concur that
the retaliation is delayed when the incumbent has low market share. An incumbent that
does not react to entry to markets where its profits lie may send signals of weakness to
potential competitors and invite further entry. “If the strategy fails and entry occurs,
consequences for the incumbent firm can vary depending on which strategy was chosen
[prior to entry]. Certain strategies may leave the incumbent in a worse competitive
position, whereas others may lead to a stronger posture after entry ” (Gruca and
Sudharshan 1995, p.44). For example, “the airline industry has certain characteristics that
make a predatory theory more than plausible” (Nannes 1999). To match a move is a
strong signal by itself, indicating unwillingness to give up a position without escalating
the war to mutually destructive levels (Chen and MacMillan 1992). Robinson (1988)
128
observed that the current strategy may influence future as well as the competitive
conduct. Therefore,
H10: The magnitude of the incumbent’s response to inter-entry will be
positively associated with its pre-entry pricing strategy.
4.2.4 Exit Decision
As detailed in previous sections, the inter-type entrants are typically subject to
aggressive price cuts when they enter strategic markets. The incumbents generally target
their responses so that the inter-entrants are driven out of their key markets before they
can establish their structure and inflict considerable damage to the incumbents. However,
since the inter-entrants are presumably more efficient than the incumbents due to their
operation and cost structures, their exit patterns would also be affected by their financial
resources and the price elasticity of the market. When discounters enter markets, the
passenger volume typically increases significantly more than the percentage decrease in
average price. Therefore, the decision to exit the market will be effected by the
incumbent’s response and the post-entry strategic assessment of the entrant.
Incumbent’s choice for reaction strategy sends a signal to the entrant as to how
determined the incumbent is to deter entry (Heil and Walters 1993). A price matching
move is a strong signal in its own right (Shelling 1960), however this move becomes
even more powerful given the context that the entrant is a discounter with less a
comprehensive value proposition. Accurately perceiving the intent of competitive
reactions enhances firm performance (Clark and Montgomery 1996; Day and Nedungadi
1994). This has further implications than just sheer financial impact of price reductions.
A sharp price cut may cause the entrant to leave the markets upon strategic assessment
129
even though it might have made more economic sense to fight back due to a more
efficient structure of the entrant (Gundlach 1995). Successful retaliation has been
associated with holding a competitive advantage (e.g., brand equity), low scale of new
entry, and low to medium access to resources by the new entrants (Robertson and
Gatignon 1991), and all these conditions are met in the inter-type competition framework.
In essence, the entrant weighs the strategic benefits against the threats posed by the
incumbent’s retaliation and makes the decision to exit or stay. Therefore,
H11: The likelihood of inter-category exit from a given market and the magnitude
of the incumbent’s competitive response will be positively associated.
H12: The likelihood of inter-category exit from a given market and the entrant’s
strategic assessment will be negatively associated.
4.2.5 Policy Consequences
Increased long-term social welfare is the ultimate goal of antitrust. Lower prices
and competition increase the welfare of the consumers and the society in general (Grewal
and Compeau 1999, p.3). However, if the consumers feel that the price they have to pay
is unfair, then social harm may occur (Guiltinan and Gundlach 1996a). The courts
consider economic harm to consumers as the best way of assessing harm to society (Baer
1996). However, a long-term assessment of social welfare should not be simplified to
allocative efficiency (Guiltinan and Gundlach 1996a). For example, Bloom and
Gundlach (2001a) identified thirteen paths through which marketing affects consumer
welfare.
130
Economic Efficiency
Perspective:
Chicago:
• Allocative Efficiency
• Productive Efficiency
Post-Chicago
• Dynamic Efficiency
Marketing Oriented
Perspective:
• Affordability
• Variety
• Satisfaction
• Convenience
• Price Stability
• Perceived Value
Innovation
Competition
Figure 4.2: Marketing Extension for Consumer Welfare Assessment
Consumer
Welfare
Figure 4.2: Marketing Extension for consumer Welfare Assessment
The long-term consequences of predatory pricing (or any anti-competitive action
for that matter) on consumer welfare are central to antitrust analysis. Marketing insights
for a more informed consumer welfare analysis are presented in Figure 4.2 and Table 4.1.
The shortcomings of just focusing on allocative efficiency for assessing consumer
welfare, and marketing’s potential contributions for its operationalization and conception
have been observed by many scholars in marketing and elsewhere (cf. Gundlach et al.
2002).
131
Table 4.1: Dimensions of Consumer Welfare and Marketing Insights
Dimensions of
Consumer
Welfare
Marketing Insights Public Policy
Implications
Efficiency Local inefficiencies
may disrupt the general
efficiency of the
markets
An antitrust analysis that
goes beyond economic
efficiency
Affordability Focus on net income is
too narrow
Affordability is a better
measure of welfare than
income
Satisfaction Established scales such
as SERVQUAL exist
Track satisfaction from
market as well as market
concentration
Variety Broader definition of
variety
Discount for seemingly
variant offerings. Let the
consumer define variety.
Price Stability Fluctuating prices are a
great nuisance to
consumers. Price
confusion is
detrimental to social
welfare
Active public
communication/advertising
and legislation to prevent
price confusion
Convenience Non-cost based tests
may be relevant; the
level of convenience
can serve as such a
measure
Monitor the level of
convenience before/after
predatory action as well as
price levels. (side note:
investigate unjustified
convenience charges)
Innovation Product Life Cycle
concept and impact of
line/brand extensions
versus radical
innovations
Emphasis to prevent bad-
lock-ins to obsolete
technology. Support
superior technology by
adoption/promotion
Competition Track market share and
sales but also
signaling, reputation
effects, and strategic
decision making
Fund and utilize research
that quantifies the impact
of signaling, reputation,
and strategic decision
making
132
4.2.5.1 Consumer Welfare
Affordability. There is more to measuring social welfare than efficiency. Purchasing
power is a classic variable considered by the Chicago School/courts in terms of price
levels. From a marketing perspective, purchasing power goes beyond the notion of the
basic price itself, but also encompasses the notion of affordability which includes credits
terms, payment options, etc.
Variety. Variety (choice) is a variable that is of paramount importance for consumer
welfare (Guiltinan 2002; Lande 2001). A high number of players in an industry does not
necessarily imply variety in commoditized markets. Therefore, perceived (buyer
defined/unobservable) variety mix is more important to capture than simply supply
variety. Marketing implications for a choice-centered antitrust policy have been identified
from both demand and supply perspectives (see Guiltinan 2002).
Satisfaction. Buyer satisfaction is a key construct in marketing, yet considered to be too
difficult to deal with and ignored in welfare analyses by economists. Marketing has a
significant history of research with the satisfaction construct for both products and
services and has established scales to measure it. Historical data can be used to estimate
the impact of a predatory strategy on buyer satisfaction in the long run.
Convenience. Overall convenience associated with the use of a product or consumption
of service can also be an important dimension of satisfaction and welfare. Related
dimensions include decision-making, access, transaction, benefit and post-benefit
convenience (Berry et al. 2002). Levels of convenience provided to the buyers before the
alleged predatory action should be compared to those of the period following it.
Price stability. Predatory pricing typically results in drastic price fluctuations especially
133
in network industries. For example, in one DOJ alleged case of predatory pricing,
American Airlines cut its prices by 26 per cent, but then raised it by 84 per cent after
driving out the competition (Carney and Zellner 2000). Fluctuating demand and prices
can prove fatal especially for small manufacturers that do not hedge their risks
effectively. Similarly, unstable prices can be a source of great inconvenience and
confusion for the consumers. Buyers may feel the price they paid is unfair (Smith and
Nagle 1995; Zeithaml 1988) or they may even feel betrayed. At the same time, it is
common for price-discriminating businesses to fuel price confusion among buyers in
order to avoid competing on price (Grewal and Compaeau 1999). Since each of these
options diminishes consumer welfare, the social implication is to communicate/advertise
and sometimes even intervene to prevent price confusion and unnecessary price
fluctuations.
The positive impact of lower prices on consumer welfare is generally accepted.
However, predatory pricing is detrimental to consumer welfare in the long run because
once the competitors exit the market, the predator raises prices with the intention of
collecting supra-normal profits and recouping its predatory investment. Moreover, the
problem with predatory pricing is not limited to harm to consumers through the increase
of prices back to monopoly levels. Predatory pricing, successful or not, can potentially
reduce incentives for investment and innovation, and prevent new entry or expansion by
more efficient firms. There are special implications for network industries such as
telecommunication and software where the value of the product/service increases along
with the number of users. Innovation can be stifled when predatory prices induce
134
consumers to continue to use an old technology as opposed to a superior alternative
offered by a new entrant (Guiltinan and Gundlach 1996a).
Grewal and Compeau (1999) suggested that marketing researchers have not
engaged in public policy implications of pricing until recently and that a focus on this
issue is long overdue. They argued that developments such as the internet, global
markets, mega-corporations, and cooperative marketing arrangements created the
necessity of taking a closer look at the pricing and public policy interaction with
consumer welfare in mind. After all, courts consider economic harm to consumers as the
best way of assessing harm to society (Baer 1996). Guiltinan and Gundlach (1996a)
argued that marketing was in a unique position to help form public policy guidelines with
comprehensive measurement and modeling procedures, and that predation and predatory
pricing have not been addressed by marketers until recently. Gundlach (1995) suggested
that the marketing discipline had the potential to further the understanding needed for the
development of a more suitable antitrust policy.
In all cases of exit by inter-entrant there will be a loss to consumer welfare
stemming from the assumption that the inter-entrant was presumably more efficient. The
loss to consumer welfare (mainly in terms of purchasing power and price stability) occurs
because the incumbents are typically able to increase the prices to pre-entry levels and
sometimes even higher to recoup losses incurred during predation. The long run effect of
predation is higher prices for the consumers.
It appears that there is an increasing gap between the insights from the modern
economic theory and the enforcement of current judicial policy. Government
enforcement concern has never been higher in many years. The new economy requires
135
new rules for the assessment of predation because of the growing importance of
intellectual property (e.g., Microsoft Case). Increasing market concentration in many
industries and number of mergers, are also of concern (Bolton et al. 2000). Obviously,
measuring social welfare with comprehensive measures itself constitutes a dissertation
topic and is well beyond the scope of this dissertation. However, the market power of the
incumbent can be practically used as a proxy of long term social welfare and the
following hypotheses can be tested. Therefore,
H13: The inter-entrant’s exit will result in loss to consumer welfare (as
measured by the increase in the incumbent’s post-exit market power).
If supported, the above hypothesis would undoubtedly generate public policy
implications. The overall premises of the proposed framework are summarized in Table
4.2.
136
Table 4.2: Summary of the Framework Premises
Context Theoretical Base
H1-
H7
Antecedents of Pricing Strategy:
• Market power leads to higher price
premiums
• Incumbents consider threat of potential
entrants when deciding strategy
• High barriers to entry prevent market
contestability
• Firm specific barriers more influential
than market specific barriers
Industrial Organization, Strategy,
Resource Advantage, Market
Segmentation
H8-
H9,
H12
Antecedents of Market Entry:
• Entrants will consider price levels as
indicators of efficiency and potential
• Entrants will conduct strategic analyses
before entry and exit
Industrial Organization,
Signaling, Game Theory, Entry
Deterring Prices, Reputation
Effects, Multi-market competition
H10
Consequences of Entry:
• Incumbents will respond differently
depending on the market of entry
Competitive Interaction, Inter-
Intra-type Competition, Multi-
market competition, Game
Theory
H11
Determinants of Market Exit:
• Entrant is likely to exit in the face of
drastic response by the incumbent
Competitive Interaction,
Signaling, Resource Advantage,
Game Theory
H13
Consequences of Market Exit:
• Exit by will have negative
consequences on long and short-term
consumer welfare.
Post-Chicago Economics, Public
Policy and Marketing Interface
137
CHAPTER 5
CONTEXT: AN INQUIRY OF THE DYNAMICS OF COMPETITION IN THE
U.S. AIRLINE INDUSTRY
It was a love of the air and sky and flying, the lure of adventure, the
appreciation of beauty. It lay beyond the descriptive words of men –where
immortality is touched through danger, where life meets death on an equal
plane; where man is more than man.
Charles Lindbergh, The Spirit of St. Louis. 1953
This is a nasty, rotten business.
Robert Crandall, American Airlines, 1994 (Petzinger 1995).
The airline industry has certain characteristics that make a predatory
theory more than plausible (Nannes 1999).
An inquiry of price competition in a key network industry –airlines, provides a
unique perspective into the dynamics of competition. The main contributions of this
chapter are two-fold: it introduces the deregulated passenger airlines as an ideal context
for empirical research, and describes the history and competitive landscape of the
industry.
Several unexpected consequences were observed as a result of the airline
deregulation in 1978. These included increased range of services at the expense of overall
service quality, the dominance of hub-and-spoke systems, and the subsequent fare
structure that penalized passengers flying out of major hubs in the form of hub premiums.
Contrary to popular belief, the effect of deregulation on decreasing fares was not robust.
In comparison to pre-deregulation era, tickets today cost less on longer haul routes
138
regardless of the nature of competition, but not on shorter routes where low cost carrier
(LCC –such as Southwest or Air-Tran) competition is absent. Waves of LCC entrants
were driven out because of cut-throat competition and/or predatory practices of the major
carriers. The existence of such practices is suggested by the observations and actions of
the DOT and DOJ.
The eighties saw the emergence of yield management systems, frequent flier
programs and the widespread use of computer reservation systems. The primary
distribution and sales channel was the travel agencies, which lost much of their power
after the explosion of the Internet, and the subsequent movement to cut their commissions
throughout the industry.
After September 11, the LCCs have performed better than the major carriers,
which could not respond to increased pressure to cut costs. Anticipated near future trends
in the airline industry include consolidation attempts among the major carriers,
subsequent antitrust action, revision of the hub-and-spoke system and re-organization,
restructuring of corporate travel policies, and increased market access by LCCs aided
either by federal or local governments.
5.1 Historical Perspective
5
As the phrase “America on wheels” indicates, cars are often referred to as the
foundation of American culture. However, by the 1990’s the number of adult individuals
who owned cars were less then those who had flown (Petzinger 1995, p.i13). Convenient
access to air service is one of the key considerations for choosing business locations. This
5
The historic facts in sections 5.1.1, 5.1.2 and intro of 5.1.3 have been adopted from the Air Transport
Association web site (www.air-transport.org) except where cited.
139
chapter will present the rich history of the U.S. Commercial Aviation with its ups and
downs and bring the reader to date. An overview is provided in Table 5.1.
Table 5.1: The U.S. Airline Industry Timetable
The Early Years
1903 First flight by Wright Brothers at Kitty Hawk, NC.
1914 The first scheduled air service starts between St. Petersburg and Tampa,
FL.
1918 First Airmail delivery
1925-26 The Contract Air Mail Act and The Air Commerce Act of 1926 passes
1933 The first modern passenger airliner the Boeing 247 can carry 10
passengers
1934 Air Mail Act of 1934 is enacted due to scandal regarding the mail bidding
process
1938 The Civil Aeronautics Act of 1938 is passed. CA Board commences
1940 The first plane with pressurized cabin, the Stratoliner is introduced by
Boeing
1943 The U.S. builds its first jet plane Bell P-59
The Jet Age
1958 Boeing remodels and introduces its KC-135 jet tanker as the first U.S.
passenger jet, the 707 (capacity 181 passengers)
The Federal Aviation Act is passed to monitor the growing airline industry
1963 Sabre Computer Reservation System introduced by American
1967 Department of Transportation (DOT) is created
1969 Boeing launched its first widebody jet the 747 in 1969. It could carry 450
passengers. The first supersonic plane the Concorde is introduced
1978 The Airline Deregulation Act is passed. A new era begins…
New Entrant Low Cost Carriers (LCCs) start to compete for markets
1981 American offers the first frequent flier program: AAdvantage
1982 Braniff goes bankrupt, the first case for a major airline since 1938
1984 Deregulation is considered to be successfully completed, and the CAB is
abolished
Most remaining functions are transferred to the DOT
Accusations of predatory pricing occur
Antitrust ruling enforces equal listing on Computer Reservation Systems
1986 People Express goes bankrupt mainly as a result of American’s revenue
management system
1986-89 Mega-Mergers era: Concentration increases and fortress hubs are formed
1992 American ignites a price war with its “value pricing” campaign and causes
many bankruptcies and the worst performance in the industry
1993 The peak year for Travel Agency share in bookings
1995 Explosion of the internet
140
ValueJet Crash causes misfortune for discounters
Majors employ aggressive/predatory pricing to contain/destroy LCCs
1997-98 Competition stagnation: No new entrants in any markets
1998 DOT proposes guidelines to prevent anticompetitive practices in the
industry
1999 DOJ reviews complaints and files a lawsuit against American
May 2001 American case dismissed by summary judgment. DOT appeals
September
2001
Terrorist attacks cause turmoil in the industry that was already sending
alarming signals
2002- Majors carriers drop commissions to travel agents
LCCs gain market share and force majors to become more efficient
5.1.1 The Propeller (Early) Era
The first flight was accomplished by Orville and Wilbur Wright on December 17,
1903 at Kitty Hawk, North Carolina. All lasted 12 seconds for 120 feet distance but it
was the first flight with the exception of balloons and gliders. The first U.S. passenger
flew with Orville Wright five years later in 1908 and the transportation industry would
never be the same.
The first scheduled air service started in 1914 between St. Petersburg and Tampa,
Florida. The seaplane had a capacity of one passenger and made two flights a day, and
the service cost $5 one-way. The company ceased operations after only four months.
Early flights were always great publicity but commercial aviation did not become
widespread until much later.
World War I necessitated more powerful engines and resulted in planes that could
fly at 130 mph. Increased power also enabled larger aircraft. However, the end of war left
a huge surplus of planes (i.e., no demand for new planes) and many aircraft builders went
out of business. The railroad was still the primary means of transportation for Americans.
Table 5.1: (continued)
141
The U.S. government decided to put the surplus planes to use. In 1917, The Congress
approved $100,000 for an experimental service and the first airmail was delivered from
New York to President Wilson in Washington on May 14, 1918. By 1920, the Post Office
was able to save almost a day on transcontinental deliveries through airmail.
The fleet, able to fly at night by mid-1920s, flew an average of 2.5 million miles
and delivered 14 million letters annually. However, the Post Office typically contracted
private companies for mail transportation, and had no intention of staying in the airmail
business. The Contract Air Mail Act (also referred to as the Kelly Act after its main
supporter, Rep. Clyde Kelly of Pennsylvania) was passed in 1925.
The Air Commerce Act of 1926 gave the Secretary of Commerce the authority to
designate air routes, and license pilots and aircraft. National Air Transport (owned by the
Curtiss Aeroplane Co.), Varney Air Lines, Western Air Express, Colonial Air Transport
and Robertson Aircraft Corporation were the initial contract winners. They would form
the core of the U.S. private airlines: “National and Varney would eventually become
important parts of United Airlines. Western would merge with Transcontinental Air
Transport (TAT), another Curtiss subsidiary, to form Transcontinental and Western Air
(TWA). Robertson would become part of the Universal Aviation Corporation, which in
turn would merge with Colonial, Southern Air Transport and others, to form American
Airways, predecessor of American Airlines. Juan Trippe, one of the original partners in
Colonial, later pioneered international air travel with Pan Am - a carrier he founded in
1927 to transport mail between Key West, Florida, and Havana, Cuba. Pitcairn Aviation,
yet another Curtiss subsidiary that got its start transporting mail, would become Eastern
Air Transport, predecessor of Eastern Air Lines.”
142
Henry Ford was also among the early contact winners. He carried mail from
Detroit to Chicago on the same planes that transported spare parts for auto
manufacturing. In 1927, Ford introduced the Trimotor (also referred to as the Tin Goose).
The “Tin Goose” was the first plane designed specifically to carry people and had a
capacity of 12 passengers. The Ford brand was assuring for the public but it took Charles
Lindbergh’s historic flight across the Atlantic to bring flying to unprecedented public
attention. Lindbergh became an instant hero when he completed his non-stop trip from
New York to Paris on May 21, 1927. The Air Age had commenced.
The 1930 Watres Act (after one of its main supporters, Rep. Laurance H. Watres
of Pennsylvania) authorized the Post Office to sign long-term airmail contracts based on
space or volume as opposed to weight, and to consolidate airmail routes where necessary.
The idea was to promote larger, hence stronger airlines enabling more frequent and faster
mail delivery. The Post Office held a number of meetings attended by select larger
airline executives (thus, these meetings were later referred to as the Spoils Conference).
The idea was to have one company operating on each of three transcontinental mail
routes as opposed to mail changing hands between several smaller airlines until it reached
its destination.
Following the victory of the Democrats in 1932, smaller airlines complained that
there had been unfair practices in the bidding process. In one case, it was discovered that
a big airline was favored over a smaller airline even though its bid was three times
higher. Congressional hearings followed, and by 1934 the scandal had reached such
proportions that President Roosevelt cancelled all mail contracts and gave the job to the
Army. Unfortunately, the Army pilots were not familiar with the routes and the
143
corresponding weather conditions. A number of accidents took place during practice runs
and President Roosevelt had to retreat from his plan just one month later. The result was
the Air Mail Act of 1934. The airmail services would be run by the private sector but the
former contractors were not allowed to bid, and the bidding process was re-structured to
be more competitive. The resulting lower margins mandated that the airlines pay more
attention to the passenger business. The government also put a halt to vertical integration
in the industry and manufacturers and operators were separated (e.g., Boeing, Pratt &
Whitney, and United Airlines) resulting in a more focused and re-organized industry.
Arguably, the 1930s were the most innovative period in aviation history. Safer,
larger, and faster planes were necessary to have a feasible airline business. Air-cooled
engines, better cockpit instruments (e.g., improved altimeters, airspeed indicators, rate-
of-climb indicators, compasses) and the introduction of artificial horizon were among the
innovations of the period. Naturally, the wide spread use of radio was also of prime
importance. Eighty-three radio beacons across the country were fully operational in 1932.
The first air traffic control tower was established at Newark, New Jersey in 1935.
Launched in 1933, the first modern passenger airliner is considered to be Boeing
247. The 247 had a capacity of 10 passengers and could fly 155 miles per hour. United
Air Lines purchased sixty 247s. TWA wanted to outdo United and the search for a better
alternative led them to the Douglas Aircraft Company. Douglas’ prototype DC-1
incorporated and improved upon many of Boeing’s innovations. The longer version the
DC-2 could accommodate 14 passengers and was a big hit. However, DC-3, later called
the plane to change the world, had a capacity of 21 passengers, and was considered cost-
efficient, safer, more comfortable, with more powerful engines. It could complete a coast-
144
to-coast trip in 16 hours. Following the marketing concept of product development,
Douglas got American Airlines heavily involved in the design process of DC-3. The
result was the first plane that enabled airlines the make money from passenger business.
DC-3 became very popular and introduced many new travelers to the joys of flying.
Despite the success of DC-3, a technical difficulty remained: The airlines wanted to fly
higher so that they could avoid air turbulence and storms at lower altitudes. Motion
sickness was also a problem for many passengers. However, they could not fly higher
than 10,000 feet because the lack of oxygen made passengers dizzy or even unconscious.
In 1940 the solution was called Stratoliner by Boeing, the first plane with pressurized
cabin. First embraced by TWA, the Stratoliner had a capacity of 33 passengers, could fly
at 20,000 feet at 200 miles per hour.
There were also political problems resulting in economic problems for the
airlines. Before the Civil Aeronautics Act of 1938, many government agencies were
involved with the airlines resulting in bureaucracy and no long-term policy for the
industry. All the airlines were losing money due to the reduced mail revenues since the
1934 Airmail reform. The wish of the airlines for a rational government regulation
through an independent agency was granted through the Civil Aeronautics Act (Rhyne
1939):
Sec. 2. In the exercise and performance of its powers and duties under this act, the
Authority shall consider the following, among other things, as being in the public interest,
and in accordance with the public convenience and necessity:
a. The encouragement and development of an air transportation system properly
adapted to the present and future needs of the foreign and domestic commerce of
the United States, of the Postal Service and of the national defense;
b. The regulation of air transportation in such manner as to recognize and
145
preserve the inherent advantages of, assure the highest degree of safety in and
foster sound economic conditions in, such transportation and improve the
relations between, and coordinate transportation by air carriers;
c. The promotion of adequate, economical and efficient service by air carriers at
reasonable charges, without unjust discrimination, undue preferences or
advantages, or unfair or destructive competitive practices;
d. Competition to the extent necessary to assure the sound development of an air
transportation system properly adapted to the needs of foreign and domestic
commerce of the United States, of the Postal Service, and of national defense;
e. The regulation of air commerce in such manner as the best promote its
development and safety; and
f. The encouragement and development of civil aeronautics.
The Civil Aeronautics Authority (CAA) was founded with the Civil Aeronautics Act.
CAA was empowered to regulate fares, mail rates, inter-line agreements, mergers, and
routes. CAA had a mission with a double edge. It needed to hold rates at reasonable
levels for the public but also strengthen the financially weak airline business to help
develop the commercial air transportation business. First, Air Safety Board was created to
investigate accidents. However, that function was also transferred to CAA in 1940. CAA
was then renamed Civil Aeronautics Board (CAB). The CAB was modeled after the
Interstate Commerce Commission so that public utility type of regulation could be
imposed and the airline industry would not be harmed because of “cut-throat”, wasteful,
destructive, excessive, unrestrained competition (Dempsey 1989, p.18).
There was one remaining factor before the commercial aviation could take off, the
World War II. Interestingly, just as aircraft helped efforts of war, warfare helped the
aircraft industry. In 1939, there were less than 300 air transport planes in the U.S., but by
146
1945, 50,000 planes were produced annually. While the U.S. focused on mass
production, the breakthrough innovations –radar and jet engines- took place in Europe.
5.1.2 The Jet (Post-war) Era
The first jet engine was designed by a British pilot in 1930; but the first to build
and test a jet plane were the Germans in 1939. However, it would take five more years
for them to perfect the design –too late to change the outcome of the war. Nevertheless,
the jet age had arrived. The U.S. built its first jet plane, the Bell P-59 in 1943. The
breakthroughs in military aircraft were eventually applied to commercial aircraft. For
example, Boeing remodeled and introduced its KC-135 jet tanker as the first U.S.
passenger jet, the 707 in 1958. The 707 had a capacity of 181 passengers and could fly
550 miles per hour. It burned kerosene, which cost half as much as the gasoline the
traditional planes were using at the time. Pan Am was the first customer of the legendary
Boeing 707. The same year the Federal Aviation Act was passed to monitor the growing
airline industry. With this Act, Federal Aviation Agency was founded to establish and run
the air traffic control system, and to monitor the safety of the overall flights (its name was
later changed to Federal Aviation Administration when the Department of Transportation
(DOT) was created in 1967). CAB still had authority over economic issues such as routes
and fares.
Boeing launched its first widebody jet the 747 in 1969. The 747 had a capacity of
450 passengers. Pan Am was the first customer. Douglas and Lockheed jumped on the
widebody jet bandwagon with the DC-10 and L1011 respectively. However, these planes
147
were smaller in size, seating about 250 passengers. The supersonic plane, the Concorde
flew the same year.
5.1.3 Deregulation Era
Deregulation will be the greatest thing to happen to the airlines since the
jet engine.
Richard Ferris, President, United Airlines (Peterson and Glab 1994, p.49)
Before deregulation the U.S. airline industry was run much like a public utility.
Civil Aeronautics Board (CAB) was in charge determining who would fly which routes
and how much they would charge for the service. The CAB would not permit most new
companies to fly. For example, it delayed its decision on a proposal to fly coast-to-coast
for less than half the going rate for eight years and then dismissed it (Peterson and Glab
1994). As a matter of fact, some of the practices in the regulated airline industry would be
considered illegal in other industries. For example, the CAB approved a war chest fund in
which the airlines agreed to support any airline that was suffering from a union strike. By
mid-70s the mutual aid had amounted to $350 million (Peterson and Glab 1994). In 1970,
American, TWA, and United jointly decided to cut the capacity on their coast-coast
services so that they could achieve higher load rates. There were no other competition on
coast-to-coast routes and this would constitute a deliberate antitrust case for any other
industry except the airlines. Instead, the CAB approved it and granted antitrust immunity
for one year (Peterson and Glab 1994, p.30). Similar agreements followed. Regulatory
structure enabled airlines to fly at half capacity to capture market share. Since carriers
148
could not compete on price, they were competing on non-price terms such as “sandwich
wars.” For example, Delta was alleging that Northeast airlines did not have their steaks
“cooked to order” as their advertisements stated (Peterson and Glab 1994, p.30). This
dispute continued until Delta’s acquisition of Northeast (Kuttner 1996). This kind of
subtle competition did not create winners but an industry with low profitability.
The notion of deregulation was becoming stronger as studies showed that
unregulated interstate fares in Texas and California were significantly lower than
intrastate flights after controlling for distance (Levine 1987). For example, Southwest
was able to avoid Federal regulations by servicing the Dallas-San Antonio-Houston
triangle, which enabled it to offer fares much cheaper than that of the incumbent airlines.
It was asserted that regulation gave consumers excessive service, but a lack of price
competition, and inflated fares (Dempsey 1989). Moreover, the increased use of the
widebody jets coincided with the OPEC oil embargo in 1973. The fuel prices and
inflation increased drastically. The result was increased capacity and cost but falling
airline traffic. In order to assure a reasonable rate of return for the airlines, the CAB
allowed airlines to increase fares. It also decided to not approve any new service on any
route for four years and limited the overall capacity on major routes. However, the poor
performance of the airline industry continued despite the fact that it cost more for the
public to fly. The industry became a target for the Ford Administration, which was after
regulatory reforms. Hearings of the Senate Subcommittee on Administrative Practice and
Procedure, chaired by Senator Edward Kennedy, concluded that airline prices would fall
automatically at the absence of government-imposed limits (1977). The major airlines
had not taken the hearings seriously and had sent their junior executives. On the other
149
hand, Kennedy had enlisted elite academics (e.g., Alfred Kahn of Cornell) and discounter
heroes such as Freddie Laker (of U.K.’s Laker Airways) (Peterson and Glab 1994).
Kennedy recognized the political importance of bringing the airfares down and it
economically seemed to make sense. Alfred Kahn had written two volumes entitled the
Economics of Regulation and had concluded that even imperfect competition was
preferable to inherently imperfect regulation (Kahn 1970). Deregulation would allow for
new and innovative services, increased productivity and efficiency resulting in higher
consumer welfare. The CAB admitted to similar conclusions (1975). The airline industry
“was naturally competitive, not monopolistic” therefore entry and price constraints could
no longer be justified. Led by John E. Robson, the CAB started to loosen its grip on its
own. Cornell University Economics Professor Alfred E. Kahn became the chairman of
the CAB in 1977. Kahn was not happy with the existing CAB regulation. He argued that
the system inflated fares, and caused misallocation of resources, carrier inefficiency,
excess capacity, and acute range of services and prices (Dempsey 1989, p.20).
Kahn actively participated in an effort for reform by effectively using the
preliminary results of flexible regulation under his leadership for convincing political
figures and media. Behind the scenes, United Airlines
6
, and Federal Express were also
supporting the idea of deregulation. Each trusted that the size of their fleet would give
them an edge over competition in an unregulated industry (Petzinger 1995).
The theory of contestable markets in which potential market entry would prevent
monopolies was going to be tested with deregulation. Kahn was confident that the results
6
Worldwide, only Aeroflot had a larger fleet than United Airlines at that time.
150
would benefit all constituents including consumers, all communities, and the airlines --its
employees, stockholders, and creditors (1978, p.8).
Kahn would be dubbed the father of the U.S. airline deregulation a year later in
1978. Deregulation first commenced on the cargo side of the business. The express
package delivery service experienced tremendous growth since deregulation. Fed-Ex,
UPS, DHL are all considered owners of major airlines today.
The Airline Deregulation Act:
The Airline Deregulation Act, a milestone for the airline industry, was approved
by Congress on October 24, 1978 and signed soon after by President Jimmy Carter.
Government Controls on domestic routes and schedules were discarded and the free
market economics took over. Congress had scheduled that route and rate regulations to be
phased out in four years. However, CAB moved very liberally with Kahn on board, and
all restrictions on routes were practically abolished within one year.
There was also much deregulation in the international arena. The U.S. had signed
45 “Open Skies” agreements by mid-2000. “Open Skies” agreements abolish limitations
on routes including those on capacity, frequency, and provide flexibility for pricing,
charters, cooperative marketing agreements, and other joint operations. The DOT
continues to monitor the remaining regulated international routes.
After deregulation was successfully completed, the CAB was abolished at the end
of 1984 and most of its remaining functions were transferred to the DOT. Among the
most important of these inherited functions was to review and grant antitrust immunity to
merger and acquisition activities. However, the role of the government on safety has not
151
been abandoned, and the FAA continues to regulate safety issues. Deregulation sparked a
wave of change in many aspects of the airline business including many that were not
forecasted. For example, the rise and fortification of the hub-and-spoke networks, more
complex fare structures, and the survival of major firms at the expense of entrants were
all unforeseen (2001g). These changes are discussed next.
5.1.4 Consequences
Transportation was the first industry in the nation to be regulated, and the first to
enjoy significant deregulation (Dempsey 1989). While the evaluation regarding the
success of deregulation has been mixed, the evidence has been encouraging. Yet the
experience was far from perfect, and there were criticisms:
Of the six intellectual assumptions behind the airline deregulation, four
have been proven completely false. Deregulators believed that airline size
was not critical to efficient operations. The marketplace, to the contrary,
has ruled that bigger is better. Deregulators believed that barriers to entry
are low in the airline business. Experience has demonstrated that they are
very high. Deregulators believed that increased competition would
produce low unrestricted fares. In fact, it has produced a bewildering
array of discriminatory prices. Deregulators believed that travel agencies
were obsolete as well as potentially misleading channels of information
and distribution. But travel agencies became more powerful than ever. A
fifth assumption, that antitrust laws would restrain competitive abuses,
has been negated by the policy default of two administrations…(Kuttner
1996).
Despite the above criticisms “Deregulation has been one of the most successful
regulatory and economic policy reforms in the late 20
th
century” (Kasper 1998). Kahn
himself admitted that he had surprises in the deregulation process including the
turbulence and painfulness of the process, the reconcentration of the industry, the
intensification of price discrimination and the deterioration in quality of airline service
152
(Kahn 1988). He asserted that deregulation resulted in lower fares, increased range of
price-quality options and improvements in efficiency but that the competition was
unevenly distributed across markets and that congestion and delays were causing
problems (Kahn 1988). Several consequences of deregulation are investigated next.
5.1.4.1 Service and Quality
Serious concern about the decreasing level of safety and service quality and
increasing concentration levels after deregulation has been voiced (Dempsey 1989):
The industry rapidly became an oligopoly, with an unprecedented wave of
mergers consolidations, and bankruptcies. Today, the top 8 airlines
dominate more than 94% of the domestic passenger market… [F]unneling
of aircraft into “hub-and-choke” bottlenecks… have significantly
narrowed the margin of safety and sent the number of near misses
skyrocketing. Airline service has gone to hell in the 1980s. We are headed
aboard aerial slums, served cardboard food, overbooked, bumped, and
misconnected. Our luggage is routed through the Twilight Zone, never to
be seen during our natural lives… We can either spend an arm and a leg
or sleep in a strange city on a Saturday night.
The number of departures have increased by 50% for small community, 57% for
medium, and 68% for large community airports since deregulation (1996a). The DOT
indicated that the number of domestic passengers tripled since deregulation and that more
than 80% of domestic passengers enjoy two or more carriers alternatives (2001g).
However, Dempsey (2000a) points put that 61% of the non-hub communities suffered
from decreases in service. 28% of them lost all the service they had and only 6% enjoyed
new services (Dempsey 2000a). Meanwhile customer satisfaction was also deteriorating
(Dempsey 1989). The following were reported to be the top customer complaints ranked
in order:
153
Flight Problems: Cancellations, delays or other deviations,
Baggage handling: Claims for lost, damaged, or delayed baggage; charges for
excess baggage; carry-on problems; difficulties with claim procedures,
Refunds: Problems in obtaining refunds for unused or lost tickets or fare
adjustments,
Customer Service: Rude or unhelpful employees, unpleasant meals or cabin
service, and treatment of delayed passengers
Reservations, ticketing and boarding: Airline or travel agent mistakes in
reservations and ticketing; problems in making reservations and obtaining tickets
due to busy phone lines or waiting in line; delays in mailing tickets; and problems
boarding the aircraft,
Oversales: Bumping problems, whether or not the airline complied with DOT
oversale regulations,
Other: Cargo problems, security, airport facilities, claims for bodily injury, and
other miscellaneous problems,
Fares: Incorrect or incomplete information about fares, discount fare conditions
and availability, overcharges, fare increases, and the level of fares in general,
Smoking: Inadequate segregation of smokers, failure of the airline to enforce no-
smoking rules,
Advertising: Ads that are unfair, misleading, or offensive.
Source: (Coleman 1987) in (Dempsey 1989)
One of the major concerns about deregulation was that smaller communities
would lose air service after deregulation. Despite the ongoing program, (Dempsey 1989)
notes that many smaller communities have either lost all air services or the services were
downgraded to commuter carriers and suffered from a sharp decrease in quality. Forty
percent of small communities have suffered from decreasing service and increasing
prices after deregulation (Moore 1986).
154
As a potential remedy, the DOT administers the Essential Air Service Program,
which was launched with deregulation. Essential Air Service Program was designed to
provide subsidies to carriers to fly to certain smaller communities to which service would
be not feasible otherwise. Section 419 of the Federal Aviation Act ensures that smaller
communities remain linked to the national aviation system. This program was initially
approved by the Congress for a period of ten years (expiring in 1988) but was later
extended for another ten years (until 1998). Seventy-eight communities were still being
subsidized under this program as of May 1998 (1998d). In 1998, the end date for the
program was also abolished and its annual budget was increased and set to $50 million
with Rural Air Service Survival Act. The program continues to this day. Some 113
communities were being subsidized as of October 2001 (2001b; 2001j). The carriers
under the Essential Air Service program are usually assigned for a period of two years.
The following is expected from the basic essential air service:
(a) service to a hub airport, defined as an FAA-designated medium- or large-hub
airport,
(b) service with no more than one intermediate stop to the hub,
(c) service with aircraft having at least 15 passenger seats at communities that
averaged more than 11 passenger enplanements a day in any calendar year from 1976-
1986,
(d) under certain circumstances, service with pressurized aircraft, and
(e) flights at reasonable times taking into account the needs of passengers with
connecting flights (1998d).
Currently, communities are eligible to be a part of this program if they are further
than 70 driving miles of an FAA-designated Large or Medium Hub airport, unless the
subsidy per passenger exceeds $200 (with certain exception for communities that are
further than 210 highway miles from the nearest Medium or Large Hub) (2001f).
155
In conclusion, it can be said that while service in hub markets increased in
frequency, some community markets suffered from a loss of service as a result of
deregulation. While the range of the quality of services (e.g., first class, business, coach)
increased, it can be argued that the average service level also decreased. However, this
issue should be taken into consideration along with the fare levels, which generally were
reduced due to heavy competition following deregulation.
5.1.4.2 Fare Levels
Numerous low cost carriers (LCCs) challenged the major carriers following
deregulation. Major carriers
7
responded by price cuts, and airfares (adjusted for inflation)
fell by 33% from 1976 (the dawn of deregulation) to 1993 (Morrison and Winston 1995).
Thus, benefits of new entrants included lower fares and increased frequency. Dropping
fares ignited demand: “For example, when AirTran entered the Atlanta-Buffalo market,
average fares declined by 36%, from $185 to $119, and the number of the passengers in
the market increased by 65%, from 23,000 per month to 38,000 per month. Similarly,
when Vanguard re-entered the Kansas City-Minneapolis market in late 1996, average
fares declined by 59%, from $246 to $101, and the traffic more than doubled, increasing
from about 12,000 passengers per month to 25,000 per month” (2001g, p.6). The DOT
estimated that consumers saved $6.3 billion per year due to low fare carriers (1996c).
“Most upstarts have average seat mile costs in the 7 cent range, compared to the 10 cent
7
A major airline is defined by the DOT as airlines with annual operating revenues of over $ 1,000,000,000.
(Oster and Strong 2001) There were 12 major U.S. passenger airlines in 2000: Alaska, America West,
American, American Eagle, American Trans Air, Continental, Delta, Northwest, Southwest, Trans World,
United and US Airways. In addition, three all-cargo airlines were classified as majors: DHL Airways,
FedEx and United Parcel Service.
156
range of traditional carriers. Upstarts maintain their low costs by a number of strategies,
including lower labor costs, direct marketing and “no frills” service” (Fones 1997). Thus,
discount airlines have approximately 30% cost advantage over network carriers.
However, the increase of concentration in hub markets limited the extent of
benefits that discount carriers could offer. The average number of carriers per route was
2.2 (most routes are served by major carriers only) and those routes that Southwest flies
were 47.2% cheaper than comparable routes (Kuttner 2000). DOT reported that
passengers paid $54 less, on average, if a low cost carrier served the same market
(Dempsey 2000a). On the other hand, while the public has been getting better deals, it
was reported that unrestricted fares (i.e., fares that mainly business class passengers pay)
have increased by 73% after deregulation (Dempsey 2000a, p.485).
157
Figure 5.1: Demonstration of fare level differences 1998Q2
Note: SIFL stands for Standard Industry Fare Level. It is the pre-deregulation fare level (adjusted
for inflation). Source: Domestic Airline Fares Consumer Report 1998 Q2 (1998a)
158
Figure 5.1 powerfully demonstrates the impact of low fares carriers and
deregulation has had on fare levels. The 100% base line represents the prices in 1978
(pre-deregulation) adjusted for inflation.
The implication is that the overall fare levels are indeed lower than they would
have been under regulated pricing for distance blocks longer than 750 miles. The prices
are lower than the regulation era for low-fare carrier markets for all distance blocks.
Finally, the major carriers’ fares are lower than regulation era only for distance blocks
longer than 1500 miles. The figure applies to the second quarter of 1998 but the
implications can be generalized to other time frames. For example, Figure 5.2 presents
the same notion for the third quarter of 2001.
Figure 5.2: Demonstration of fare level differences: Top 1000 markets, 2001 Q3
Note: SIFL stands Standard Industry Fare Level. It is the pre-deregulation fare level
(adjusted for inflation). Source: Domestic Airline Fares Consumer Report 2001 Q3
(2001d)
159
5.1.4.3 Market Concentration and Power
Despite the early success, ten years into deregulation, the increasing levels of
concentration were causing concern. The DOT observed that the number of carriers had
increased from 39 to 131 from 1978 to 1987 (1987a). However, this observation was
misleading since “nearly two-thirds of city-pairs were airline monopolies and another
20% were duopolies” (Dempsey 1989, p.87). “The 11 major airlines have shrunk to eight;
the eight local former local service carriers are now two and they are trying to merge; the
eight original low-cost charter airlines have been reduced to one, through bankruptcy and
abandonment; 14 former regional airlines have shrunk to only four; over 100 new upstart
airlines were certified by the CAB and about 32 got off the ground and most of these
crashed, leaving only a handful still operating; of the 50 top commuters in existence in
1978, 29 have disappeared… Today, the top 50 carriers who constitute 90 percent of that
industry are captives of the major airlines and relegated to serving the big airlines at their
hubs” (1987b, p.61-62).
A major reason for increased market concentration was the DOT’s relaxed merger
evaluations in the 80’s. Particularly in the second half of the 80s, the DOT approved a
number of mergers based on the notion that there were no barriers to entry even tough the
mergers resulted in dominant markets shares. The analysis overlooked the issue of
whether or not those potential entries would be economically feasible (Nannes 1999).
Despite the warnings in the Airline Deregulation Act of 1978 against “unreasonable
industry concentration, excessive market domination”, the DOT “never met a merger it
didn’t like” (Dempsey 1989, p.87):
DOT approved them all. It approved Texas Air’s (i.e. Continental and
New York Air) acquisition of both People Express (which included
160
Frontier) and Eastern Airlines (which included Braniff’s Latin American
routes); United acquisition of Pan Am’s transpacific routes; American’s
acquisition of AirCal; Delta’s acquisition of Western; Northwest’s
acquisition of Republic: TWA’s acquisition of Ozark; and US Air’s
acquisition of PSA and Piedmont, to mention only a few. This has sharply
increased national levels of concentration to the point that the eight
largest carriers control over 94% of the domestic passenger
market…Under deregulation, [charter flights] virtually vanished
(Dempsey 1989, p.88)
A critical review can indeed establish that the mergers in the Reagan era
started an irreversible process and were an important reason of the increase in
market concentration in the airline industry. Table 5.2 demonstrates that twelve
mergers took place in a matter of two years:
Table 5.2: Airline Mergers in the 1985-87 Period
Acquiring
Airline
Passengers
(thousands)
Acquired Airline Passengers
(thousands)
Final Bid
Southwest 10698 Muse 1980 March 11, 1985
Piedmont 14274 Empire 1084 October 3, 1985
People 9100 Frontier 7068 October 9, 1985
Northwest 14539 Republic 17465 January 24, 1986
Texas 19640 Eastern 41662 February 24, 1986
TWA 20876 Ozark 5541 February 28, 1986
Alaska 3132 Jet America 774 September 8, 1986
Delta 39804 Western 9062 September 10, 1986
Texas 19640 People 11907 September 16, 1986
American 41165 Air Cal 4451 November18, 1986
Alaska 3132 Horizon 942 November20, 1986
US Air 19278 Pacific Southwest 9049 December 9, 1986
Us Air 21725 Piedmont 22800 March, 1987
Braniff 2557 Florida Express 1415 December 15, 1987
Source: DOT Air Carrier Traffic Statistics (Oster and Strong 2001, p.7).
Other major mergers since deregulation have included (1993a, p.459):
American: TWA, Air Cal, Eastern (Latin America), Reno
United: Pan Am (transpacific), Pan Am (Latin America), Pan Am (Heathrow)
161
Delta: Pan Am (Europe)
Continental: Texas International, Frontier, New York Air, Rocky Mountain, Britt, PBA
Pan Am: Pan American World, National, Ransome
Eastern: Braniff (Latin America)
Republic: North Central, Southern, Hughes Airwest
US Airways: US Air (Allegheny), PSA, Empire, Henson
Southwest: Morris Air
TWA: TWA, Ozark
Former CAB chairman Alfred Kahn heavily criticized DOT for their hasty
approval of the mergers: “It is absurd to blame deregulation for this abysmal
dereliction.”(Kahn 1988). Empirical research has also indicated that the airline
consolidation has led to market power (Kim and Singal 1993). The ten largest airlines
accumulated 88% of the revenue passenger miles flown; that figure had risen to 94% by
1990 (Sheehan 1993). General Accounting Office reported that consumers flying from
small to major airports had to pay 34% more if the major airport was concentrated and
42% more if both airports were concentrated (Dempsey 2000a, p.485). However,
mergers were probably not the primary reason for the increased concentration in the 80’s.
Market concentration and the resulting power was a consequence of the rise of the hub-
and-spoke system.
5.1.4.4 Hub-and-Spoke Networks
I never heard a word spoken about hub and spoke in the entire debate
leading up to deregulation.
Congressman James Oberstar, Ranking Minority Member, U.S. House of
Representatives Aviation Subcommittee (Dempsey 2000a)
One of the unforeseen results of airline deregulation was the emergence of the
hub-and-spoke networks. The hub and spoke concept lies at the heart of any network
162
carrier. Securing hubs enables the dominant carrier to charge “hub premiums” (Hecker
2001). The following statistics should clarify their importance: U.S. General Accounting
Office (GAO) compared prices at concentrated hub airports and relatively unconcentrated
airports, and found that prices were 27% higher in the concentrated hubs (Dempsey
2000b). Adjusting for the average trip distance and the size of the markets, the
concentrated hub fares were on average 18.7% higher than for similar markets in other
airports. In the absence of LCC competition, the major carrier was able to charge fares
that exceeded its fares in non-hub markets of comparable distance and density by
upwards of 40% (Dempsey 2000b).
The incumbent hub carrier has an advantage over their competitors because they
are able offer a wider range of flights and services. In effect, they are able to attract a
larger portion of business travelers and obtain a higher yield than their rivals (2001g).
There is also higher brand recognition and the advertising costs are spread across more
markets than that of the rivals (Levine 1987). Robert Crandall, CEO of American airlines
summarized the benefits of the system as follows:
While a hub and-spoke system is admittedly more expensive to operate
than a comparably sized system of point-to-point routes, the system’s
incremental costs are more than offset by its enormous revenue benefits.
For example, we estimate that there are fewer than 500 city pair markets
in the United States big enough to adequately support point-to-point jet
service. However, our hub-and-spoke system makes it possible for
American to effectively serve over 10,000 markets –and realize a large
revenue per available seat mile premium relative to point-to-point
carriers. (1993a, p.3)
However, the efficiency of using hub and spoke networks for short-haul flights is
questionable. For example, Southwest has an average of 20.4 minutes ground time as
163
opposed to American’s 50.3 minutes (Dempsey 2000a). The result is a 22% better
utilization of aircraft for Southwest plus additional gains from personnel productivity
(Dempsey 2000a). The hub-and-spoke system also led to the use of smaller aircraft,
which meant relatively poor seat-mile cost efficiency. The same pattern of increasing
concentration of fortress hubs has been observed over and over again (Dempsey 1990;
1997). Table 5.3 demonstrates the increase of market share by hub carriers particularly
during the decade following deregulation. These figures continued to be alarmingly high
in the 90’s.
Table 5.3: Single Carrier Market Shares at Major Airports Pre and Post Deregulation
Airport 1977 1987 1997*
Baltimore/Washington 24.5% US Air 60.0% US Air 33.7% US Air
Cincinnati 35.0% Delta 67.6% Delta 79.4% Delta
Detroit Metropolitan 21.2% Delta 64.9% Northwest 63.4% NW
Houston Intercontinental 20.4% Continental 71.5% Continental 39.1% CO
Memphis 40.2% Delta 86.7% Northwest 52.2% NW
Minneapolis/St. Paul 45.9% Northwest 81.6% Northwest 69.7% NW
Nashville Metropolitan 28.2% American 60.2%American 29.3%SW
Pittsburgh 43.7% US Air 82.8% US Air 73.3% US
St. Louis – Lambert 39.1% TWA 82.3% TWA 48.7% TWA
Salt Lake City 39.6% Western 74.5% Delta 55.5% Delta
Average 33.8% 73.2% 54.4%
Source: Extended from Consumer Reports (June 1988), at 362-367 (Dempsey 1989, p.89)
* 1997 figures were retrieved from the DBP database and may not be fully compatible
with earlier figures. The loss of market share at the hubs were typically due to Southwest
competition.
Hubs enable the incumbent to have more frequent service on the spokes than
would otherwise be economically feasible (Levine 1987). Therefore, they are generally
beneficial for the public, however, their disadvantages include increased barriers to entry
(2001g; Anderson 1997). The resulting market power may also lead to supra-competitive
164
pricing (i.e., higher prices than would be considered normal for the market) and decrease
the welfare of consumers especially for those living in hub markets (Brown 1991). A
major concern regarding competition was that the major carriers stifled competition in
their hub airports. The incumbents carry the majority passengers in and out of their hub
markets (2001g). “The hub carrier dominates city pairs it serves directly from its hub,
except to other cities that are also hubs for other carriers, in which case the two carriers
providing hub service dominate. Entry by a major carrier on a point-to-point basis into
another carrier’s hub has become very much the exception” (Nannes 1999, p. 4). One
2001 DOT report concluded that remaining passengers pay 41% more than those flying
in hubs with low-cost carrier presence (2001e). The situation in short-haul hub markets
was even worse, with its 54% price premium over comparable routes with low fare
competition.
5.1.4.5 Hub Premiums
Federal Aviation Administration categorizes airports into four categories: large
hubs, medium hubs, small hubs and non-hubs (Morrison and Winston 1997):
Categorization Total U.S. Traffic Accounted by Hub
Large Hubs 1% or more
Medium Hubs 0.25% to 0.99%
Small Hubs 0.05% to 0.24%
Non-hubs less than 0.05%
When the fare levels are compared to non-hubs, fare premium effects are robust
and present. A hub premium is the increased fares that the consumers have to pay for
flying from a certain hub and is defined in percentages. Premiums have been observed to
165
be 62% for Charlotte, 51% for Cincinnati and Pittsburgh and 41% for Minneapolis/St.
Paul (Oster and Strong 1996). These premium effects were “persistent” over time
(2001g).
Interestingly, hub premiums are either mild or non-existing for hubs that are
served by Southwest (hence, the phrase “Southwest effect”) (2001g). For example, one
year after Southwest began serving Providence, RI markets, fares fell by almost 50% and
traffic more than tripled (Slater 2001). In summary, hub-and spoke networks enable the
use of market power and act as barriers to entry (Leigh 1990). Table 5.4 demonstrates
the consistent increase of hub premiums for the leading hub airports in the nation on an
airline basis. For example, it shows that consumers in Atlanta have had to pay 20-40%
higher fares for the privilege of flying out of Hartsfield airport in comparison to the rest
of the nation for the same distances.
Table 5.4: Changes in Hub Premiums over Time
Source: (Oster and Strong 2001, p.33)
166
The trends in Tables 5.4 and 5.5 indicate that the hub premium effect is persistent
over time. It should be noted that the occasional negative premiums may be either due to
presence of discounters or due to low entry barriers and limit pricing (e.g., St. Louis,
Baltimore).
Examining the HHI index at the national level can be misleading. A more
accurate assessment would be possible through an examination at the hub airport level. It
was found that the concentration for the airline industry had increased to an alarming
3877 (above 1800 is considered highly concentrated by the DOJ) and thirty-three airports
were assessed to be highly concentrated (1997a). It should be noted that only La Guardia
and Los Angeles International airports are below the 1800 threshold for high
concentration. There has been a general trend for more concentration fueled by the
mergers as explained previously.
Table 5.5: Sample HHI Index for the 10 largest US Airports 1985-1996
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
Chicago O’Hare 2449 2836 3263 3491 3526 3591 3617 3427 3557 3474 3505 3445
Atlanta 4190 4468 4569 4619 5420 5489 7586 7604 7058 6223 5585 6233
Dallas/Fort Worth 4143 4500 4519 4759 4906 4927 4833 5008 4541 4451 4581 4684
Detroit 2189 2151 3903 3793 4655 5124 5557 5973 6188 5949 6240 6465
Los Angeles 852 911 1025 1066 1131 1156 1297 1376 1393 1379 1528 1524
Denver 2356 3015 3767 3674 3504 3562 3587 3661 3787 4307 4980 4892
Phoenix 1601 2000 2395 2459 2507 2700 2849 2679 2645 2556 2544 2451
San Francisco 1437 1723 1847 1715 1903 2134 2552 2685 3073 3448 3760 3831
Newark 2821 2395 2154 2400 2815 2933 3092 3143 3222 3287 2929 3060
NewYork
LaGuardia
1368 1384 1320 1175 990 1102 1293 1662 1763 1733 1688 1777
Source: DOT Form 11 Data (1997a)
Professor Allvine observed that the major airlines widely employ predatory
pricing in the name of defending their “fortress hubs”:
167
Government study after study shows that airline industry is not perfectly
competitive… These studies show that the major airlines employ many
monopolistic practices that contribute to the market power to raise and
maintain prices above the competitive level. In imperfect competitive
markets, it makes perfectly good sense for large firms to use predatory
pricing to destroy competition that threatens the monopoly prices charged
(cf. Dempsey 2000a, p.474).
5.1.5 Marketing
Deregulation has resulted in stiff competition for the airlines and marketing
efforts became key as airlines scrambled to differentiate themselves from competition.
The airlines spent in excess of $10 billion for domestic sales and promotion activities in
2000 which was almost twice their operating profits (2001c). The following section
investigates the impact deregulation had on important elements of airline marketing.
5.1.5.1 Yield Management
“It was a typical flight in the era of deregulation: United Airlines flight
815 from Chicago to Los Angeles, with 204 tickets sold at almost as many
prices.”
(the range of prices on that flight was from $87.21 to $1258.51)
NY Times Reporter Matthew Wald (Wald 1998)
Bob Crandall of American Airlines allegedly is the innovator of the airline yield
(a.k.a. revenue) management. Based on the concept that the cost of additional seats on
scheduled flight was negligible, he ordered his staff to study the demand and price
elasticity patterns from its computer reservations system “Sabre”. The result came in the
form of 35% “Super-Saver” discounts in 1977 all around the nation. Unable to compete
without a price advantage many charter services were soon bankrupt (Petzinger 1995).
168
American training manual stated the new objective “to sell the highest priced product that
the customer is willing to buy” (Peterson and Glab 1994, p.59). The minds behind
regulation and economists had anticipated that deregulation would lead to a simpler fare
structure (2001g). However, the yield management systems led the industry in the
opposite direction.
Yield management essentially uses the power of price to communicate with the
customer and to lure them from one service (own or competitors’) to another. The
demand for different products/services is monitored continuously and price and
promotional adjustments are made to maximize revenues, thus the profits of a company
through segmentation of the market. Many companies in other industries have also
realized the benefits of revenue management. Examples include the hotel industry,
railroads, telecommunications, and broadcasting. Certain conditions need to be satisfied
for revenue management to be useful (Daudel and Vialle 1994). The product should be
perishable, it should be possible to price target different customers. It should be possible
to sell the product in advance. The variable costs should be low. The demand should be
cyclical or it should vary so that it can be smoothed by revenue management. Typically,
segmentation by time is key. Needless to say, the airline industry is ideal for the
application of revenue management. Today, virtually every successful airline is using
yield management in some form.
5.1.5.2 Frequent Flyer Programs
Among the innovations that deregulation inspired, a most important marketing
tool became the frequent flier programs. In its purest form, the frequent flier programs
169
attempt to create brand loyalty by inducing a passenger to fly on only the owner of the
program or one of its affiliates (through code-sharing). As with yield management, the
Frequent Flier Miles concept was also innovated by American Airlines (Peterson and
Glab 1994, p.60) in May 1, 1981. Five days later United Airlines responded by
introducing Mileage Plus (Woodyard 2001). All major airlines developed their own
programs shortly thereafter.
Table 5.6 demonstrates the scope of the frequent flier programs in the U.S.. For
example, American’s AAdvantage program has induced some 35 million members
resulting in 2.3 million travel awards redeemed during 1998.
Table 5.6: A Comparison of Major Frequent Flier Programs
Airline Members
(million)
Number of
Awards
Accumulated
Number of travel
awards redeemed in
1998
American Aadvantage 35 NA 2.3 million
US Airways Dividend Miles 20 4.4 million 900K
Continental OnePass 16 NA 1 million
Northwest WorldPerks 18.5 6.1 million 1.2 million
United MileagePlus 27 6.1 million 2.1 million
America West FlightFund 2.9 NA NA
Alaska Mileage Plan 3 812K 191K
TWA Aviators 12 1.1 million NA
Delta Skymiles 24 9.6 million 1.9 million
Southwest Rapid Rewards 1.2 NA 927K
Source: InsideFlyer (Stoller 1999)
Frequent flier programs have advanced into more sophisticated forms and
consumers can earn mileage points through several means including credit card purchases
or long-distance phone calls. As perfectly acceptable as they are as a marketing tool, a
potential anticompetitive problem with the frequent flier programs is that they can be
effectively used to target new entrants. Additional bonus miles on certain routes can
170
effectively bring down the fare without being detected and then be revoked (Oster and
Strong 2001). It has been argued that having a reward program tied in with an excellent
customer service will help take the customer’s eye off the price (Mohs 1999).
A current snapshot of the frequent flier campaigns is presented below
8
:
• There are more than 120 million members worldwide, 74 million members of whom
are from the U.S.
• 27-28% of the members are active.
• AAdvantage of American Airlines remains as the largest frequent flier program with
more than 45 million members. On average, more than 11 thousand new members
enrolled in AAdvantage per day in 2001.
• The programs grew by 11% on average with the fastest growing segment being
“mileage consumers.”
• Approximately 40% of the miles earned are not from flying and credit cards are the
most popular form of earning such miles.
• An award is estimated to cost airlines $13.93 on average.
• 14 million free tickets were awarded in 2001.
• 82-87% of members have web access.
5.1.5.3 Distribution: The Rise (and Fall) of Travel Agents
There is no doubt that the travel agencies grew in importance after deregulation.
As the choice of flights available to the public increased in number, the influence that the
agents had in the decision making also increased. The airlines tried to get the leading
agents on their side by use of commissions. Two main conflicting forces are at work
when the motives of travel agents are considered. On one side is the issue of the override
commission and the percentage baseline commission, which motivates the agent to
influence the customer to buy the most expensive ticket from the airline that offers the
8
Source:http://www.webflyer.com/company/press_room/facts_and_stats/frequent_flyer_facts.php
171
best override commission, on the other hand is the need to keep their customers satisfied
and happy to succeed in the long run. Travel Agent Commission Overrides (TACOs) are
special bonus commissions paid to travel agents by a specific airline for meeting a
targeted proportion or number of passengers booked (Oster and Strong 2001). It appears
that this scenario was more beneficial to major airlines who offered better commission
rates than low-cost carriers. Another public policy concern is that travel commission
overrides were typically designed to be in favor of the carrier with the largest market
share (1996a).
The inability of the new entrants to cope with commission overrides (TACOs)
was a main reason why they exited certain markets. Southwest airlines’ decision to pull
out of Indianapolis – Detroit (one of the rare exit events for Southwest) and Midwest
Express’ exit from Milwaukee-Detroit and other markets have been linked to commission
overrides (Oster and Strong 2001). The largest travel agencies also admitted that
overrides had an important effect on the booking patterns (1996a). The sophistication in
the computer reservation systems also fueled the need for agents.
However, the growth of discounting and the explosion of the internet as a viable
distribution medium drastically changed the equilibrium as the major airlines were being
forced to cut their commissions due to increasing cost pressures. Namely, Delta,
American and Continental decided to drop the commissions altogether and others
followed suit (Brannigan and Stringer 2002). Some major airlines have announced that
they will start for charging extra for paper tickets. Major airlines also realize that selling a
ticket through their own web site costs them 25% of what it cost through a travel agent,
172
therefore the travel agencies who typically sold 80% of the tickets face a gloomy future
(Fonti 1999). This trend can be readily observed in Figure 5.3.
Figure 5.3: US Travel Agency Overall Airline Commission Rates and the Number of
Agency Offices, 1976-2000
Source: Harris / Travel Weekly, 2000 Travel Agent Survey
5.1.5.3.1 Computer Reservation Systems (CRS)
The CRS systems had tremendous impact on the development of complex pricing
strategies and revenue management systems. These systems enabled the travel agents to
track fare and capacity changes instantaneously. Although several competing systems
were initially introduced (American’s Sabre vs. United’s Apollo), many of these systems
(as well as the largest travel agencies) later merged. Some airlines chose to join existing
systems and pay fees for listings rather than develop their own systems. Table 5.7
presents the travel agents’ market share of competing systems:
173
Table 5.7: 1999 Market Share for CRS
CRS Market
Share
Sabre 34%
WorldSpan 24%
Apollo 24%
System One 20%
Note: Some travel agents use more than one system
Source: Harris / Travel Weekly 2000 Travel Agent Survey
The DOJ actually sued the airlines and owners of the systems because of
signaling and price fixing (1994b). That the systems were designed to give their owners
certain (anti-)competitive advantages was a concern:
An airline whose CRS is used by travel agents has access to a very
accurate picture of both its own and its rivals’ business patterns. Through
the CRS an airline can track the effect of price changes, see roughly how
much of a rival’s seat inventory is assigned to a given discount fare
classification, measure how much full-fare business it attracts compared
to rivals, and track changes in city-pair flows… It can even see how loyal
its own frequent flyers are. A CRS owner can then use this information to
distort market signals to its rivals, leading them to make incorrect
decisions. When a CRS owner sees travel agents making bookings on a
rival airline’s flights, it can intervene through targeted incentive programs
in an attempt to switch business. By responding selectively, it can
temporarily distort signals the market sends to competitors, in order to
persuade the rivals to abandon fares, schedules, or even routes where,
absent these secret interventions, its offerings would be preferred by
customers (Kuttner 1996, p.261).
Computer Reservation Systems (CRS) provide much of the information needed
for a competitive response. Competitors, prices, capacity and availability can be observed
through these systems. The incumbents can even gather the scope of new entry because
the schedule and fares are filed before they are put in effect (Oster and Strong 2001).
Therefore, the incumbent has much of the knowledge it needs for a competitive response
174
decision. It can target its direct (e.g., price, capacity on route) and indirect predatory
weapons (e.g., frequent flier miles, commission overrides) discussed earlier to coordinate
a successful campaign against the entrant (Oster and Strong 2001). Since these efforts can
be highly targeted, they do no not signal hostility to provoke retaliatory response from
other major carriers.
5.1.5.4 Discounting and Competition
“Today, one of every seven domestic passengers is flying because of the
increased competitiveness resulting from low fare service.” (1996c)
In April 1996, the DOT released a report full of hope for the airline industry. Its
title read “The Low Cost Airline Service Revolution.”(1996c). One conclusion of the
report was that the consumers heavily benefited from the efficiency and competition that
the low cost carriers (LCCs) were bringing in the airline industry. It was also suggested
that there was evidence that network (major) carriers and low cost carriers could co-exist.
The fundamental cost advantages that low cost carriers have over network carriers were
emphasized and the global implications were discussed. Other positive implications
included those on industry labor force (e.g., union relations) and economic growth and
benefits to consumers, communities, travel related industries and the aerospace industry
in general. The report also included a warning that the premiums at network hubs where
there was no low-cost competition were high and increasing.
It can be argued that the discounter LCCs entered the market in two waves. The
first wave emerged right after deregulation in 1978 and continued roughly until 1982.
Even though these discounters helped reshape the industry and brought many benefits to
175
consumers, the pre-deregulation incumbents did not let the LCCs steal away their
customers. They responded with price discrimination utilizing sophisticated revenue
management techniques. Some of their deep-discount fares were coded “FU” and the
signal to the new entrants could not get clearer than that! (Kuttner 1996). As a result
many LCCs went out of business before they got a fair chance to compete.
The classic example of a discounter during the first wave was undoubtedly People
Express. People Express was founded in April 1981. Don Burr’s strategy was simple:
very low operating costs and offering flight experience to the masses (Peterson and Glab
1994). The service might not be extensive but it would be warm (Petzinger 1996). Every
employee was required to have a second job and many employees (e.g., the CFO) served
as a flight attendant in their second jobs. Don Burr was a charismatic leader and the
employees embraced his style of his leadership. He announced the precepts –the code of
behavior for every employee of People Express:
One: Service –Commitment to the growth and development of our people.
Two: To be the best provider of air transportation.
Three: The provide the highest quality of leadership.
Four: To serve as a role model for others.
Five: Simplicity.
Six: Maximization of profits. (Petzinger 1996, p.134)
As the most successful airline launch ever, the company grew from 3 to 17 planes
and was flying to 17 destinations out of Newark by the end of the same year. In five
years, it had 217 planes operating in domestic and international routes (Pearson 1996).
Their target market was the middle-class, working class, and students. “At one point it
was cheaper to fly to Florida than to take a bus or drive” (1994a, p.35; Pearson 1996).
The prices were fixed and you bought tickets as you boarded the plane, on a first come,
176
first serve basis. However, unable to respond to fierce price competition and revenue
management by American, it was already bankrupt in 1986. In fact, the same fate was
typical of the LCC of the era, among the 58 start-ups launched between 1978 and 1990,
only America West and Midwest Express have survived (McBride 1999).
The second wave of LCC entries happened after the industry once again became
profitable in 1993 and continued until the ValueJet crash in 1996. Ironically, the typical
example of a second wave entry would be ValuJet which started its operations in Atlanta
in 1993 and had grown from 5 to 30 cities until the crash. Second wave entries were also
effectively stopped by the major players because they were already fortified in their hubs
and were not about to give up the premiums. The typical competition pattern, also called
the “homicidal cycle” (Dempsey 2000a; Dempsey 2000b) was observed in the airline
industry in these five steps:
1. Incumbent (major carrier) monopolizes a hub and raises prices to supra-
competitive levels.
2. Low Cost Carrier (discounter) enters the market with low fares.
3. Incumbent responds by matching fares regardless of its cost structure, and often
couples the response with increased capacity and/or frequency. Incumbent also
uses TACOs (commission overrides) and other tactics to drive the entrant out
before it can establish a foothold in the market.
4. LCC, not having the deep pockets as the incumbent is forced to withdraw.
5. Incumbent increases price to previous or higher levels and decreases capacity.
As the same pattern was observed over and over again, the reputation effects
become more than a theory and new entry to markets were stifled. New market
entries diminished in the nineties (Dresner et al. 2001) and there were no new entry
applications to DOT during 1997 and 1998 (see Figures 5.4 and 5.5 below). However,
the tremendous effect discounting has had on fares levels is probably most apparent
from reviewing actual figures.
177
0
1
2
3
4
5
6
1992 1993 1994 1995 1996 1997 1998
New Entry Applications
Figure 5.4: New Entry Applications
Data Source: (1999b)
Market Entry Patterns in the Airline Industry
0
1
2
3
4
5
6
7
8
9
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
Figure 5.5: Entry/Exit Patterns in the Airline Industry
Note: The bottom (blue) stack represents new entering and remaining airlines for
the year, the top (red) stack represents the entry and exits in the same year.
178
Table 5.8: Average Fares in Markets With and Without Low-Fare Competition
Market Type 1994 1995 1996 1997
Without Low-Fare Competition $168 $183 $177 $180
With Low-Fare Competition $86 $95 $91 $100
% Difference 97% 92% 93% 80%
Source: Domestic Airline Fares Consumer Report 1997 (Q4)
Low-fare carriers usually do not fly longer distances so it may be argued that a
comparison between long haul and short haul market fares may not be justified.
However, a deeper look in any given year indicates that the benefits of low-fare
competition exist over the range of flights:
Table 5.9: Average Fares Sorted by Distance
Mileage Block Market Type
0-250 251-
500
501-
750
751-
1000
1001-
1500
1501-
2000
Over
2000
Without Low-Fare
Competition
$142 $165 $171 $167 $179 $198 $223
With Low-Fare
Competition
$67 $75 $101 $121 $133 $163 $176
% Difference 111% 121% 70% 38% 35% 22% 27%
Source: Domestic Airline Fares Consumer Report 1997 (Q4) (1997c)
The above Tables 5.8 and 5.9 demonstrate that discounting has a robust negative
impact on the prices. Focus on specific markets is also telling in that it reveals the
competitive impact on prices and passenger levels.
179
Case in point: Air-Tran Entry and Exits
Table 5.10: Atlanta Markets with AirTran Entry
Atlanta to:
Passengers per
day
Average Fare
97 Q3 98Q3
Change in
Passengers
97 Q3 98 Q3
Change
in
Average
Fare
Buffalo, NY 269 443 65% $169 $119 -30%
Dayton, OH 261 521 100% $178 $116 -35%
Greensboro,
NC
310 575 85% $229 $113 -51%
Hartford, CT 514 885 72% $242 $124 -49%
Houston, TX 974 1409 45% $211 $130 -38%
Richmond, VA 426 700 64% $223 $118 -47%
Group
Average
459 756 65% $213 $122 -43%
Table 5.11: Atlanta Markets with AirTran Exit
Passengers per
day
Average Fare
Atlanta to:
97 Q3 98Q3
Change in
Passengers
97 Q3 98 Q3
Change
in
Average
Fare
Charlotte, NC 875 623 -29% $110 $181 65%
Columbus, OH 728 546 -25% $114 $182 60%
Louisville, KY 589 391 -34% $101 $188 86%
Group
Average
730 520 -29% $108 $184 70%
Source: Domestic Airline Fares Consumer Report 1998 (Fourth Quarter); (2001e)
LCCs grew rapidly over time in spite of anti-competitive practices in the industry.
Their market share grew to almost 15% in 1999 from 8.8% in 1986 (2001a). This growth
is also evident in Table 5.12 below:
180
Table 5.12: Airports with Most U.S. Destinations Served by Low Fare Airlines
Low Fare Services 1992 2002
Phoenix 48 84
Las Vegas 33 59
Chicago Midway 9 54
Orlando 5 40
Atlanta 2 39
Baltimore 3 38
Kansas City, Mo. 6 35
Denver 2 33
Tampa 2 32
Nashville 4 28
Houston 14 25
Los Angeles 6 25
Fort Lauderdale 1 23
Albuquerque 11 22
Oakland 7 22
St. Louis 13 20
New Orleans 3 19
New York JFK 3 19
Austin 7 17
Columbus, Ohio 11 17
Seattle/Tacoma 3 17
Indianapolis 6 16
Salt Lake City 2 15
San Diego 6 15
Detroit 2 14
Portland, Oregon 3 14
Dallas Love Field 13 13
El Paso 12 13
San Antonio 6 13
Birmingham 2 12
Dallas/Fort Worth 2 12
San Jose 1 12
Source: (Adams 2002)
181
5.1.5.5 Price Wars
“I knew the industry would be ruggedly competitive, but I did not expect
its leaders to engage in extended kamikaze behavior.”
Warren Buffet regretting his investment in USAir in 1989 to his
shareholders (Labich 1993).
The competition between the major airlines seems to have stagnated during the
last ten years or so. However, this was not always the case, and irrational price wars have
also taken place in the colorful history of the airlines. Dempsey (2000b) describes one
such war between American and Northwest. In April 1992, American Airlines launched
its “value pricing” campaign. They had concluded that the price structures had become
too complex for the consumer to comprehend. Their new value pricing had only four
levels (first class, regular coach, 14-day advance purchase, 21-day advance purchase).
The unrestricted coach fares went down by almost 40% on average. The Economist
argued that American was pricing below operating costs and could cause more chapter 11
files in the industry (1992a). American anticipated that the newly stimulated demand
would more than offset the reduction in prices. However in May 1992, Northwest
announced its “Grown-Ups Fly Free” campaign (i.e., an adult flying with a fare paying
child would fly for free) to steal the thunder from American. American responded by
cutting its lowest fares by 50%. As a result, the prices went down and the industry
experienced the worst times in its history. Northwest and Continental sued American
airlines with claims of predatory pricing, but as it has been the case in almost any
predatory pricing case since Areeda-Turner, they could not get a verdict against the
182
defendant. However, American had to spend in excess of $20 million to cover legal costs
(Clouatre 1995). Eventually, complex fare structures resurfaced. Accusing American
Airlines with predatory pricing, Northwest CEO John Dasburg then stated:
In the long run, predatory pricing will reduce the number of airlines,
ultimately cutting the number of flights and choices available, particularly
in smaller markets. This will leave the few surviving airlines free to price
just as high as they want for just as long as they want.
(Dempsey 2000b, p.3).
A Relatively Recent Incident:
American seems to take the lead in most price changes but sometimes with
differing results. In March 2002, American eliminated a three-day advance fare
(Woodyard and DeLollis 2002). The impact of this would be to force the customers who
used to buy this class of tickets (i.e., mostly business travelers) into buying a higher class
fare (i.e., 10% fare increase). However, other major carriers, which usually readily went
along with price increases, did not think that the increase was justified this time. They
already had low load factors, and with the exception of Continental, they did not comply
with American’s price increase. Surprisingly, rather than taking the increase back to its
previous level, American “retaliated” to the refusal to follow suit by dropping its three-
day-advance fares by as much as $99 on routes where it competed with the rebel carriers
and required no Saturday-night stay (Woodyard and DeLollis 2002). It offered these
discounted fares in ten markets that it competed with United, Delta, and US Airways each
(McCartney 2002), but not with Continental. Normally, this type of big brother
punishment would be expected if a member breaks the rules of a cartel such as OPEC.
This was a case of blatant signaling and retaliation in the face of refusal to comply. This
183
incident demonstrates the unhealthy nature of competition in the airline industry.
However, this time the end result was beneficial to consumers. Northwest would not take
any of American’s swashbuckling and reacted by offering round trips for $198 in 10
American markets, and when American did not give in, it offered the same discounts in
20 markets. American insisted on its discounts so Northwest finally offered $189 round
trips in some 160 American markets (Woodyard and DeLollis 2002) while American’s
comparable prices were averaging $1650 (McCartney 2002). Southwest, always in search
of an opportunity, joined the battle by re-offering its “friends fly for free” campaign after
five years (Woodyard and DeLollis 2002).
Southwest Effect: Southwest undoubtedly proved to be the most successful LCC after
deregulation. It actually had begun its operation as an intra-state carrier before
deregulation but quickly extended its operations to more states. Southwest does not use
major hubs and relies on secondary airport for its operations. It also typically does not
serve long distance markets and focuses on short and medium haul instead. Most entries
by Southwest are to markets with connecting service by incumbents. Southwest enters
these markets with non-stop service and lower prices. The result is an overall decrease in
the fare levels in the market, also called the “Southwest effect.” (1998a).When the price
wars got tough in mid 90’s because of competing discount services such as Calite
(Continental), Shuttle (United), and Reno Air, Southwest’s Pilots Union accepted no
increases for five years in return for options, which in effect enabled the airline to cut its
costs by hundreds of millions of dollars (Banks 1995). At the height of the battle, United
Shuttle even purchased 1-800-Southwest and used it as its reservations number.
184
The major carriers were not successful in their campaigns against Southwest
because Southwest had both lower costs and deep pockets. “In the 1998 calendar year,
the total domestic operating cost in cents per available seat-mile for the network airlines,
adjusted for distance, ranged from 7.737 cents for America West and 9.123 cents for
Delta to 11.582 cents for US Airways. The comparable costs for the low-fare airlines
ranged from 6.083 cents for Southwest to 8.626 cents for Frontier. Thus every low-fare
airline had adjusted costs per available seat-mile that were significantly below the costs
of any network airline except America West.” (2001g). Therefore, the major carriers
admitted failure against Southwest’s strategy and attempted to contain it rather than
defeat it. “When major network airlines were subject to entry by either Southwest or by
another major network airline, the response was typically either a very slight fare
reduction with no significant increase in capacity or a fare increase. We did not find cases
where the response was as aggressive as when a new-entrant low fare carrier entered a
market.” (Oster and Strong 2001, p.15).
A Word on a Market Niche: It should be noted that not every new entrant in the airline
industry has attempted to be a discounter. Despite the tough competition with the
incumbents, several start-ups have attempted to carve themselves a niche in the first class
business travelers’ market. For example, Midwest Express has secured itself a niche in
the tough industry. Its formula is offering superior service at slightly higher fares. The
seats are leather, and there is no center row. As a result, there are 30% fewer seats on its
DC-9s than expected (Oliver 1995). Meals come in china with silverware and “crystal
salt and pepper shakers” (though how this was impacted after September 11 remains a
185
question) (Oliver 1995). Midwest spends twice as much on its meals than other airlines
(Oliver 1995). Midwest had been owned by Kimberly-Clark Corporation but went public
in September 1995 (Oliver 1995). Recently, JetBlue, National, and Legend were among
the other names who attempted to capture the same niche in different markets. So far,
JetBlue seems to be successful, and Legend is already out of business due to aggressive
responses by American.
5.2 Competitive Outlook
“Capacity is how we compete in this business”
Bob Crandall CEO, American (Labich 1993)
The revenue of the U.S. airline industry was around $97 billion in 2001. The
profit margins have not been impressive. Revenue management is used to subsidize
discounted tickets with the gains from first and business class fares. When the demand
falls for cyclical or other reasons, the industry experiences red ink. The following section
exposes the industry with some factual information.
5.2.1 Industry Snapshot
On the day the Wright Brothers flew at Kitty Hawk if there had been a capitalist
down there, the guy should have shot down Wilbur.
Warren Buffet. (Smith 1995)
Due to increased competition from LCCs and price wars, the profitability of the
airline industry has displayed a highly cyclical pattern over the years. Table 5.13
illustrates this point. It should be noted that the price wars in 1982 and 1992 happened
while the industry was already in a recession as the airlines tried to increase their load
186
factors.
Table 5.13: The Cyclical Nature of the U.S. Airline Industry
Cycle Years Duration Airlines
Boom 1960-68 9 years Expansion
Bust 1969-74 6 years Recession
Boom 1975-79 5 years Expansion
Bust 1980-82 3 years Recession
Boom 1983-89 7 years Expansion
Bust 1990-94 5 years Recession
Boom 1995-2000 6 years Expansion
Bust 2001-2005 5 years Recession
Note: 1995-2000 expansion was a forecast which eventually proved to be accurate. 2000-
2005 recession was predicted by AirTran CEO Joe Leonard at his speech at Georgia
Pacific Headquarters on Aug 19, 2002.
Source: Aviation Week and Space Technology, 13 March 1995
One recent trend has been international growth through global alliances. For
example, American Airlines and British Airways; United Airlines and Lufthansa; Delta
Air Lines, Air France and Aero Mexico all joined forces for international expansion. The
industry appears to be leading to a new wave of consolidation. However, the merger
between United and U.S. Airways was blocked by the DOJ. Alternative expansion
strategy for the majors has been controlling some or all of regional start-ups. For
example, Delta owns Delta Express, Atlantic Southeast, Comair and has launched Song.
Another interesting response by major airlines to low cost carriers was in the form
of founding their own low-cost versions. United Airlines was the first airline to try this
strategy in October 1994, and the service was called “The Shuttle” by United (Oster and
Strong 2001).
9
Delta started Delta Express and US Airways started Metrojet in October
9
Continental had had an earlier effort with Continental Lite but the initiative lacked a consistent strategy
187
1996 and June 1998 respectively (2001g; Oster and Strong 2001). These efforts were
generally defensive in nature but also allowed the majors to use the low-cost model of the
entrants through their flexible labor agreements. Pilot unions ultimately restrained the
expansion of most of these initiatives by imposing limits on total hours that could be
flown as a proportion of total flights.
Table 5.14 illustrates the aggregate income statement of the industry over time.
The expansion and recession periods of the industry can be tracked within this Table as
well.
and eventually failed (Oster and Strong 2001).
Table 5.14: Aggregate Industry Snapshot Over time
Source: Air Transport Association Annual Report (2001c)
1
8
8
The table illustrates that the profit margins of the industry have not been
impressive even in good years. Airlines have earned a net profit between one and two
percent on average, compared to an average of five percent or more for other industries.
Only 15 have survived among the 43 pre-deregulation airlines, and two thirds of 226
post-deregulation airlines have not been able to succeed (Dempsey 2000a). On the
positive side, the airlines were highly profitable in the late 90’s and their net profits
totaled $23 billion from 1995 through 2000 Q3. The Average passenger load was around
72% percent for the same year while the break-even load for most carriers was around
65% (2001g).
More than 130 airlines have gone bankrupt since deregulation, and currently there
is little incentive for new entrants (Wysocki Jr 2001). Airlines already schedule more
flights than the eight most crowded airports in the nation can handle, which only adds to
the congestion problems at the hubs (Wald 2001). The recent activities (e.g., reports,
DOT proposed guidelines, complaints and finally the DOJ lawsuit) in the airline industry
strongly suggest that anticompetitive practices (e.g., predatory pricing) may indeed be an
issue.
The unique nature of the airline industry enables the network carriers to engage in
predation in a variety of ways. Observed tactics have included: “dropping prices sharply;
eliminating advance purchase and Saturday night stay-over restrictions; expanding the
inventory of low-fare seats offered; increasing the number of flights and/or the size of
aircraft; scheduling departures in close proximity to the new entrant’s flights, sometimes
boxing them in; offering passengers bonus frequent flyer miles; paying travel agent
commission overrides to steer traffic toward the incumbent in the new entrant’s markets;
189
190
paying higher upfront commission rates on routes where it competes with a new entrant;
biasing its computer reservations systems against non-affiliated interline connections;
refusing to enter into ticketing-and-baggage, joint-fare, and code-sharing relationships
with the new entrant; refusing to lease gates, provide services, or sell parts to the new
entrant; restricting airport operators with majority-in-interest clauses to prohibit the
construction of gates and other infrastructure for new entrants; and prohibiting affiliated
regional feeder airlines from entering into marketing agreements with the new entrant”
(Cooper 1999; Dempsey 2000b, p. 24). Half of the informal complaints received by DOT
between 1993-1999 were regarding unfair pricing and capacity increases, where as more
than 30% were regarding restrained access to gates and other facilities or services
(1999d).
Increasing complaints and actual observations of unfair conduct motivated
the DOT to propose a set of guidelines to protect competition in the airline
industry in 1998. The guidelines essentially proposed that predation would be
inferred if:
(1) the major carrier adds capacity and sells such a large number of seats
at very low fares that the ensuing self-diversion of revenue results in lower
local revenue than would a reasonable alternative response,
(2) the number of local passengers that the major carrier carries at the
new entrant’s low fares (or at similar fares that are substantially below
the major carrier’s previous fares) exceeds the new entrant’s total seat
capacity, resulting, through self-diversion, in lower local revenue than
would a reasonable alternative response, or
(3) the number of local passengers that the major carrier carries at the
new entrant’s low fares (or at similar fares that are substantially below
the major carrier’s previous fares) exceeds the number of low-fare
passengers carried by the new entrant, resulting, through self-diversion, in
lower local revenue than would a reasonable alternative response
(1998c).
191
The responses for the above guidelines were mixed, and generally the
parties who would be distracted by the application of it opposed to it while those
who suffered from/disturbed by anticompetitive practices supported it. Therefore,
the major airlines, their trade association, certain academics, and unions opposed
the guidelines (Gattuso and Boudreaux 1999), whereas the low-fare airlines,
GAO, and Alfred Kahn (1998) supported them.
An interesting counter-proposal to the DOT guidelines came from Foer (1999).
He argued that the DOT guidelines would not be effective since the courts take too long
to resolve the cases (if ever) and that the low-cost carriers go out of business long before
then. He emphasized the need for greater predictability and proposed a “Safe Harbor”
option. A major carrier (defined as 50% or more marker share from a hub) could take the
Safe Harbor option when challenged by a new entrant. A major carrier that chose to
comply with the Safe Harbor option would not be subject to any predation lawsuits. That
option would require them to make a public statement that they will commit to their
response level (low) fares and increased level of capacity for at least two years. They
would have to file a public notice to the DOT. And with certain exceptions, they would
not be able to increase their prices or decrease their capacity for a period of two years. He
suggested that the majors would not alter their price and capacity unless they were
willing to live with it and suffer the losses for two years. The pattern where the major
goes back to the high (monopoly) levels of pricing and capacity would be effectively
constrained with the Safe Harbor (Foer 1999). Naturally, ongoing antitrust law would
have to be modified before Safe Harbor could become successful. Otherwise, no major
carrier would opt for it when they have the option to predate and not be found guilty
192
under existing laws.
Canada, facing similar predatory problems in their airline industry passed
Regulations Respecting Anti-competitive Acts of Persons Operating a Domestic Service,
S.O.R./00-324 in year 2000. Predation by incumbent would be detected based on two
rules with respect to AAC:
(a) operating capacity on a route or routes at fares that do not cover the
avoidable cost of providing the service; and
(b) increasing capacity on a route or routes at fares that do not cover the
avoidable cost of providing the service (cf. West 2000).
Even though an average avoidable cost test has its own complexities such as
determining which costs were avoidable and the appropriate time frame, it has been
successfully employed in at least one case in Canada (West 2000).
An impartial body, the Transportation Research Board of the National Research
Council encouraged DOT to continue their investigations. They observed that “a cursory
review revealed some actions that were difficult to reconcile with fair and efficient
competition. Particularly difficult to reconcile were cases in which incumbent carriers
added nonstop service in low- to moderate-density markets they had not previously
served directly, coincident with a new entry. In some of these cases, the incumbent
bypassed its own hub to initiate the service, a strategy seldom employed outside of high-
density markets. The logical inference is that such responses are probably temporary—
possibly calculated to protect the incumbent’s hub traffic and to dissuade similar
challenges elsewhere—and would seem to warrant additional scrutiny” (1999d, p.33).
193
Inspired by DOT’s guideline proposal, DOJ decided to file a lawsuit against
American Airlines in 1999. This was the first DOJ predatory pricing lawsuit in the airline
industry since deregulation (Nannes 1999). DOJ alleged that American had used
predatory tactics in order to drive Vanguard, Western Pacific, and Sunjet out of its
markets (1999e, para 23-28):
American has monopoly power in most of its DFW city pairs and faces
little current competition and little prospect of entry on those routes. Its
monopoly power allows it to charge supracompetitive fares. American’s
fares on DFW city pairs are substantially higher then its fares on
otherwise comparable routes where it faces competition….[W]hen a [low
cost carrier] entered a DFW route and it appeared the [low cost carrier]
would be economically viable if American simply followed a profit-
maximizing business strategy, American would instead saturate the route
with enough additional capacity at low fares to keep the entrant from
operating profitably. American would also take further steps, such as
matching the [low cost carrier’s] connecting fares with its own nonstop
fares, to keep traffic away from the [low cost carrier]. To evaluate the
success of its strategy and determine whether to intensify its response,
American would investigate the financial resources of [low cost carriers],
determine their break-even factors, and conduct head counts at the
departure gate to monitor their passenger loads.
As a result of these practices all three entrants had to leave the markets
they entered, and all were bankrupt before September 11, 2001.
Table 5.15: American Airlines’ Actions Relevant for the DOJ Lawsuit
American Before During After
From Dallas
To
Flights Price Flights Price Flights Price
Kansas City 8 $108 14 $80 11 $147
Long Beach 0 --- 3 $86 0 ---
Colorado
Springs
5 $150 7 $81 6 $137
Source: (Carney and Zellner 2000).
194
Vanguard markets: American operating at 65% load factor charged $108 before
Vanguard’s entry on DFW to Kansas route. They reduced their rates to $80 following the
entry matching the fares. American also added 6 more round trips to the route, driving
Vanguard out of the market in the process. After Vanguard pulled out, American
decreased the daily round-trips to 11 from 14 and increased the prices by 80%. It was
also documented that the route went from the worst performing to the best performing
American route after the exit of Vanguard.
SunJet markets: American responded to the entry announcement of SunJet to the DFW –
Oakland route. They matched SunJet’s fares and also entered the DFW-Long Beach route
in which it also matched Sun Jet. SunJet was forced to move out of all DFW routes.
American increased prices by over 30% in the two months following Sun Jet’s exit.
Western Pacific markets: American increased its capacity significantly to put pressure on
Western Pacific. The frequency was increased from 5 to 7 flights per day. As a result of
this response, Western pulled out of the routes and declared bankruptcy. American’s fares
were in the range of $81-105 when Western was in the market. After they pulled out
American went back to its original frequency of 6 and increased the prices to $137 on
average.
As it was discussed previously, cases of predatory pricing a very hard to prove in
the airline industry, and the judge of DOJ’s American lawsuit dismissed it by summary
judgment in May, 2001. DOJ appealed the decision but cold not overturn the ruling.
195
The DOT issued a series of reports in January 2001 re-stating their concern about
the anti-competitive practices in the airline industry (2001e; 2001g; Oster and Strong
2001). They concluded that “incumbent airlines at times have responded to new
competition with fare cuts, capacity increases, and other practices that are apparently
designed to eliminate or reduce competition” (2001g). The DOT reminded that they were
authorized by the Congress to prevent unfair practices in the airline industry and warned
that they had the right and the obligation to prevent such practices even if they do not
violate antitrust laws. After studying thousands of responses their proposed guidelines
generated, the DOT concluded that they would examine each case independently due to
the complexities and differentials in the marketplace, and one governing guideline would
be too simple of an approach. Instead, the DOT expressed a wish to develop a “body of
caselaw based on a more thorough examination of cases of apparent predatory-type
behavior” (2001g, p.11).
The second report was entitled Predatory Practices in the U.S. Airline Industry by
Professors Oster and Strong (2001). It described the patterns of entry and competition,
and included an analysis of predatory conduct and predatory pricing with examples from
the airline industry. This report examines the potential for predatory practices, or unfair
methods of competition, in the U.S. domestic airline industry. One of their two main
conclusions was that “predatory practices may have occurred in the past and are a
recurring possibility in the U.S. domestic airline industry.” They detected high use of
market power in the airline industry, hub premiums and recoupment of losses suffered
during predation against new entrants. Their second conclusion was that the antitrust
laws, as they have been applied in other industries, may not be sufficient to identify some
196
types of predatory practices in the airline industry. Their work is highly relevant to
understand he competitive context of the airline industry, and several of their figures have
been cited and reproduced in this chapter.
The third report issued by the DOT in January 2001 was entitled Dominated Hub
Fares (2001e). This report summarized, updated, and provided new support for the
existence and demonstration of hub fares and the effects of LCC service on fare levels. It
argued that quality service and reasonable fares can co-exist just like major carriers and
LCCs can. Atlanta and Salt Lake City were provided as examples for co-existence. The
report concluded that “the key to eliminating market power and fare premiums is to
encourage entry into as many uncontested markets as possible.”
The DOT received 32 informal complaints from 1992 until 1999. Half of these
were allegations of unfair pricing and capacity responses –the dumping of low-fare
capacity in the city-pair market and, in some cases, added flights. Others were about
unfair marketing, airport handling relationships (higher travel agent commissions) (Oster
and Strong 2001, p.10). The administration concluded that “apparent unfair practices
have occurred in the airlines industry”(2001g, p.114). “Free markets do not exist in a
state of nature. Free markets are things that have to be defined by custom and law”
(Dempsey 2000a, p.487).
5.2.2 The Role of Barriers to Entry
It has been shown that the theory of market contestability did not apply to the
airline industry because of high barriers to entry (Baker and Pratt 1989; Hurdle et al.
1989). The threat of potential entry may not force the airlines to cut their prices since
197
they can readily and effectively react to any entry with price cuts and capacity increases
(Levine 1987). Instead, the incumbents may attempt to raise the barriers to entry and
develop a reputation for predation by competitive and aggressive reactions to entry to
their hub markets. Therefore, these barriers gains special importance when conducting
research in this industry. Among the six generic barriers identified by Porter (1980),
several either do not apply or take on a proxy form in the airline context. For example,
cost advantage of incumbents’ would not necessarily apply since the LCC entrants are
presumably more efficient within this framework. However, barriers such as product
advantages of incumbents are captured by hub formation and code-share agreements.
Both hubs and code-share agreements have been associated with higher fares (2001e;
Hassin and Shy 2000). Figure 5.6 presents proposed proxies for Porter’s barriers to entry
for the airline context.
Porter’s six generic barriers to
entry:
Barriers as reflected in the Airline Industry:
Cost advantages of incumbents The LCC entrants have cost advantages
Product advantages of incumbents Hub Formation
Capital requirements Airport Congestion, Reputation Effects
Customer switching costs Code-Share agreements, Frequent Flier Miles
Access to distribution channels Gate, Slot and Noise Controls and Availability
Government policy Gate, Slot, and Noise Controls
Figure 5.6: A Comparison of Generic and Industry Specific Barriers to Entry
As discussed in a prior section, Karakaya and Stahl (1989) have proposed an
extended list of generic barriers to entry. Several of the items among their extended list
are of interest and captured either in different components of the proposed framework
(e.g., market concentration, price) or under the current proxies of barriers to entry (e.g.,
gate and slot controls and availability is a proxy for limited access to essential facilities).
198
Another barrier to entry suggested by Karakaya and Stahl (1989) is incumbent’s expected
reaction to entry and this has been included in the framework in the form of reputation
effects of incumbent’s predation. It might also be interesting to consider advertising,
product differentiation of incumbents and brand name or trademark as potential barriers
to entry from a marketing perspective. However, these involve further complications in
terms of measurement.
Slot and Gate controls: Certain airports are considered high-density airports. This
indicates that in these airports, an airline must have slots to schedule a flight (take-off or
landing) between 6 a.m. and midnight. For example, FAA’s high density rule permits 48
slots at LaGuardia from 6 am to midnight every day, resulting in (48x18) 864 slots per
day (1995b). Slot controls have been imposed in four major airports (Chicago O’Hare,
New York La Guardia, New York Kennedy, and Washington National) since late 1960s.
It has been estimated that the fares in slot-controlled airports have been 11-15% higher
than non-controlled comparable airports with the exception of New York Kennedy
(Morrison and Winston 1997). Congestion pricing (charging aircraft for their take-offs
and landings according to the delays it imposes) has been offered as an alternative to slot-
controls (Morrison and Winston 1997).
Joe Leonard, CEO of AirTran, has noted that they are able to serve Atlanta with
cheap fares because Eastern Airlines went bankrupt and freed up 22 gates for AirTran
(2001h). Similarly, JetBlue is serving New Yorkers because they were awarded 75 slots
at JFK airport by the FAA (2001h). Landing slots in most major airports are not
available. The incumbent is rarely, if ever, willing to lease extra slots to a new entrant.
199
Even in the rare case of lease, monopoly rents may be collected. The FAA “Buy-Sell”
slot rule lets deeper pockets ensure market share by getting the slots and enjoy
monopolistic pricing (Dempsey 1989). Slot control regulation needs to be significantly
revised so that LCC access to markets is possible and not left to coincidence or awards by
officials. Before the DOT permitted the landing slots to be resold, the largest eight
airlines controlled about 70% of the slots at Chicago O’Hare, New York La Guardia,
Kennedy, Washington National. In a couple of years the figure increased to 96%
(Dempsey 2000a, p.450).
Table 5.16: Percentage of Domestic Air Carrier Slots Held by Selected Groups
Airport Holding Entity 1986 1991 1996
O’Hare American and United 66 83 87
Other Established airlines 28 13 9
Financial Institutions 0 3 2
Post-deregulation airlines 6 1 1
Kennedy Shawmut Bank, American, and Delta 43 60 75
Other Established airlines 49 18 13
Other Financial Institutions 0 19 6
Post-deregulation airlines 9 3 7
LaGuardia American, Delta and US Airways 27 43 59
Other Established airlines 58 39 14
Financial Institutions 0 7 20
Post-deregulation airlines 15 12 2
National American, Delta and US Airways 25 43 59
Other Established airlines 58 42 20
Financial Institutions 0 7 19
Post-deregulation airlines 17 8 3
Note: Some Financial institutions have taken control of slots when incumbent carriers
went bankrupt.
Source: Anderson (1997)
Similarly gate controls are an issue, and the dominance of major carriers in
controlling the gates is illustrated in Appendix D (1999a). Low Fare carriers who cannot
200
secure access to gates are left to the mercy of the major carriers and have to pay hundred
per cent to two hundred per cent premiums to use the gates (Woellert 1998). Spirit
Airlines was reported to face such a fate since it had been outbid by the majors in Detroit
continuously for almost a decade (Ingersoll 1999). Thus, slot and gate controls can also
effectively serve as barriers to entry and lead to increased fares (Hurdle et al. 1989;
Morrison and Winston 1990).
5.2.3 After September 11, 2001
The general belief about the airline industry is that it is not really a profitable
industry and that it should be left alone. Although examples of apparently unfair practices
seem to be common, this general belief has provided the industry some relief with respect
to antitrust investigations. Such empathy regarding the industry is higher than ever after
September 11. The industry has received around $15 billion worth in direct aid and
guaranteed loans. The total direct aid payments by October 2002 have exceeded
$4,299,852,435.3 to some 382 airlines. Table 5.17 illustrates the direct aid that the larger
carriers have been provided by the federal government as of July 25, 2002.
Table 5.17: Direct Aids to Major Airlines by the Federal Government
UNITED AIR LINES, INC. $724,485,485.8
AMERICAN AIRLINES, INC. $656,032,849.3
DELTA AIR LINES, INC. $594,894,536.4
NORTHWEST AIRLINES, INC. $405,525,526.3
CONTINENTAL AIRLINES, INC. $343,398,895.9
US AIRWAYS, INC. $287,219,674.0
SOUTHWEST AIRLINES CO. $264,642,089.9
AIRTRAN AIRWAYS, INC. $27,690,969.2
Source:http://www.dot.gov/affairs/carrierpayments.htm
201
The airlines also laid-off employees and cut back on services in response to
falling demand and increasing security costs. However, the future of the major airlines
still looks gloomy. The consumers save $6.3 billion annually due to low-fare airline
competition (1996). On June 18 2002, The Wall Street Journal reported that ‘Wal-Mart’
airlines were crunching the biggest carriers. The number of passengers carried by the five
major airlines was 10% lower in May 2002 than it was in 2001. One would consider this
a result of September 11, and the turmoil it created, but the traffic of the five biggest low-
cost carriers increased by 11% in the meantime.
The future of the surviving LCCs looks brighter than it did before September 11.
Obviously, the same conclusion cannot be drawn for the major carriers who have yet to
re-structure their costs. Union relations play a big role in restructuring and labor expenses
represent up to 40% of the revenue of a major airline but only 25% of the revenues of
low-cost carriers (Trottman and McCartney 2002). Another advantage of low-cost
carriers is their higher aircraft utilization due to their point-to-point and secondary airport
operations. The transparency of (low) prices and distribution of tickets (e-tickets) have
also helped the low cost carriers. More than 40% of Southwest tickets are booked online
whereas the same figure is around 5% for majors (Trottman and McCartney 2002).
The fare difference between flying business class on a major carrier and
discounter has become so wide that some employers allow their employees to purchase
two discounter seats so that they can get some extra space to stretch out (and still spend
less money than it would cost on a major carrier) (Trottman and McCartney 2002). While
the passenger traffic in April 2002 was down by 10.5% from two years earlier, the major
202
carriers had increased the average cost of a 1000 mile coach ticket had by 14.7%
(Trottman and McCartney 2002).
5.2.3.1 Anticipating Trends
Our major competition over the last fifteen years was other people doing
the same things we do. That’s not the case anymore.
Don Carty, CEO, American Airlines (2002)
In July 2002, Delta announced that the competitive threat is not from the big
players such as American but from the discounters (Fonti 2002). It devised low-fare
competition strategy with McKinsey and Co. (Fonti 2002). In the meantime, Southwest
announced an aggressive expansion plan “connecting the dots” between city pairs
common for business travel (e.g., Baltimore to LA, the highest Southwest round-trip fare
was $598 as opposed to $1127 of United) (Trottman 2002).
In August 2002, Delta, Northwest and Continental announced a marketing
agreement which links flight schedules, frequent flier programs, and access to airport
clubs. A similar deal was earlier announced by United and US Airways (Woodyard
2002). In the meantime, the share of revenues from premium passengers (i.e., business
and first class), average industry yield and revenues have been decreasing (Hazel 2003).
Deriving from the above sections, I anticipate that the following trends could be
observed in the industry in the near future.
1. Industry Consolidation Attempts: As the competition and pressure for
profitability increases, the majors will attempt to consolidate. US Airways is
not in a position to compete with either majors or LCCs. It is likely that it will
203
either shrink to its roots as a regional carrier or be taken over by a major
carrier (Sheth and Sisodia 2001). If the second option is realized, other
mergers would follow. The CEO of America West Airlines recently predicted
that United, Delta and Northwest would remain as major network carriers, and
AirTran, Southwest, and Jet Blue would remain as discount carriers following
anticipated mergers (Sunnucks 2005).
2. Antitrust Action: In relation to the above scenario, the DOJ will have to revise
its criteria for evaluating mergers. DOT reports have shown the effect of the
presence of LCCs on fare levels, and their current measure (HHI) does not
take this into consideration. A new set of criteria could enable consolidation
action as long as there is sufficient LCC presence in the markets.
3. Revision of the Hub-and-Spoke System: The efficiency and cost of managing a
large scale hub-and-spoke system is being questioned. Major carriers may be
better off concentrating on medium and long haul markets and leaving the
short-haul spoke (feeding) operations to regional operators that they own or
control. Regional carriers have better cost structures that can effectively
compete with other LCCs.
4. Restructuring of Corporate Travel Policies: As audio and video conferencing
becomes widespread, corporations will revise what constitutes a need for
travel for whom and at what cost. More and more companies are already
booking flights on LCCs (Woodyard 1999). LCCs will want to get a larger
share of the corporate pie, and they will arrange umbrella deals with them. For
example, ProAir offered unlimited business trips for a flat rate, and Vanguard
204
offered 10-15% discounts for businesses that buy 10 or more trips (Woodyard
1999).
5. Increased Market Access and Growth for LCCs: Permission of the mergers
with special access conditions for LCCs could serve to increase their presence
in key markets. The Federal government has already asked the airports to
reveal their plans for increased competition. More than a dozen hub airports
with high concentration were requested to file “competition plans” in order to
secure federal expansion funding or increase facility charges (Pinkston 2000).
These plans would include new gate plans, ticket counters, and strategy to
attract new (i.e., LCC) airlines. Slot controls are also likely to be revised and
relaxed for the benefit of LCCs. However, the presence of LCCs in bringing
lower fares and stimulating economic growth in communities is well
understood. If the Federal government acts slowly in opening up the markets
to LCCs, the communities are likely to take the matter in their own hands. The
following section demonstrates such a case. This is a very important
phenomenon as it demonstrates the insufficiency of public policy at the
federal level.
Case in point: Communities want LCC Service. Sam Williams, the President of Metro
Atlanta Chamber of Commerce, commissioned a study to examine the fare levels with
benchmark cities and stated “We want more airlines brought in here” (Saporta 1998).
More and more communities are becoming aware of the substantial effect that the
discounters have on prices. Since the public policy makers have been slow to open up
markets to LCCs, some communities have decided to be proactive and have taken the
205
matter in their own hands. For example, the business community in Wichita, Kansas have
raised $4.7 million so that they could attract AirTran (Fonti 2002). The money would be
solely committed for ticket purchases on AirTran for the next two years (Fonti 2002).
Air-Tran would also be able to draw funds if ticket sales fall below expectations. Finally,
the community would spend $600K for marketing AirTran flights in Wichita (Fonti
2002). Wichita was previously served by three majors: Delta, American, and United, and
the community’s multimillion dollar investment would bring much lower fares in return.
Similar deals with AirTran have already been successfully employed in six other markets:
Gulfport/Biloxi-Mississippi, Pensacola-Florida, Tallahassee-Florida, Grand Bahama,
Newport News, Va., and Rochester-New York (Fonti 2002). Admitting the failure of
current policy, the DOT itself stated that communities that have little or no service by
LCCs should actively seek their presence. The communities would have to market
themselves to LCCs, and it was advised that they seek advice from aviation consultants
(1996c).
5.2.4 European Landscape
European policy makers have observed the deregulation process in the US
closely. Unrestricted cross-border flights have been allowed since 1993 (McCormick and
Field 1997). The new start-ups in Europe have adopted Southwest’s business model and
tried to emulate its success. Successful European low cost carriers include Ryanair,
EasyJet, Virgin Express, and Debon Air:
10
10
This section draws heavily from (McCormick 1997) March 31 3B.
206
Ryanair: RyanAir is the oldest of the LCCs in Europe. It took advantage of early
deregulation and started its operation in 1985 flying between Britain and Ireland. It is the
leading carrier on the Dublin-London route. Continental Airlines chairman David
Bonderman owns 20% of Ryanair. Like Southwest, Ryanair intentionally ignores the
major airports due to their congestion and aircraft turnaround times. No frequent flyer
miles, no free snacks. However, this structure enables Ryanair to be able to break even at
50% utilization (and its load factors are around 75%) (Michaels 2000). Ryanair uses only
one type of aircraft (the Boeing 737) and is based in Dublin.
EasyJet: It appears that Easyjet is trying to emulate a combination of People Express and
Southwest. It offers a no-frills service: seats are not assigned (first come first serve), no
free meals, and casually dressed flight attendants. EasyJet also does not use travel agents
or a CRS system for ticket sales. The result is 25% cost savings per ticket and also fewer
employees per aircraft. Its base is London-Luton airport.
Virgin Express: Virgin Express was created when Richard Bronson (Founder and
Chairman of Virgin Group including Virgin Atlantic Airways) purchased European
Business Airlines. The initial problem Virgin Express faced was the high tax and cost
structure in its base, Brussels, Belgium. They also faced the problem of not getting the
slots they needed for efficient operations to their destinations. However, Richard Bronson
is reputed as a tough competitor, and Virgin Express is a strong brand throughout the
European Union. Its current base is Brussels-Zaventem airport.
207
Debonair: Debonair attempts to use a hybrid strategy: quality service at an affordable
price. They do offer quality snacks on board but no meals. Debonair has in-flight
entertainment at every seat and plans to introduce video-on-demand and in-flight gaming.
Like EasyJet, Debonair’s base is London-Luton airport.
As a result of increasing pressure from the emerging LCCs, some large players
who could not adopt to the new competitive field are facing trouble (e.g., Iberia, Alitalia,
Olympic). On the other hand, British Airways has aggressively made acquisitions in
Germany and France to increase its presence. Concentration and prices are expected to
decrease in the coming years in Europe as the benefits of deregulation are greeted by
more consumers.
208
CHAPTER 6
DATA AND METHODOLOGY
6.1 Data Set
The Airline Origin and Destination Survey (DB1B) is a 10% sample of airline
tickets from reporting carriers collected by the Office of Airline Information of the
Bureau of Transportation Statistics. This database is used to determine air traffic patterns,
air carrier market shares and passenger flows. The data includes origin and destination,
distances, passengers per day, average one way and round trip fares for all carriers, fared
pasangers, market shares per city-pair and airport, average yields, market ranks, and
revenues. The data can be aggregated for airport, for markets within an airport or at city-
pair detail. The Database Products Inc. makes the data commercially available in the
form of several CD-ROMs. Access was gained to these databases through industry
contacts. This available data of 28 quarters (from 1
st
quarter of 1993 to fourth quarter of
1999) allows for analysis of aggregate and individual markets. From this analysis, one
could understand and identify market defense behavior in light of the principles discussed
in this dissertation. The content of the databases has been summarized as follows in Table
6.1 (Dixit 2000):
209
Table 6.1: Contents of O&D databases
Variables/Database O&D Plus
Market
O&D Plus
Carrier
Onboard
Passengers yes yes yes
Market Share yes yes yes
Distance yes yes yes
Revenues yes yes no
Yield (cents per
passenger mile)
yes yes no
Onboard Passengers no no yes
Seats available no no yes
Market Rank Yes yes yes
Fared passengers yes yes no
Fare yes yes no
Enplaned Passengers yes yes yes
Other data for code-share agreements among carriers, gate and slot controls and
usage agreements, carrier financial resources, employees etc. come from various DOT
reports (see section on measures) and have been manually and built into the database
extracted through a cumbersome process.
6.2 Methodologies
Most empirical work to date seems to have utilized standard approaches such as
regression analysis as their methodology (e.g., Abunassar 1994). One exception is Dixit
(2000) who used neural networks and binary logit analysis. I have considered a wide
array of methodologies to ensure robust and unbiased results and to strengthen the
validity of the findings a la triangulation. Sophisticated methods such as binary and
ordered logit analysis, multi-level modeling with mixed coefficients (HLM), and event-
history analysis were determined to be feasible. Logit models enabled the analysis of
limited dependent variables; HLM enabled to account for the hierarchical structure in the
210
data; and even-history analysis overcomes the issues with right-censoring (i.e., exits
unobserved in the database). Split sample and double-cross validation have been
employed for validation purposes where applicable.
6.2.1 Logit models
The applicability of logit models to the airline industry context has been shown by
Dixit (2000). The model is quite robust, and its use is familiar in marketing.
The ordinal categorical/binary nature of the dependent variables in the model
(e.g., supra-competitive, competitive, entry/no entry; exit/no exit) facilitates the use of
ordered/binary logit models. Ordered logit implies that the dependent variable can be
ranked but the distance between the categories is not known (Aldrich and Nelson 1984;
Menard 2002). Binary logit implies a dependent variable with two known states.
The assumptions of the limited dependent variable models, similar to multiple
regression models, are as follows (Long 1997):
a. The x’s and ? are not correlated ( \ ) 0 E x ? =
b. Error terms have constant variance
c. Error terms are normally (or logistically) distributed with a mean of zero.
An ordered regression model such as logit (or probit) can be derived when an
infinite range for y is mapped to an observable variable and cut-off points (thresholds) are
established (Long 1997). For example, an ordered logit model with three levels for the
dependent variable would be specified as follows (Borooah 2002):
Yi = 1, if Fi ? ?
1
Yi = 2, if ?
1
?Fi ? ?
2
211
Yi = 3, if Fi ? ?
2
where ?
1,
?
2
? 0; ?
1
< ?
2
, and represent unknown parameters. Te classification for pricing
strategy depends on whether the observed fares surpass the respective thresholds. The
probabilities for Yi for each pricing strategy are:
Pr(Y
i
=1) = Pr(Z
i
+?
i
? ?
1
) = Pr(?
i
? ?
1
– Z
i
)
Pr(Y
i
=2) = Pr(?
1
? Z
i
+?
i
? ?
2
) = Pr(?
1
– Z
i
< ?
i
? ?
2
– Z
i
)
Pr(Y
i
=3) = Pr(Z
i
+?
i
? ?
2
) = Pr(?
i
? ?
2
– Z
i
)
The ordered logit model assumes that the error term (?
i
) is logistically distributed
(on the hand, an ordered probit model would assume the error term to be normally
distributed. Greene (2000) suggests that while it is hard to specify this distribution based
on theory, the results are equivalent in most applications). Therefore, the cumulative
distribution function of the random variable X is:
exp( ) 1
Pr( ) ( )
1 exp( ) 1 exp( ))
x
X x x
x x
? = ? = =
+ + ?
and the logit model becomes:
1
1
1
Pr( 1) ( )
1 exp( )
i i
i
Y Z
Z
?
?
= = ? ? =
+ ?
2 1
2 1
1 1
Pr( 2) ( ) ( )
1 exp( ) 1 exp( )
i i i
i i
Y Z Z
Z Z
? ?
? ?
= = ? ? ? ? ? = ?
+ ? + ?
2
2
1
Pr( 3) 1 ( ) 1
1 exp( )
i i
i
Y Z
Z
?
?
= = ? ? ? = ?
+ ?
212
The estimates for
1 2
, ,
k
? ? ? are estimated through the maximum likelihood
function. When the intercept is explicitly included, the following general notation
is revealed:
0 0
1
ik
K
i k i i i
k
F X Z ? ? ? ? ?
=
= + + = + +
?
The binary model is simpler in that only one threshold point has to be defined for
the dependent variable.
6.2.2 Multi-level mixed coefficient models (Hierarchical linear modeling)
Multilevel analysis focuses on the analysis of data with complex patterns of
variability, with a focus on its nested sources (Luke 2004; Snijders and Bosker 1999).
Much of the phenomena in social and behavioral sciences consist of nested layers (Kreft
and Leeuw 2002). Variables may relate to individuals, groups, teams, departments,
organizations, countries, trading blocks and so forth. Multilevel models have been
developed to deal with hierarchically structured data. These models are referred to as
mixed-effect and random effect models in biometric applications, random coefficient
regression models in econometrics, and covariance component models in statistics
literature (Raudenbush and Bryk 2002). It is also referred to as Hierarchical linear
modeling in the social sciences because of popular software package with the same name.
A random coefficients multi-level model is said to exist when there are at least two
levels, and the coefficients at the lowest (micro) level are treated as random variables at
the second level. In contrast, mixed effect models assume that some of the coefficients
are fixed and some are random (Snijders and Bosker 1999).
213
Adding new level(s) to the model renders it more general and often more useful
(Kreft and Leeuw 2002). Analysis from data aggregated at different levels may produce
different results when a hierarchy exists. This can be determined by intra-class
correlations. If intra-class correlation is practically zero, then group differences do not
exist and can be ignored in the analysis. However, by assuming and modeling this
correlation, the implicit nested structure of the data is accounted for. On the other hand, if
this correlation is substantial and is ignored, the reliability of the results become
questionable (Kreft and Leeuw 2002). Therefore, the choice for the level of aggregation
used to analyze the data becomes important. If we aggregate the data, we may potentially
discard meaningful data and also introduce an aggregation bias. On the other hand, if we
conduct our analysis at the individual (i.e., micro) level, we violate the independence
assumption.
Overall, multi-level models help the researcher in three main ways (Raudenbush
and Bryk 2002, p.7):
1. Improved estimation of effects within individual units (e.g., developing an
improved estimate of a regression model for an individual school by borrowing strength
from the fact that similar estimates exist for other schools)
2. The formulation and testing of hypotheses about cross-level effects (e.g., how
varying school size might effect the relationship between social class and academic
achievement within schools)
3. Partitioning of variance and covariance components among levels (e.g.,
decomposing the covariation among set of student-level variables into within- and
between-school components).
214
These features enable researchers to construct more realistic models and test
hypotheses across levels with greater accuracy and reliability. The use of multilevel
models has been recommended unless there is reason to believe that hierarchical structure
in the data is not an issue. The method is emerging and more and more researchers in
marketing have been utilizing it (e.g., Malhouse et al. 2004; Van den Bulte 2000).
Luke (2004, p.13-14) categorized multilevel models in three main classes:
1. Unconstrained
1:
ij oj ij
Level Y r ? = +
00 0
2:
oj j
Level u ? ? = +
Mixed-Effects model becomes
00 0 ij j ij
Y u r ? = + +
This model is used as a null model to estimate between group effects.
2. Random intercepts:
1:
ij oj ij
Level Y r ? = +
00 01 0
2:
oj j j
Level W u ? ? ? = + +
or with explanatory variable
1
1:
ij oj j ij ij
Level Y X r ? ? = + +
00 0
2:
oj j
Level u ? ? = + ;
1 10 j
? ? =
Mixed-Effects Models are
00 01 ij j oj ij
Y W u r ? ? = + + + and
00 01 ij ij oj ij
Y X u r ? ? = + + +
respectively.
3. Random intercepts and slopes
0 1
1:
ij j j j ij
Level Y X r ? ? = + +
0 00
2:
j oj
Level u ? ? = + ;
1 10 1 j j
u ? ? = +
215
with cross-level interaction terms (W
j
represents the interaction term):
0 1
1:
ij j j j ij
Level Y X r ? ? = + +
0 00 01
2:
j j oj
Level W u ? ? ? = + + ;
1 10 11 1 j j j
W u ? ? ? = + +
The mixed effects models are
00 10 0 ij ij j ij ij ij
Y X u u X r ? ? = + + + + and
00 01 10 11 0 1 ij j ij j ij j j ij ij
Y W X W X u u X r ? ? ? ? = + + + + + + respectively.
6.2.3 Event-History analysis
11
An event is described as a qualitative change that occurs at a specific point in
time. Therefore, while a price change no matter how drastic, does not necessarily
constitute an event, market entry and exit are considered events. It has been suggested
that the best way to examine the causes and consequences of such events is through
event-history analysis. If standard methods are applied to event-history data, the results
can be severely biased due to censoring and time-varying explanatory variables. The
event-history analysis has found many applications including but not limited to
unemployment studies, consumer behavior studies (i.e., brand choice over time), medical
studies on the course of illness, learning experiments in psychology and instruction
research, insurance and accident studies, studies of migration, analysis of family
formation and fertility, criminology studies ad legal research, organization and
management research (cf. Blossfeld et al. 1989). There are several dimensions that need
to be considered before one can apply event-history analysis. I consider these in view of
my model and data set.
11
This section is based on Allison (1984) except where cited otherwise.
216
Distributional versus regression methods. Most of the early research that used
event-history analysis has focused on the distributional issues such as the time until an
event or the time between events. More recently, the focus has shifted to regression
models in which the event is the dependent variable where its occurrence depends on
explanatory variables. Similarly, the factors that lead to market entry/exit rather than the
time that passes until a market entry is of interest for the current research.
Repeated versus nonrepeated events. Events can be repeated (i.e., marriage) or
non-repeated (i.e., death). Not surprisingly, repeatable events are more complex to model
but can be handled with event-history analysis. There can be more than one entry/exit in a
given market in a longitudinal study or one can simply assume that it is a one-time event
for a market.
Single versus multiple kinds of events. Events can be treated as identical or
distinguished from one another (i.e., voluntary versus involuntary job termination). As
such, market entry and exit in my models can be treated as identical or categorized as
inter-entry and intra-entry.
Parametric versus nonparametric methods. Non-parametric methods make few or
no assumptions about the distribution of the data. Alternatively, the researcher can
assume a specific distribution (e.g., Weibull). Hybrid approaches have also been
developed in which the regression model has a specific functional form but the
distribution of the event times is not specified (resembles the linear models where no
distribution is assumed for the error term).
Discrete versus continuous time. If the researcher is able to measure the “exact”
event time or assume a continuous scale, then continuous models can be used. Otherwise,
217
discrete units (e.g., quarterly data) can also be handled by the event-history method.
Discrete time models are also easier to implement.
In simple form event-history analysis is conducted in the form of life-tables. This
essentially implies that the survival time data is grouped.
If we define time intervals as I
j
where j=1,….J : I
j :
[t
j
, t
j+1
), where
D
j
: the number of failures observed in interval I
j
M
j
: the number of censored spell endings observed in interval I
j
N
j
: the number at riskof failure at start of interval
S
j
, F
j
: survival and failure functions for interval j respectively
If we assume that the transitions are evenly spread, then the notation is (Jenkins 2004):
2
j
j j
d
n N = ? and
1
ˆ
( ) (1 )
j
k
k k
d
S j
n
=
= ?
?
since S(t)=1-F(t):
1 1
ˆ ˆ ˆ ˆ
( 1) ( ) ( ) ( 1)
ˆ
( )
j j j j
F j F j S j S j
f j
t t t t
+ +
+ ? ? +
= =
? ?
hazard rate is estimate for mid-point of the interval becomes
ˆ
[ ( )]
ˆ
( )
( )
f j
j
S j
? =
where
ˆ ˆ
( ) ( 1)
( )
2
S k S k
S k
+ +
=
While the notation becomes more complex when we introduce explanatory
variables to the model, a Cox proportional hazards model with two time-constant
variables can be simply generalized as:
Log h(t) = a(t) + b
1
X
1
+b
2
X
2
;
where a(t) can be any function of time (Allison 1984).
218
The main methods that are utilized are contrasted with each other and other
alternatives in Table 6.2 in order to demonstrate their relative advantages.
Table 6.2: Comparison of Alternative Methodologies
Methodologies Pros Cons
Multiple Linear
Regression
• Easy to interpret and
communicate
• Interaction and dummy
variable coding
• Not suitable for binary
dependent variables
Logit Models • Can effectively deal with
binary/ordered/ multinomial
dependent variables
• Arbitrary choice for
dependent variable
• Cannot deal with censored
observations
Hierarchical Linear
Modeling
• Considers the nested structure
of the data and cross-level
correlations
• Time-consuming to prepare
data
Event-History
Analysis
• Can effectively deal with
censored observations
• Utilizes full information for
the dependent variable
• No assumptions on the nature
and shape of the hazard
function (Cox)
• Can deal with time-varying
explanatory variables
• Data hard to get and time
consuming to prepare
• Feasible only for qualitative
changes with known times
Structural
Equation Modeling
• Tests the full model rather
than a stepwise approach
• Not feasible when the
variables have single
indicators or formed of
formative (not reflective)
scales
Time-Series Models • Considers autocorrelation in
the data and can establish
cause and effect relationships
• Data hard to get and very
time consuming to prepare
Game-Theoretic
experiments
• Enables precise modeling of
timing and information levels
• Difficulty with complex
models prohibits
construction of a
generalizable model
• Hard to find executive
subjects
6.3 Measures
The operationalization of the variables for data analysis are described in this
section.
219
Relevant markets: An inquiry of market defense immediately raises the question as to
what constitutes a relevant market in the airline industry. One possible approach would
be to assess the whole nation as the market. It may be argued that even though no carrier
serves every city, there are no such restrictions prohibiting from doing so. However, with
this definition of the relevant market --six major carriers, three of which are about equal
size, and dozens of smaller carriers-- one would be misled to conclude that the industry is
indeed not concentrated. For all practical purposes, considering regional markets as the
relevant market level does not clarify the customer’s dilemma either. Let us consider the
case of the Southeast region: For a passenger who wants to fly from Atlanta to Savannah,
the competitive service that Air-Tran offers on its Atlanta-Orlando flights provide little
relief. A product/service market should have its own elasticity of demand as derived from
consumer preferences. Therefore, from a marketing perspective, the relevant market for
the consumers is the substitutable airport/city pairs. It should be noted that several DOT
reports and recent literature have also taken this view regarding the definition of the
relevant market in the airline industry (e.g., 2001d; Dempsey 2000b; Dixit 2000; Nannes
1999; e.g., Oster and Strong 2001). “We have consistently found that relevant airline
markets are generally no larger than city-pair routes” (1996b, p.5).
12
Pricing Strategy: This was the average yield (Average price / distance) for the incumbent
for the market in question for a given quarter.
Distance: is the physical distance in miles between the two cities that make the market
12
The relevant airline markets may be narrower than city pairs (1996b, p.6). A relevant
market investigation at the segment level (e.g. business class city-pair market, coach class city-
pair market) would be fully justified. Unfortunately, such an inquiry was not possible with the
available data.
220
Market Size: Number of total passengers in the market for a given time frame (i.e.,
quarter).
Market Power: The market power of the incumbent was measured by the Herfindahl-
Hirschman (HHI) index. The index is essentially the sum of the squared market shares of
all firms in a market (i.e., revealing 10000 maximum for a pure monopoly). The primary
advantage of the HHI is that it is more sensitive to the size distribution of firms than a
four firm concentration ratio. HHI performed better than simple market share during
preliminary analysis. An alternative measure to the HHI index was developed but did not
perform better (see Appendix B).
Barriers to Entry: Barriers to entry were formed as a composite variable. The factors that
were considered in the formation of the variable were as follows:
Market-specific barriers: Whether any of the cities that form the market is a hub; is
subject to slot and noise controls; whether the gates are utilized at capacity; whether the
gates are predominantly leased for exclusive use by one carrier (1999a; 1999d; Dresner et
al. 2001).
Firm-specific barriers: Content analysis was undertaken in trade press to identify carriers
with a reputation for predation for a period of ten years (1989-1999, i.e., pre-dating
DOJ’s American Airlines case). Keyword searches were conducted for “predatory”,
“aggressive”, “anticompetitive”, and “antitrust.” The results consistently revealed the
same pattern of frequency associations. American Airlines had its own a cluster with the
highest frequency count (3), Continental, Northwest and United were grouped together
next (2), Delta, TWA, and U.S. Airways (1) were grouped third.
221
Code-share agreements that were in effect for the major carriers were identified
from the DOT reports. Code-share agreement barriers were coded to exist if the
incumbent and one of its code-share partners operated in the same market.
Strategic Assessment: Strategic assessment was the sum of the composite variables for
entrant resources, strategic fits, and market growth.
Resources: This was the sum of the full-time employees of the low cost carriers that
operated at either end of the market (i.e., either city pair). It was transformed into a
composite variable. Alternative financial measures such as resource slack (working
capital over total assets) were employed as well but did not produce satisfactory results.
For example, the working capital for Southwest airlines was consistently negative,
whereas some smaller airlines’ were positive. Therefore, I have opted for employees as a
measure of potential entrant resources.
Strategic fit: This was the sum of the aggregate market shares of the low cost carriers that
operated at either end of the market. This was transformed into a composite variable.
Market growth: For quarter n, this was measured as the percentage of growth in the
number of total quantity of passengers flying in the market:
1 5
5
n n
n
QP QP
QP
? ?
?
?
. This was
transformed into a composite variable.
Entry event: Entries by the following carriers, JetBlue Airways, Frontier Airlines, Tower
Air, AirTran Airways, ValueJet Airlines, Kiwi International, Carnival Airlines, Nations
Air Express, National Airlines, Vanguard Airlines, Spirit Airlines, Pro Air, Reno Air, Sun
Country Airlines, American Trans Air, Western Pacific Airlines, Air South, and Casino
222
Express (list acquired from the DOT officials, identical with that in the DOT special
feature study (2003a)) with more than twenty passengers per day (2000c). Pioneer entries
into markets that did not exist before were excluded. Entries by Southwest airlines were
excluded because of the previously discussed “Southwest effect.” Competitors do not
retaliate against Southwest and Southwest very rarely exits from markets it enters.
Southwest’s market capitalization is more than double the rest of the industry combined
(Hazel 2003, p.13).
Incumbent’s response: This was calculated as the percentage difference between the
average incumbent yield before entry (Q
n-1
), and the average of the incumbent yield at
entry and four quarters following entry (Average Q
n
to Q
n+4
). The signs were reversed so
that they measure retaliation rather than change, which makes it easier to interpret.
Incumbent’s retaliation response for exit hypothesis was the percentage difference
between the average yield of the incumbent in the quarter before exit (or the quarter of
last observation) and the average incumbent yield before entry. Use of average responses
(i.e., average of yields after entry) would not necessarily support hypothesized
relationships for the retaliation–exit linkage because an incumbent may keep prices high
for a few quarters and then decrease it sharply (in multiple markets) to drive the entrant
out (which was an observed pattern in the data). Such predatory actions would not be
captured by an average response measure.
Other variables:
Hubs: Major and regional airline Hubs for the purposes of this research are listed in
Appendix C.
HubPower: Weighed average HHI of all markets linked to the hub.
223
HubPremium: Weighed average yield for the hub
HubSpecificBarriers: These are market specific barriers excluding hub formation.
224
CHAPTER 7
DATA ANALYSIS AND RESULTS
H1: Market power of the incumbent and the market price premium will be
positively associated.
H2: The positive relationship between the market power of the incumbent and the
market price premiums will be positively moderated by barriers to entry.
H3: Barriers to entry and the market price premium will be positively associated.
H4: The positive moderating effect of firm specific barriers on the incumbent’s
pricing strategy will be higher than that of market specific barriers.
H5: The positive effect of firm specific barriers on the incumbent’s
pricing strategy will be higher than that of market specific barriers.
H6: The positive relationship between the market power of the incumbent
and the market price premiums will be negatively moderated by the
incumbent’s strategic assessment (i.e., resources, strategic fit,
market growth) of the potential entrants.
H7: Incumbent’s strategic assessment and the market price premium will be
negatively associated.
These hypothesis were tested using multiple linear regression, logit and HLM
models. Top 4000 markets in the U.S. by passenger volume for the 4
th
quarter of 1999
(the most recent quarterly data available) were chosen for the analyses.
13
This was also
the quarter by which extensive code-share agreements had come in effect. The general
MLR model for hypothesis testing with moderating variables (Sharma 1981) was:
6 6
0 3
1 4
i j ij k i ik i
j k
Y X X X ? ? ? ?
= =
= + + +
? ?
Where Y
i
is the average yield for the incumbent for market i
Xi
1
is the distance between city pairs that form market i (control variable)
13
This represented a comprehensive approach as the top 2500 markets approximately account for
90% of passenger traffic.
225
Xi
2
is market size for market i (control variable)
Xi
3
is market power of the incumbent for market i
Xi
4
is strategic assessment of the incumbent for market i
Xi
5
is firm specific barriers to entry for market i
Xi
6
is market specific barriers to entry for market i
After cases with missing data and outliers were excluded, the final sample
consisted of 3948 markets. A quick preliminary run showed that the main effects existed
in the expected directions. All entered variables were significant (p
The price variable is among the most powerful instruments in the arsenal of the executives to achieve entry deterrence objectives. There are two main pricing strategies that firms may use to defend against a competitive market entry.
THE ROLE OF PRICING STRATEGY IN MARKET DEFENSE
A Dissertation
Presented to
The Academic Faculty
by
Can Uslay
In Partial Fulfillment
of the Requirements for the Degree
Doctor of Philosophy in Management
Georgia Institute of Technology
April, 2005
THE ROLE OF PRICING STRATEGY IN MARKET DEFENSE
Approved by:
Dr. Naresh K. Malhotra, Advisor
College of Management
Georgia Institute of Technology
Dr. Fred C. Allvine
College of Management
Georgia Institute of Technology
Dr. Pat H. Dickson
College of Management
Georgia Institute of Technology
Dr. Jagdish N. Sheth
Goizueta School of Business
Emory University
Dr. Richard D. Teach
College of Management
Georgia Institute of Technology
Date Approved: April 6, 2005
iii
ACKNOWLEDGEMENT
This work is dedicated to my parents, Dr. Yuksel and Nezihe Uslay.
I would like to express my heart-felt thanks to all of my dissertation committee
members. This dissertation would not have been possible without the ample guidance and
support from Professors Fred C. Allvine, Naresh K. Malhotra, Richard D. Teach, Pat H.
Dickson, and Jagdish N. Sheth. I cannot thank them enough, but let me attempt to capture
my gratitude with a few words here. Every student is lucky to have a mentor but I was
lucky to have not one but five of them!
Dr. Dickson provided the kind of constructive and useful feedback every doctoral
candidate needs. Dr. Teach always had time for me when I needed guidance and advice.
Dr. Allvine’s enthusiasm on the use of price in competitive interaction was critical for my
choice of the dissertation topic, and his industry contacts proved crucial in securing the
data. He has always trusted, supported me and served as a role model. Every minute spent
with the great Dr. Sheth has been inspiring and his vision influenced my work and life in
more ways than even he could imagine. Last and certainly the most, I would like to thank
my advisor, Dr. Malhotra. He simply is a true mentor.
I would also like to thank Dr. Goutam Challagalla, Nancy Wong, Charles Parsons,
Robert G. Hawkins, Francis M. Ulgado, Alka Citrin, Koert van Ittersum, and my friends
Tracey King, Brian Murtha, Leslie Vincent, Lan Wu, and Tim Quey who have provided
useful comments, feedback and camaraderie along the way. Naturally, I take full credit
for any errors that may remain in this work.
iv
TABLE OF CONTENTS
ACKNOWLEDGEMENT iii
LIST OF TABLES viii
LIST OF FIGURES xi
SUMMARY xiii
CHAPTER 1: INTRODUCTION 1
CHAPTER 2: THEORETICAL BASE AND CONCEPTUAL
UNDERSTANDING 5
2.1 Competitive Signaling 5
2.2 Resource Advantage (R-A) Theory 8
2.3 Game Theory 11
2.4 Marketing Strategy 13
2.4.1 Reputation effects from a marketing perspective 16
2.4.2 Inter-type and intra-type competition 22
2.4.2.1 Mutual Forbearance 24
CHAPTER 3: AN INQUIRY OF THE USE OF PREDATORY PRICING IN
MARKET DEFENSE: EVOLUTION, REVIEW,
AND MARKETING SYNTHESIS 27
3.1 Introduction to the Review 27
3.2 Pricing 32
3.2.1 Predatory Pricing 36
3.2.2 Predatory Pricing versus Predation 39
3.3 Evolution of the Theory and Practice of Predatory Pricing 41
3.3.1 A Historical Perspective 41
3.3.2 Regulatory Acts Regarding Predatory Pricing 45
3.3.3 Early Years and the Populist Era (1890-1975) 48
3.3.4 Areeda-Turner Era (1975-1982) 49
v
3.3.4.1 Counter Cost and Non-cost Proposals 51
3.3.5 Augmented Average Variable Cost Rule Era (1982-1992) 55
3.3.6 Recoupment Requirement Era (1993- ) 56
3.3.6.1 Brooke Case 57
3.3.7 The Emerging School of Thought 59
3.3.8 Differing Practices among States 62
3.3.9 Insights from European Union Regulation 63
3.4 Critical Cases for the Evolution of Case Law 64
3.4.1 Standard Oil (1911) 65
3.4.2 Utah Pie (1967) 66
3.4.3 Matsushita (1986) 67
3.4.4 Rose Acre (1989) 68
3.4.5 Brooke (1993) 69
3.4.6 American Airlines (2001) 71
3.5 Legal Elements of Proof 73
3.6 The State of the Debate on Predatory Pricing 78
3.6.1 The Chicago School 78
3.6.2 The Post-Chicago School of Thought 84
3.7 Synthesis and a Marketing Perspective 92
3.7.1 The Need for a Marketing Perspective 92
3.7.2 A Focus on Deregulated Industries 93
3.7.3 Epilogue 95
3.7.4 Reaction Framework for Competitive Price Reductions 104
CHAPTER 4: NETWORK PRICE COMPETITION FRAMEWORK 107
4.1 Network Industries 111
4.2 Hypotheses 114
4.2.1 Pre-emptive Defense 114
4.2.1.1 Market Power 114
4.2.1.2 Barriers to Entry 118
4.2.1.3 Strategic Assessment 121
4.2.2 Entry Decision 122
4.2.3 Post-Entry Defense 126
4.2.4 Exit Decision 128
4.2.5 Policy Consequences 129
4.2.5.1 Consumer Welfare 131
vi
CHAPTER 5: CONTEXT: AN INQUIRY OF THE DYNAMICS OF
COMPETITION IN THE U.S. AIRLINE INDUSTRY 137
5.1 Historical Perspective 138
5.1.1 The Propeller (Early) Era 140
5.1.2 The Jet (Post-war) Era 146
5.1.3 Deregulation Era 147
5.1.4 Consequences 151
5.1.4.1 Service and Quality 152
5.1.4.2 Fare Levels 155
5.1.4.3 Market Concentration and Power 159
5.1.4.4 Hub-and-Spoke Networks 161
5.1.4.5 Hub Premiums 164
5.1.5 Marketing 167
5.1.5.1 Yield (Revenue) Management 167
5.1.5.2 Frequent Flier Programs 169
5.1.5.3 The Rise (and Fall) of Travel Agents 170
5.1.5.3.1 Computer Reservation Systems (CRS) 172
5.1.5.4 Discounting and Competition 174
5.1.5.5 Price Wars 181
5.2. Competitive Outlook 185
5.2.1 Industry Snapshot 185
5.2.2 The Role of Barriers to Entry 196
5.2.3 After September 11, 2001 200
5.2.3.1 Anticipating Trends 202
5.2.4 European Landscape 206
CHAPTER 6: DATA AND METHODOLOGIES 208
6.1 Data Set 208
6.2 Methodologies 209
6.2.1 Logit Models 210
6.2.2 Multi-Level Mixed Coefficient Model (HLM) 212
6.2.3 Event-History Analysis 215
6.3 Measures 219
CHAPTER 7: DATA ANALYSIS AND RESULTS 224
CHAPTER 8: CONTRIBUTIONS 242
8.1 Managerial Implications 246
vii
8.2 Public Policy Implications 248
8.3 Limitations and Future Research 251
8.4 Conclusion 256
APPENDIX A: Survey of State Law regarding Predatory Pricing 258
APPENDIX B: Comparison of HHI and Proposed MP Index 260
APPENDIX C: U.S. Regional and Major Airline Hubs 265
APPENDIX D: Gate Usage at Large Hubs 1998 266
APPENDIX E: Transformations and Other Analyses 267
APPENDIX F: HLM Analysis Supplements 272
APPENDIX G: Logit Analysis Statistics 274
APPENDIX H: Event-History Analysis Supplements 276
APPENDIX I: Overview of Relevant Empirical Research in the
Airline Industry 280
APPENDIX J: Graphical Representation of Inter-Type Competition 288
REFERENCES 293
viii
LIST OF TABLES
Table 2.1: The Pedigree of Resource Advantage Theory 8
Table 2.2: Characteristics of Intra-type versus Inter type Competition 23
Table 2.3: Non-competitive Pricing among Intra-type Competitors
in the Airline Industry 24
Table 3.1: A Summary of the Evolution of Predatory Pricing in the U.S. 42
Table 3.2: Criticisms regarding the Areeda-Turner rule and the Chicago School 96
Table 4.1: Dimensions of Consumer Welfare and Marketing Insights 131
Table 4.2: Summary of the Framework Premises 136
Table 5.1: The U.S. Airline Industry Timetable 139
Table 5.2: Airline Mergers in the 1985-87 Period 160
Table 5.3: Single Carrier Market Shares at Major Airports
Pre and Post Deregulation 163
Table 5.4: Changes in Hub Premiums over Time 165
Table 5.5: Sample HHI Index for the 10 largest US Airports 1985-1996 166
Table 5.6: A Comparison of Major Frequent Flier Programs 169
Table 5.7: 1999 Market Share for CRS 173
Table 5.8: Average Fares in Markets With and Without Low-Fare Competition 178
Table 5.9: Average Fares Sorted by Distance 178
Table 5.10: Atlanta Markets with AirTran Entry 179
Table 5.11: Atlanta Markets with AirTran Exit 179
Table 5.12: Airports with Most U.S. Destinations Served by Low Fare Airlines 180
ix
Table 5.13: The Cyclical Nature of the U.S. Airline Industry 186
Table 5.14: Aggregate Industry Snapshot Over time 188
Table 5.15: American Airlines’ Actions Relevant for the DOJ Lawsuit 193
Table 5.16: Percentage of Domestic Air Carrier Slots Held by Selected Groups 199
Table 5.17: Direct Aids to Major Airlines by the Federal Government 200
Table 6.1: Contents of O&D databases 209
Table 6.2: Comparison of Alternative Methodologies 218
Table 7.1: Multiple Linear Regression Results 227
Table 7.2: Main Logit Model Results 228
Table 7.3: Logit Model Results for Alternative Cut-off Points 229
Table 7.4: Main HLM Results 232
Table 7.5: Logit Model Results for Hypotheses 8 and 9 233
Table 7.6: Logit Model Classification Table 234
Table 7.7: MLR Results for Hypothesis 10 235
Table 7.8: Logit Analysis Results for Hypotheses 11 and 12 236
Table 7.9: Bivariate Correlations for Hypotheses 11 and 12 238
Table 7.10: Cox Regression Results for Hypothesis 11 and 12 238
Table 7.11: Summary of Results 239
Table 8.1: Overall Contributions through Key Marketing Concepts,
Methods, and Data Sources 244
Table 8.2: Issues/Research Perspectives on the Use of Pricing Strategy
in Market Defense 255
Table A1: Predatory Pricing Survey 258
x
Table B1: Bivariate Correlations with Full Sample 262
Table B2: Bivariate Correlations with Final Sample 263
Table E1: Correlation Table 270
Table E2: MLR Model results with full sample 270
Table E3: Reduced MLR Model Results 271
Table G1: Results with Dummy Variables 274
Table H2: Event-History Analysis 279
Table I1: Overview of Relevant Empirical Research in the Airline Industry 280
xi
LIST OF FIGURES
Figure 2.1: A Simple Game-Theoretic Construct in the Airline Context 13
Figure 2.2: A Typology of Reputation and Image 18
Figure 2.3: Framework for Assessing the Incumbent’s Reputation for 21
Predation and its Impact on Market Entry
Figure 3.1: Perception of Predatory Pricing under Current Law 38
Figure 3.2: Approximated Predatory Pricing Plaintiff Court Success over Time 44
Figure 3.3: Cost Relationships 50
Figure 3.4: Decision-Making Framework in the U.S. Federal Courts 83
Figure 3.5: Framework for Deregulation and Predatory Pricing Relationship 93
Figure 3.6: Incumbent’s Reaction Framework for Competitive Price Reductions 106
Figure 4.1: Pre-entry and Post Entry Network Price Competition Framework 110
Figure 4.2: Marketing Extension for Consumer Welfare Assessment 130
Figure 5.1: Demonstration of fare level differences 1998Q2 157
Figure 5.2: Demonstration of fare level differences: Top 1000 markets, 2001 Q3 158
Figure 5.3: US Travel Agency Overall Airline Commission Rates
and the Number of Agency Offices, 1976-2000 172
Figure 5.4: New Entry Applications 177
Figure 5.5: Entry/Exit Patterns in the Airline Industry 177
Figure 5.6: A Comparison of Generic and Industry Specific Barriers to Entry 197
Figure 7.1: Graph for Survival Function 237
Figure 7.2: Modified Network Price Competition Framework 241
xii
Figure B1: Comparison Graph 1 263
Figure B2: Comparison Graph 2 264
Figure B3: Comparison Graph 3 264
Figure E1: Histogram for Raw Yield Data 267
Figure E2: Histogram after Ln Transformation on Yield 268
Figure E3: Histogram after Log10 Transformation on Yield 268
Figure E4: Histogram after final transformation 269
Figure E5: Normal Probability Plot for Yield 269
Figure H1: Survival Function 278
Figure H2: One Minus Survival Function 278
Figure J1: Incumbent: Northwest; Route MSP-MCI; Entry: 1995 Q2 289
Figure J2: Market Category comparison (1997) 290
Figure J3: Market Category comparison (1998) 290
Figure J4: Market Category comparison over time (0-200 miles) 291
Figure J5: Market Category comparison over time (200-400 miles) 291
Figure J6: DFW-Wichita Fare Patterns 292
Figure J7: Detroit –Boston 1996 Q1- 1997 Q1 292
Figure J8: Detroit –Philadelphia 1996 Q1 – 1997 Q 293
xiii
SUMMARY
The price variable is among the most powerful instruments in the arsenal of the
executives to achieve entry deterrence objectives. There are two main pricing strategies
that firms may use to defend against a competitive market entry. The first of these
options, limit pricing (or entry deterring price), may be utilized prior to competitive entry.
The second option, aggressive (predatory) pricing, may be executed post-entry. The
effectiveness of both of these options is still controversial. For example, the Chicago
School proponents argue that these strategies are anecdotal in nature. On the other hand,
the rationality of such conduct has been reliably simulated by Post-Chicagoans in game
theoretic settings. The potential contributions of the marketing discipline have been
recognized and called upon to help resolve the conflict.
With this dissertation, I attempt to shed light on the role that price plays in
preemptive and post-entry market defense of firms. As such, the questions tackled
include but are not limited to: how effective is price as an entry-deterrence tool; in
conjunction with firm and market specific barriers to entry; and as a post-entry retaliation
mechanism? What are the facilitating conditions for limit, aggressive (predatory),
competitive and supra-competitive pricing? What are the (long-term) consequences of
these strategies? Following a multi-disciplinary literature review, I present a dynamic
process model and test my hypotheses in a key network industry – the airline industry.
Building upon the advantages of multiple methods a la triangulation, I find that both limit
pricing and predatory pricing can serve as effective strategies for the incumbents’ market
defense. Predatory use of pricing in network industries may diminish consumer welfare.
xiv
Results also suggest that firm specific barriers have a more significant role in market
defense than market specific barriers. Insights and frameworks based on the marketing
philosophy are also presented with the hope of advancing the ongoing debate between the
Chicago and Post-Chicago Schools of thought.
1
CHAPTER 1
INTRODUCTION
The contemporary competitive process resembles military campaigns. Firms need
to advance and protect their strategic positions often at great cost. Conflict with too many
competitors quickly diminishes the scarce resources and depresses the viability of firms.
Resource advantages need to be fortified, exploited, and enhanced before competition can
emulate (cf. Barney 1992). Therefore, the attention to market defense mechanisms has
been increasing (Kuester et al. 1999), and managers consider market entry deterrence to
be a major strategic issue (Smiley 1988). Scholars from economics, industrial
organization, and marketing have acknowledged the important managerial and public
policy implications of market defense and provided guidance for this problem (e.g., Bain
1956; Gatignon et al. 1989; Gruca et al. 1992; Han et al. 2001; Kuester et al. 1999;
Milgrom and Roberts 1982a; Ramaswamy et al. 1994; Sullivan 1977).
A firm may retaliate to competitive market entry with several options such as
increased promotional spending (to consumers or trade members), new product launch
and price reductions, simultaneously or otherwise. There is evidence that as the number
of the marketing mix instruments used (breadth of reaction) increases, market defense
becomes less successful (Gatignon et al. 1997). The most common reaction pattern is
reaction by a single variable (Robinson et al. 1988). The price variable is among the most
powerful weapons in the arsenal of the executives to achieve entry deterrence objectives.
It is inherently flexible, can be modified quickly, and is directly associated with
profitability. In one survey, pricing was deemed to be “extremely important” by seventy-
2
eight per cent of the respondents and ranked third among fifteen key marketing issues (cf.
Srinivasan et al. 2000). Chen and McMillan (1992) reported that the likelihood of
competitive response is higher, the response delay is shorter, and the likelihood of a
matching response is higher for price cuts than they are for other competitive actions.
Incumbent firms often reduce price when they encounter new market entry (Calantone
and di Benedetto 1990).
There are two main pricing strategies that firms may use to defend against a
competitive market entry (LeBlanc 1992). The first of these options, limit pricing (or
entry deterring price (Porter 1980)), may be utilized prior to competitive entry. The
second option, predatory (aggressive) pricing, may be executed post-entry. The
effectiveness of both of these options is still controversial. The Chicago School
proponents argue that they are “more anecdotal than actual” (Gilbert 1989, p.125) or an
outright myth (DiLorenzo 1992; Koller 1971; Lott 1999). This has been the basis that
Federal courts have used for not attributing much credit to unfair pricing claims for the
past three decades. On the other hand, the Post-Chicagoans approach the same facts with
different assumptions (e.g., information asymmetry) and deem the same set of options
rational, alive and well (Jung et al. 1994).
1
Lively debates between Chicago and Post-
Chicago scholars continue (e.g., Edlin 2002; Edwards 2002; Elhauge 2003; ten Kate and
Niels 2002).
1
Price competition is the core element of free markets. Even though it has become much
more complicated during the past two decades, it still remains the variable that is the least under
the control of the firm under (perfectly) competitive market conditions. However, when the
marketplace is not competitive the reverse becomes true and one or few firms may enjoy the
benefits of controlling the going price in a market through the use of their market power.
Ironically, drastic price cuts and price wars that are ingredients for healthy competition are also
associated with unfair competitive conduct (Gundlach and Guiltinan 1998). The abuse of market
power in terms of price manipulation is considered illegal and is subject to antitrust sanctions.
3
The disciplines of economics, law, public policy, and more recently strategic
management have been heavily involved in this stalemate debate between Chicago and
Post-Chicago proponents as to how intense (if any) should the antitrust sanctions be. The
potential contributions of the marketing discipline have been recognized and been called
upon to help resolve the conflict (e.g., Bloom and Gundlach 2001b; Grewal and Compeau
1999; Guiltinan and Gundlach 1996b; Gundlach 1995; Helgeson and Gorger 2003; Ursic
and Helgeson 1994). Foer (2002, p.227) suggested that marketing and strategic
management could serve as “the third leg that gives the antitrust stool stability.”
With this dissertation, I attempt to shed light on the role that price plays in pre-
entry and post-entry market defense of firms. As such, the questions I tackle include but
are not limited to: 1. how effective is sole price as an entry-deterring tool; in conjunction
with other barriers to entry; and as a post-entry retaliation mechanism? 2. What are the
facilitating conditions for limit and predatory (aggressive) pricing? 3. What are the long
term consequences of these strategies? After a thorough multi-disciplinary literature
review, I develop a process model and test my hypotheses using multiple methods. Logit
analysis enables me to examine different pricing strategies, market entry and exit as
dependent variables. Event-history analysis enables me to consider the conditions that
lead to or delay market exits. Finally, multi-level mixed coefficients modeling
(hierarchical linear modeling) allows me to go beyond the framework, take the nested
nature of the data into account, and illuminate the role that different types of barriers play
for entry deterrence. Building upon the advantages of these methods a la triangulation, I
attempt to contribute to the debate between Chicago and Post-Chicago Schools of
thought, and provide empirical ammunition towards this objective.
4
The remainder of this dissertation is as follows: In Chapter 2, I undertake a
literature review of the relevant theoretical base and develop a conceptual understanding.
In Chapter 3, I review and synthesize the literature on predatory pricing from a multi-
disciplinary perspective. The in-depth discussion of predatory pricing here leads to the
development of a network price competition framework i.e., Chapter 4 where I introduce
network industries and develop and present my hypotheses. In Chapter 5, I discuss the
context for my network data --the airline industry, in detail. Chapter 6 presents a review
of the methods and measures I intend to use. In Chapter 7, I analyze the data, present,
and discuss the results. In Chapter 8, I conclude with limitations and future research
avenues. The dissertation also includes ten Appendices in support of the eight chapters.
5
CHAPTER 2
THEORETICAL BASES AND CONCEPTUAL UNDERSTANDING
The relevance of competitive signaling, resource advantage theory, game theory,
and marketing/strategy literature for building a market defense framework is reviewed in
this chapter.
2.1 Competitive Signaling
Signaling theory serves as an important foundation to understand limit and
predatory pricing because both represent (potentially) costly signals to potential and
actual entrants. Limit pricing may signal that the cost structure of the incumbent is low
and/or is willing to protect the market through sacrifice, whereas predatory pricing may
indicate both a low cost structure for the incumbent and/or that the incumbent has deep
pockets and is willing to fend off an entrant at all costs (LeBlanc 1992). Similarly, all
price modifications can convey competitive signals, depending on the amount, timing,
and the context. Distinguishing among temporary, evolving and structural changes in
prices is important because they are likely to produce different results (Srinivasan et al.
2000). Facing a price cut, an incumbent firm may choose to accommodate, to defend
market share or to counter-attack with deeper discounts. The assumption on the degree of
(imperfect) information is key in distinguishing Post-Chicago from the Chicago School
of thought (Lande 1993).
A market signal is an action that conveys indications of intentions, motives, goals
or abilities (Porter 1980). Signals can be directed at customers, channel members,
6
competitors or other stakeholders (Prabhu and Stewart 2000). Firms usually infer
meaning from the signals and decide on their competitive responses. The same signal can
gain different meanings depending on the interpretation of the message. Characteristics of
the signal (i.e., clarity, consistency, and aggressiveness) are considered by the receivers
(Heil and Robertson 1991). For example, a price increase can be interpreted as a response
to market demand or a weakness on the sender’s side (Heil and Walters 1993; Prabhu and
Stewart 2000). Moore (1992) found that managers’ beliefs about the nature of the signal
affected the nature of their responses. If the particular signal was perceived to be
cooperative, the managers were more likely to cooperate and not to retaliate.
Signaling can render predation rational when there is imperfect information
(Hilke and Nelson 1987; Kreps and Wilson 1982; Milgrom and Roberts 1982a; Milgrom
and Roberts 1982b). In many markets, the incumbent is more informed about the
characteristics and conditions (e.g., demand, technology, production) of the market than
the new entrant. The utilization of the asymmetry of information can be predatory if the
incumbent influences the belief and expectations of a rival in a way to alter its decision
regarding entry, exit, price and output levels. This may especially be the case, if the prey
is more efficient than the predator but still decides not to enter a market as a result of
signaling, signal jamming, or reputation (Gundlach 1995). It was also suggested that
when competitive intelligence is not reliable, managers have incentives to overreact when
they face competitive actions (Leeflang and Wittink 1996).
Signal Jamming: Signal jamming refers to the unobservability of a sender’s actions
regarding key variables such as cost. A dominant firm might send signals to current and
7
prospective rivals that its costs are low and the chances for profitable entry are dim
(Milgrom and Roberts 1990). For example, a competitor may alternate its marketing
variables (e.g., sales promotions) to steal the short term demand for a potential entrant’s
product when they are conducting market tests (Gundlach 1995). The potential entrant
may not realize the ploy and decide not to launch the product. False/pre-mature product
announcements (i.e., vaporware) can mislead competitors, potential entrants, and buyers
in their decisions (Bayus et al. 2001; Eliashberg and Robertson 1988; Robertson et al.
1995). Thus, signal jamming may result in an exit, a decision to not enter, or have other
anti-competitive effects (Fudenberg and Tirole 1986; Grout 2000).
Reputation: As will be detailed in a later section, repeated interaction between firms
facilitates the formation of perceptions and beliefs of competitors. A hostile reputation
can be an asset for the firm in the long run (Weigelt and Camerer 1988). A reputation for
predation can deter entrants as well as potential investors of competitors. Burns (1986)
showed that alleged predation depressed the acquisition cost of the victims and others
through reputation.
Case in point: The Department of Justice (DOJ) filed a suit against the Airline Tariff
Publishing Company (ATP) in 1992. DOJ argued that the carriers were able to negotiate
and agree on prices by signaling them through the ATP computer reservation system for a
future date. Another charge was that carriers were able to negotiate the elimination of
discounted tickets via signaling within the system. The case was settled in 1994 after the
airlines agreed to restrictions in using the system (1994b).
8
2.2 Resource Advantage (R-A) Theory
Nonetheless, as an evolutionary, process view of competition, R-A theory
provides a viable starting point from which to debate antitrust issues. It
also provides public policy researchers with many potentially fruitful
avenues for empirical investigation (Hunt and Arnett 2001 p. 23-24)
R-A theory is a dynamic, process theory of competition developed by Shelby D.
Hunt and Robert M. Morgan. It has been developed through more than a dozen articles
over time (e.g., Hunt and Morgan 1995; Hunt and Morgan 1996) and is detailed in two
books (Hunt 2000; Hunt 2002). It draws from evolutionary economics, Austrian
economics, heterogeneous demand theory, differential advantage theory, historical
tradition, industrial-organization economies, resource-based tradition, competence based
tradition, institutional economics, transaction cost economies, and economic sociology
(Hunt 2000). Table 2.1 provides a summary of characteristics that R-A synthesis shares
with other theory.
Table 2.1: The Pedigree of Resource Advantage Theory
Research Tradition Representative Works Affinities with R-A Theory
Evolutionary Economics (Marshall 1890)
(Schumpeter 1934; 1950)
(Alchian 1950)
(Nelson and Winter 1982)
(Langlois 1986) (Dosi et al. 1988)
(Witt 1992) (Foss 1993)
(Hodgson 1993)
Competition is an evolutionary,
disequilibrating process. Firms
have heterogeneous competences.
Path Dependencies can occur.
Austrian Economics (Mises 1920; 1949)
(Hayek 1935; 1948)
(Rothbard 1962)
(Kirzner 1979; 1982)
(Lachmann 1986)
Competition is a knowledge-
discovery process. Markets are in
disequilibrium. Entrepreneurship
is important. Value is subjective.
Intangibles can be resources.
Heterogeneous Demand Theory (Chamberlin 1933)
(Smith 1956)
(Alderson 1957; 1965)
(McCarthy 1960) (Myers 1996)
Intra-industry demand is
substantially heterogeneous.
Heterogeneous supply is natural.
“Product” should be defined
broadly.
Differential Advantage Theory (Clark 1954; 1961)
(Alderson 1957; 1965)
Competition (a) is dynamic, (b) is
both initiatory and defensive, and
9
(c) involves a struggle for
advantages. General equilibrium
is an inappropriate welfare ideal.
Historical Tradition (North 1981; 1990)
(Chandler 1990)
(Landes 1998)
History “counts.” Firms are
entities that are historically
situated in space and time.
Institutions influence economic
performance.
Industrial-organization
Economies
(Mason 1939)
(Bain 1954; 1956)
(Porter 1980; 1985)
Firm’s objective is superior
financial performance. Market
place positions determine relative
performance. Competitors,
suppliers, and customers
influence performance.
Resource-based Tradition (Penrose 1959)
(Lippman and Rumelt 1982)
(Rumelt 1984)
(Wernerfelt 1984a)
(Dierickx and Cool 1989)
(Barney 1986; 1992)
(Conner 1991) (Grant 1991)
(Pfeffer and Salancik 1978)
Resources may be tangible or
intangible. Firms are historically
situated combiners of
heterogeneous, imperfectly
mobile resources. Firms are
constrained by a network of
resource inter-dependencies with
other entities.
Competence-based Tradition (Selznick 1957) (Andrews 1971)
(Hofer and Schendel 1978)
(Hamel and Prahalad 1989;
1994a; 1994b)
(Prahalad and Hamel 1990; 1993)
(Teece and Pisano 1994)
(Day and Nedungadi 1994)
(Aaker 1995) (Sanchez et al.
1996) (Heene and Sanchez 1996)
(Sanchez and Heene 1997)
(Christensen 1997; Christensen
and Bower 1996)
Competition is disequilibrating.
Competences are resources.
Renewal competences prompt
proactive innovation. Firms learn
from competing. Firms are
embedded.
Institutional Economics (Veblen 1899; 1904)
(Commons 1924; 1934)
(Hamilton 1932) (Kapp 1976)
(Neale 1987) (Mayhew 1987)
(DeGregori 1987)
(Ranson 1987) (Hodgson 1994)
Competition is disequilibriating.
“Capital” is more than just
physical resources. Resources
have “capabilities.”
Transaction Cost Economies (Coase 1937)
(Williamson 1975; 1985; 1996)
Opportunism occurs. Many
resources are firm specific. Firm-
specific resources are important.
Economic Sociology (Parsons and Smelser 1956)
(Granovetter 1985; 1994)
(Etzioni 1988) (Coleman 1990)
(Zukin and DiMaggio 1990)
(Powell and Smith-Doerr 1994)
(Smelster and Richard 1994)
(Scott 1995) (Uzzi 1996)
(Fligstein 1996)
Institutions can be independent
variables. Social relations may be
resources. Economic systems are
embedded.
Table 2.1 (continued)
10
Drawing from the above theory, Hunt and Morgan (1997) present the
foundational premises of R-A Theory as follows:
1. Demand is heterogeneous across industries, heterogeneous within industries
and dynamic.
2. Consumer information is imperfect and costly.
3. Human motivation is constrained self-interest seeking.
4. The firm’s objective is superior financial performance.
5. The firm’s information is imperfect and costly.
6. The firm’s resources are financial, physical, legal, human, organizational,
informational, and relational.
7. Resource characteristics are heterogeneous and imperfectly mobile.
8. The role of management is to recognize, understand, create, select, implement,
and modify strategies.
9. Competitive dynamics are disequilibrium provoking; innovation is endogenous.
Of particular interest for my purposes, Hunt and Arnett (2001) argue that R-A is
the remedy for the current antitrust stagnation. For example, they argue that a market
based advantage that stems from a contractual agreement that exclusively ties a
distributor is anticompetitive if the distributor is coerced into the agreement through the
bundling of a complementary product. In this case, the market advantage does not
necessarily come from a relational resource but through market power. They argue that
the focus on neo-classical theory and equilibrium economics has limited the exposure to
outside ideas in antitrust. The focus of Chicago School of antitrust has been solely
economic efficiency (as opposed to social welfare), which cannot be sustained through
the analysis of the static supply and demand curves which actually do not exist, and thus
cannot be calculated (Hunt and Arnett 2001).
R-A is a dynamic theory that also focuses on market segments, comparative and
competitive advantage/disadvantages and enables useful insight for framework
11
development (Hunt and Arnett 2001). It encompasses a wide body of theories (e.g.,
resource-based view (Barney 1986; Barney 1992)) and extant research (e.g., the role of
disruptive technologies (Christensen 1997; Christensen 2001)). Finally, it stresses and
avoids many of the unrealistic assumptions of the neo-classical price theory, which is also
a goal of the current research.
2.3 Game Theory
Isaac and Smith (1985) were among the first to study predation possibilities
through game-theoretic experiments involving decision making under competitive
contexts. They run different versions of a single market design experiment in search of
predatory pricing (including one with sunk costs), yet they did not detect any. However,
Jung et al. (1994) commonly observed predatory pricing in their experiment with
repetitions of a simple signaling game. This game was set up so that the potential entrant
would choose whether to enter or not, and the monopolist would choose whether to fight
or accommodate. Prospective entrants were permitted to observe monopolist’s decisions.
The entrants were better-off if they remained out of the market, unless the monopolist
chose to accommodate. Moreover, Harrison (1988) modified and implemented Isaac and
Smith’s (1985) work to a multiple market setting and found evidence for predatory
pricing. Harrison’s settings were further replicated with modifications and consistent
patterns of predatory pricing were detected in most of the markets. It was shown that
predatory pricing can be reliably simulated “both in stylized signaling games and in rich
market settings”(Gomez et al. 1999) and that it is “alive and well” (Jung et al. 1994,
p.73).
12
Kreps and Wilson (1982) analyzed the incentives of an incumbent to fight an
entrant in order to influence the beliefs of future entrants in a case of multi-market
predation. This work led to the classic work of (Milgrom and Roberts 1982b). According
to the Milgrom and Roberts model, an incumbent firm manipulates its price in order to
signal to a potential entrant, hence exerts influence on the entry decision (Milgrom and
Roberts 1982a; 1982b). In a dynamic signaling game with two-sided uncertainty, it was
shown that when the incumbent expected the entrant to be weak, predatory pricing was
chosen, and when the incumbent expected the rival to be strong, limit pricing was chosen
(LeBlanc 1992). “Predatory pricing is a rational strategy on game-theoretic models of
oligopoly, based on informational asymmetries, that take into account signaling, signal
jamming, and reputation effects” (Burns 1989, p.327). Thus, the entry deterring effects of
a reputation for predation have also been established with game theoretical experiments.
Guiltinan and Gundlach (1996a) suggested that the courts would benefit from the
competitive interaction and strategic decision-making insights. The advancement of game
theory has enabled the study of these complex issues. In that context, acting strategically
meant to consider the expected move of the opponent to come up with the best possible
move. This can be staged as a multistage game in which players intend to make
preemptive moves. Purpose and intent are also considered, and the information needed
for decision making of existing and potential players are described in strategic game
theoretic models. A simple game-theoretic framework is presented in Figure 2.1.
13
Figure 2.1: A Simple Game-Theoretic Construct in the Airline Context
Note: The expected values represent the average fare and the load factors of the incumbent and
the entrant respectively.
2.4 Marketing Strategy
Marketing strategy research (e.g., research focusing on marketing-strategy
content, formulation process, and implementation related issues) and strategy research in
marketing (research focusing on the role of marketing in the formulation of corporate and
business level strategy and knowledge management) are two broad research streams that
are getting intertwined (Bharadwaj and Varadarajan 2004). The determinants of business
performance that the academics have considered to be predominant have evolved over
time. The current paradigm appears to be the resource-based view of the firm also
adopted the by general R-A Theory. Bharadwaj and Varadarajan (2004, pp.222-23)
provide an overview of the evolution of thinking on this topic over time and complement
R-A theory.
Incumbent
Potential
Entrant
Potential
Entrant
Enter
Enter
Do not enter
Do not enter
EV: ($100, 90%;
$100, 85%)
EV: ($150, 80%; 0)
EV: ($180, 65%,
$120; 90%)
Predate
Do not
predate
EV: ($200, 65%; 0)
After reviewing these major theories, I focus on two relevant conceptualizations
from marketing strategy. The first of these works is by Gruca and Sudharshan (1995)
which focused mainly on pre-entry conditions. Their generic entry deterrence strategy
framework consisted of feedback loops that started and evolved around the competitive
environment (cost conditions, demand conditions, history, and legal climate). The loop
consisted of alternative entry deterrence strategies (at functional, business unit, and
corporate levels) leading to the anticipated entry decision, leading to anticipated
consequences for incumbent, leading to choice of entry deterrence strategy, leading to
entry decision, leading to realized consequences for the incumbent. Much of the attention
was devoted to the competitive environment which formed the core of their model. The
current research is focused on a single element (i.e., price) of the functional level entry
deterrence strategies that Gruca and Sudharshan (1995) describe. Therefore, the overall
marketing mix strategy (integrated product differentiation), building switching costs,
brand proliferation, and pre-announcements of new products are beyond the scope of this
research.
The second effort focused on post-entry competitive response options as opposed
to pre-entry deterrence. Kuester et al. (1999) examined and summarized the empirical
contributions to competitive market entry reactions to date and reported five dimensions
for it.
Instrumental: refers to the elements of the marketing mix used for reaction. If retaliation
occurs using the same instrument (e.g., counter product introduction or counter price cut)
reciprocal retaliation is said to occur.
Intensity: the weight of reaction (i.e., the funds allocated for counter promotional
activities.)
Breadth: the number (variety) of marketing instruments used,
15
Time: the speed (or time lag) of reaction.
Domain: refers to the choice of market for counterattack.
It should be noted that Kuester et al. (1999) focused on price and product
retaliation and ignored domain retaliation in the empirical examination.
The understanding of competitive responses was furthered by the “Defender”
model (Hauser and Shugan 1983). Using a zero-sum approach, empirical results have
supported that the optimal response to entry is to reduce price, advertising, and
distribution spending for non-dominant brands; to reduce price but increase marketing
budget for dominant brands (Gruca et al. 1992). This suggests that price reduction is the
de facto optimal response against competitive entries. Relative power theory suggested
that strong incumbents (i.e., those with market power) are expected to retaliate more
intensively and more often than weak incumbents (Kumar et al. 1998). The focus in
recent competitive interaction literature has been on competitive reactions at the retail
level and especially on sales (price) promotions, advertising expenditures, and store
brand sales due to the availability of scanner data. Most advertising and price promotions
do not attract retaliation by incumbents (Nijs et al. 2001). However, when there is
retaliation, it tends to focus on a single variable and often uses the same instrument of
aggression (Steenkamp et al. 2005). Despite the observation that relative levels of price
among competitors explain significant variance in retail strategy (Shankar and Bolton
2004), the clash of the relative prices in inter-type competition (rather than short-term
price reductions) remains a gap in literature.
16
2.4.1 Reputation effects from a marketing perspective
A CEO is ultimately responsible for the growth of a company as evidenced
by its financial performance, its capacity for self-renewal, and its
character. The only way you can measure character is by reputation.
Roberto Goizueta (Goizueta 1995).
If the firm can convince its rivals that it is committed to a strategic move it
is making or plans to make, it increases the chances that rivals will resign
themselves to the new position and not to expend the resources to retaliate
or try to cause the firm to back down. Thus, commitment can deter
retaliation (Porter 1980, p.101).
A vastly unexplored area of research is how to measure the reputation of a firm
and its effects on different constituents in the marketplace. A favorable reputation has
been linked to survival in crisis (Yoon et al. 1993), positive customer attitudes toward the
company’s products and salespeople (Brown 1995), enhanced buying intentions (Yoon et
al. 1993), and choice (cf. Traynor 1983; cf. Weiss et al. 1999).
Corporate Reputation has been defined as the “overall estimation in which a
company is held by its constituents. A corporate reputation represents the “net” affective
or emotional reaction –good or bad, weak or strong– of customers, investors, employees,
and the general public to the company’s name” (Fombrun 1996).
In the marketing domain, Reputation Management has been increasingly attached
to the public relations function. A good reputation may be considered the most important
asset of a company in the long run and can help it survive and even thrive during the
tough times. Reputation is relied on for many aspects of organizational decision-making
at different levels (e.g., choosing a supplier/distributor, promoting an employee).
Similarly, many organizational decisions impact the reputation of the firm, thus attention
17
to this largely ignored topic is necessary. The corporate need for a Chief Reputation
Officer (CRO) position has been proposed to manage the Reputational Capital of firms
(Young 1996). “Reputation is becoming central in the language of strategy and
competition, rather than in the old language of public relations” (cf. Garone 1998). Very
few empirical studies that study reputation effects exist (Landon and Smith 1997; Landon
and Smith 1998).
Often used interchangeably in the marketing domain, both corporate reputation
and corporate image reflect perceptions of an entity. However, they are conceptually
distinct in two main ways. Image summarizes a brand or firm’s identity (Park et al. 1986)
whereas “reputation reflects an overall judgment regarding the extent to which a firm is
held in high esteem or regard. Thus, whereas image reflects what a firm stands for,
reputation reflects how well it has done in the eyes of the marketplace. Image and
reputation are distinct concepts as each can vary independent of the other. A firm can
change its image through positioning, though its reputation remains intact” (Weiss et al.
1999). Moreover, the attractiveness of images is segment specific (e.g., Rolex for luxury),
but a favorable reputation is desirable by all customer segments (Weiss et al. 1999). For
example, Virgin Group’s businesses range from book publishing, radio and television
broadcasting, hotel management to entertainment retail, trading investment, and airlines.
“Virgin is considered the consummate specialist in all things for youth fashion and
fashionability” (Garone 1998) despite the bad publicity about the poor quality and
services of their railroad services (Bower 2000).
Sheth (in personal communication 2004) suggested two typologies to illustrate the
dynamics of reputation. First, he argued that a reputation can be represented in two
18
dimensions: strong/weak and good/bad. Therefore, similar to conjoint analysis (Malhotra
1999), the strength of the attribute (weight of attribute utility) can be multiplied with
reputation score to observe overall reputation (i.e.,
1 1
( )
i
k m
ij ij
i j
U x x ?
= =
=
??
(Jain et al. 1979)).
The same typology can be applied to the concept of image as well. Sheth (2004) also
suggested a 2X2 matrix in which firm image and form reputation form the two
dimensions. This typology is represented on the sample spatial map below (Figure 2.2):
Image +
Reputation +
Apple
Rolex
Enron
Jaguar
Wal-Mart
Nike
Harley-Davidson
Microsoft
Virgin
RJR Nabisco
Campbell
Figure 2.2: A Typology of Reputation and Image
It should be noted that reputation has a somewhat different meaning when studied
from an economics perspective. In this context, reputation has to do with the consistency
to keep promises and sticking to a particular strategy (e.g., a government with a
reputation for committing to a path for money supply) (Rogoff 1989). Expected actions
19
are anticipated through reputation (Evans and Thomas 1997). The rationality of
reputation building and its effects on decision making, bargaining and technology
adoption has been shown in game theoretical agency settings (DeJong et al. 1985;
Dobson 1993; Evans and Thomas 1997; Hendricks 1992; Park 1999). The significantly
positive reputation effects for not expropriating minority shareholders on stock prices and
IPOs have been shown in Finance (Gomes 2000).
Finally, reputation is important to examine from an antitrust perspective where it
gains a different interpretation. Limited work on reputation so far has focused on the
customer’s perspective. The classic definitions of reputation have not considered
competitors as direct constituents of reputation. However, it may be possible for a
monopolist to charge supra-competitive prices due to high barriers to entry, reputation for
predation and other signaling effects. In this context, a reputation for predation is
established by constant signaling of future intention to predate in the face of new entry.
This is best demonstrated by actions in a market. Potentially more efficient competitors in
(other) markets observe the reactions of the incumbent to entry, and often the quick
demise of the previous entrant. As a result, they may decide not to commit the high level
of resources needed to compete against the incumbent. Ceteris paribus, they would rather
enter a market where the incumbent does not have a reputation for predation. Thus, a
valid purpose of predation may be to develop a reputation as a tough competitor
(Comanor and Frech 1993; Kreps and Wilson 1982). Reputation for predation is neither
the only, nor the most effective factor affecting potential entry. However, a reputation for
predation implies strong and constant signaling to all potential competitors for all the
markets a company operates in, thus may be quite influential overall. Similarly, Weiss et
20
al. (1999) concluded that corporate reputation has a strategic influence beyond traditional
approaches. The DOJ stated that the existence of a reputation for predation could be
examined by means of industry surveys and that such a study had not been undertaken
(1999e, p.134).
Repeated interaction between firms facilitates the formation of perceptions and
beliefs of competitors. A hostile reputation can be an asset for the firm in the long run
(Weigelt and Camerer 1988). A reputation for predation can deter entrants as well as
potential investors of competitors. Burns (1986) showed that alleged predation depressed
the acquisition cost of the victims and others through reputation. Areeda and Turner
(1975) admitted that “a demonstrated willingness to indulge in predatory pricing might
itself deter some smaller potential entrants…” The reputation can develop from previous
experience with the incumbent in the market in question for potential entry or in other
markets. Similarly, the reputation may be derived from observing other firms’
competitive interactions with the incumbent in the market in question for potential entry
or in other markets. The anticipated retaliation for specific markets may differ because
the incumbent may have revealed its intent to protect a specific market(s) (e.g., fortress
hubs (Allvine 1996b)) at all costs and not react so sharply to other market entries. I
develop this notion into a framework in Figure 2.3 below:
21
Strategic
Entry
Deterrence
Aggressive Competition Experience
with Incumbent
in Specific Market
Aggressive Competition Experience
with Incumbent
in Other Markets
Observed Aggressive Competition
of Incumbent with Other Firms
in Other Markets
Observed Aggressive Competition
of Incumbent with Other Firms
in Specific Market
Framework for Assessing the Incumbent’s Reputation for Predation
and its Impact on Market Entry
Incumbent’s
Reputation
for Predation
+
+
+
+
+
Profitability
Supra-competitive
Prices
+
+
Rivalry
Familiarity
Size
Figure 2.3: Framework for Assessing the Incumbent’s Reputation for Predation and its
Impact on Market Entry
For example, in the airline context, a carrier “defending its turf” against
encroachment by a start-up carrier in a few markets can create a “reputation for
predation” that deters start-up carriers from entering its many other hub markets; this can
significantly alter the “cost-benefit” predation calculation for a hub carrier in a way
uncharacteristic of most other industries” (Nannes 1999). Frederick Reid, a former Pan
Am and American employee, who recruited to be the President and Chief Operating
Officer for Lufthansa admitted to this phenomenon: “Pan Am was effectively destroyed
by it and American was a winner. American has this reputation as a tough customer…”
(McCormick and Field 1997). “If the prey believes that the threat or promise will be
22
carried out, there is no need for actual predation. Thus, like collusion, the most successful
use of predatory threats or promises is difficult for outsiders to observe”(Comanor and
Frech 1993). Reputations are impressions of actual behavior by the constituents. Yet, the
courts continue to treat reputation effects as “industry folklore” because of a lack of
empirical evidence. The bottom-line implication of the reputation effects literature is that
it is important to consider the reputation effects as a barrier to entry when studying pre-
and post-entry market defense.
2.4.2 Inter- vs. intra- type competition:
Intra-type competition takes place between businesses of similar
cost/organization/service structure (e.g., Goldman Sachs/Merrill Lynch). By contrast,
inter-type competition is defined as competition between businesses with different
structure (e.g., Merrill Lynch/Charles Schwab) (Allvine 1996b). Examples in business-to-
business settings are commonplace as well: In IT outsourcing, IBM, EDS, and Accenture
are intra-type competitors whereas Infosys, TCS, and Wipro (i.e., Indian offshore
providers) would be considered their inter-type competitors. The competition between
large distributors (e.g., IKON for copier and printers) and the local dealers can be
characterized as inter-type. It is typical for intra-type competition to focus on non-price
factors (e.g., facilities, sales assistance, and extended warranties) and for inter-type
competition to focus on discounts. Such structural differences between inter- and intra-
type competition have long been observed (Allvine 1996; Miller, Reardon, and McCorkle
1999). Similarly, competition between Delta and American Airlines is considered intra-
type, while competition between Vanguard and American Airlines would be considered
inter-type competition. It is typical for intra-type airlines to focus on non-price factors
23
(e.g., promotion, frequent flier miles) and for inter-type airlines to compete on price.
Table 2.2 contrasts the two types of competition.
Table 2.2: Characteristics of Intra-type versus Inter type Competition
Intra-type Inter-type
Occurrence
Change
Risks
Established Methods
Innovations
Continuous
Small
Low
High
Minor
Discontinuous/random
Large
High
Low
Major
Changes in efficiency
Demand Effect
Low
Secondary
High
Primary
Leadership Style Managerial Entrepreneurial
Outcome Reduction of Uncertainty Novelty
Source: Dixit (2000)
A major take-away from the retailing literature is that inter-type businesses (e.g.,
general merchandiser (Sears), broad-line specialist (Home Depot), limited-line specialist
(Ace Hardware)) can co-exist and prosper (i.e., number of larger stores are positively
related to the size and number of smaller stores (Miller et al. 1999)). Similar formations
in many business-to-business contexts exist: in machine tools there are industry giants
such as Illinois Tool Works that supply a wide variety of equipment, but there are also
operations that focus on segments of the market (e.g., CNC tool rooms), and finally
smaller shops that only custom-build. In heavy construction equipment, Caterpillar and
Komatsu are the market leaders, but Linkbelt’s specialization in cranes pays-off.
At the absence of inter-type competition price competition typically suffers. For
example, Table 2.3 illustrates the lack of price competition between intra-type
competitors in the airline industry.
24
Table 2.3: Non-competitive Pricing among Intra-type Competitors in the Airline Industry
(Oster and Strong 2001, p.31)
The notion of inter-type versus intra-type competition is also supported by the
notion of mutual forbearance.
2.4.2.1 Mutual forbearance
Mutual forbearance theory implies that the higher the multi-market contact
between the same firms, the lesser the intensity of competition due to increased
familiarity between firms and their ability for deterrence (Jayachandran et al. 1999).
Therefore, gaining competitive intelligence becomes advantageous. Korn and Baum
(1999) observed that empirical evidence has robustly demonstrated that multimarket
contact leads to mutual forbearance.
Mutual forbearance moderates the effect of rivalry through tacit collusion and
leads to improved performance for firms. Increased deterrence (Porter 1980) and
increased familiarity among firms (Baum and Korn 1999) were observed as potential
reasons for this effect (Jayachandran et al. 1999). While direct collusion is illegal and has
25
antitrust consequences, tacit collusion is observed when firms understand the motives and
strategies of each party and implicitly coordinate so that intense competition is avoided
(Jayachandran et al. 1999). Since most discounters that provide inter-type competition are
smaller than their major counterparts they would have less market contact than intra-type
competitors. Therefore, mutual forbearance would not only explain the general intensity
aspect of inter-type competition but also help explain the exception for large inter-type
competitors (e.g., intra-type major competitors do not retaliate against large inter-type
discounters such as Wal-Mart and Southwest Airlines because of the high multi-market
contact with them).
Korn and Baum (1999) found out that firms (in the airline industry) did not
actively increase multi-market contact to achieve mutual forbearance. Perhaps the
potential for increased market contact is limited by resource constraints or regulation.
Moreover, previous studies have generally not differentiated market contacts between
inter- and intra-type competing firms. Jayachandran et al. (1999) argued that the positive
relationship between multimarket contact and the intensity of competition was moderated
by the organizational structure of competing firms, seller concentration, spheres of
influence (i.e., focal market distribution), and resource similarity (i.e., parity).
Mutual forbearance may have antitrust implications through collusion, and multimarket
reactions may have antitrust implications through predation (Jayachandran et al. 1999).
As previously discussed, an incumbent's competitive reputation can also deter market
entry in the context of multimarket competition. Clark and Montgomery (1998) have
experimentally shown that an incumbent's reputation for aggressiveness, but not
intelligence, makes a market less attractive and more risky to a potential entrant. They
26
found that reputation has a stronger effect when the degree of multimarket contact is
high.
The theoretical basis for the use of pricing strategies as a defense mechanism in
competitive interaction was inquired in this Chapter. The theories are linked to support of
specific hypotheses in Chapter 4 (also see Figure 4.4). In the next Chapter, an inquiry of
the actual use and evolution of aggressive pricing is undertaken from a multi-disciplinary
perspective and the contemporary thinking is synthesized with insights from marketing.
27
CHAPTER 3
AN INQUIRY OF THE USE OF PREDATORY PRICING IN MARKET
DEFENSE: EVOLUTION, REVIEW, AND MARKETING SYNTHESIS
3.1 Introduction to the Review
The monopolists, by keeping the market constantly understocked, by never
fully supplying the effectual demand, sell their commodities much above
the natural price, and raise their emoluments, whether they consist in
wages or profit, greatly above their natural rate.
Adam Smith, The Wealth of Nations (1776)
Aggressive competitive conduct by a monopolist, which is beneficial to
consumers, and aggressive exclusionary conduct by a monopolist, which
is deleterious to consumers, look alike.
(Blair and Esquibel 1995)
Pricing has become an increasingly complex and sophisticated marketing activity
over the last few decades. Interestingly, among all marketing mix elements, it is the one
that is the least under the control of the business under perfectly competitive market
conditions. However, when a market is not competitive, the reverse becomes true and one
or few businesses may enjoy the benefits of controlling the going price through the use of
their market power. The abuse of market power in terms of price manipulation is
considered illegal. This chapter focuses on a widely recognized and debated form of price
manipulation –predatory pricing, and describes how the marketing discipline can
contribute to its assessment.
The current stage of the evolution predatory pricing is quite important. The very
mechanism, drastic price cuts, that is associated with predatory pricing also happens to be
at the heart of healthy competition (1993b). The primary objective of public policy
28
making regarding competitive interaction is to distinguish anti-competitive conduct from
those that are pro-competitive (Scherer 1976). Despite the growing economic literature
on predatory pricing, the case law in Europe and the U.S. remains limited (Grout 2000).
This chapter reviews the extensive literature on the history of predatory pricing
and presents insights to a serious debate on whether or not a new set of ground rules for
evaluating predatory pricing antitrust cases should be adopted. The notion that the
Supreme Court’s Brooke decision has established an imperfect standard against the
plaintiffs/new entrants is also examined. The main goals of this chapter are to provide a
historical perspective on the topic, to summarize the new evaluation alternatives available
to the courts, and to present a synthesis that also integrates marketing insights.
Courts generally do not find predatory pricing to be rational and assume that
predatory pricing practices are rare. This emphasis comes from the concern for antitrust
litigation to not disturb the beneficial, competitive price-cutting behavior. At the absence
of absolute market power, the courts simply view the cases as competitive (Sheffet and
Petty 1994). The concept of predatory pricing represents a double-edged sword for the
policy makers in that if not prevented, the price competition that enhances consumer
welfare in the short-run can turn out to be disadvantageous for the consumers with return
of the supra-competitive prices in the long run (Gundlach 1995).
The positive impact of lower prices on consumer welfare is generally accepted.
However, predatory pricing is detrimental to consumer welfare in the long run because
once the competitors exit the market, the predator raises prices with the intention of
collecting supra-normal profits. Moreover, the problem with predatory pricing is not
limited to harm to consumers through the increase of prices back to monopoly levels.
29
Predatory pricing, successful or not, can potentially reduce incentives for investment and
innovation, and prevent new entry or expansion by more efficient firms. There are special
implications for network industries such as telecommunications and software where the
value of the product/service increases along with the number of users. Innovation can be
stifled when predatory prices induce consumers to continue to use an old technology as
opposed to a superior alternative offered by a new entrant (Guiltinan and Gundlach
1996).
Price competition is the core element of free markets. Lower prices and
competition increase the welfare of the consumers and the society in general (Grewal and
Compeau 1999). Yet, if consumers feel that the price they have to pay is unfair, then
social harm may occur (Guiltinan and Gundlach 1996). The line between competition and
anti-competitive conduct needs to be carefully drawn.
It appears that there is an increasing gap between the insights from the modern
economic theory and the enforcement of current judicial policy. Government
enforcement concern is high as evidenced by the DOJ lawsuits. The new economy
requires new rules for the assessment of predation because of the growing importance of
intellectual property (e.g., Microsoft litigation). Increasing market concentration in many
industries and number of mergers, are also of concern (Bolton et al. 2000). There is an
ongoing debate between two camps on how enforcement on predatory pricing should be
exercised.
The Chicago School of thought, with its more established neo-classical theory
economists and free enterprise institutes is supported by corporations and practiced by the
Supreme Court. Since 1993, the Supreme Court requires proof of below-cost pricing and
30
of recoupment of losses suffered during predation to concur a predation case, and no
predatory pricing plaintiff has been able to prevail in the courts (Bolton et al. 2000). The
Chicago School concurs that Predatory Pricing cannot be a logical business practice for a
company. They claim that it is not rational if they price below (short-run) AVC, and
perfectly legal if they are above AVC, and thus predatory pricing claims should be
ignored (DiLorenzo 1992).
Many companies are suffering and trying to get attention to what Post-Chicago
School of academics call blatant use of predatory pricing and harm to the competitive
process. The following scenario is observed in many monopolistic markets: The
incumbent signals the intention to predate once a new player announces entry. If the
entrant is bold enough to actually enter the market, the predator matches the entrant’s
(lower) price in the market and usually increases output. The predator declares an all-
front war against the much smaller (but usually more innovative and efficient) entrant
and does not budge until it is driven out of the market or out of business. As soon as the
entrant has been forced out, the monopolist ignores the newly stimulated demand,
reduces the capacity and increases its prices to levels (sometimes higher than) before the
entry. This pattern has been observed for many cases in the airline industry (1996a;
Allvine 1996b). There are many independent experts and academics, who concur that
predatory pricing can indeed be a viable business strategy for the major player in a
monopolistic market. This view validates the need for attention to this important antitrust
topic. The Post-Chicago view of predation has been on the rise in the nineties during the
Clinton Administration as demonstrated by the U.S. versus Microsoft, and U.S. versus
American Airlines cases. However, it would not be surprising if most pending cases were
31
settled during the administration of President George W. Bush, who stated during his
election campaign that he would only pursue price fixing antitrust cases (Financial
Times, 2000c).
Grewal and Compeau (1999) suggested that marketing researchers have not
engaged in public policy implications of pricing until recently and that a focus on this
issue is long overdue. They argued that developments such as the Internet, global
markets, mega-corporations, and cooperative marketing arrangements created the
necessity of taking a closer look at the pricing and public policy interaction with
consumer welfare in mind. After all, courts consider economic harm to consumers as the
best way of assessing harm to society (Baer 1996).
Guiltinan and Gundlach (1996) argued that marketing was in a unique position to
help form public policy guidelines with comprehensive measurement and modeling
procedures, and that predation and predatory pricing have not been addressed by
marketers until recently. Gundlach (1995) suggested that the marketing discipline had the
potential to further the understanding needed for the development of a more suitable
antitrust policy.
The literature on predatory pricing is overwhelming and dominated by the
disciplines of law and economics. This chapter represents a literature review of the
existing knowledge on this topic. It is an effort to thoroughly understand, summarize, and
then synthesize and present a big picture of the subject from the perspective of cases and
law, and the assumptions and development of the thinking in economics.
32
3.2 Pricing
Price can assume many meanings depending on the specific context. It can mean
rent, tuition, fee, fare, rate, interest, toll, premium, honorarium, dues, assessment,
retainer, salary, commission, wage, even bribe and income taxes (Schwartz 1981).
Merriam-Webster Dictionary defines price as “the amount of money given or set as
consideration for the sale of a specified thing.” It is also defined as the quantity of one
thing that is exchanged or demanded in barter or sale for another or the cost at which
something is obtained. More elaborate definitions have involved the concepts of value
and worth. Price can mean a fair return or equivalent in goods, services, or money for
something exchanged, or the monetary worth or value of something (Mish 1995). Simply,
it is the amount of money the customers have to pay for a product or service (Grewal et
al. 1998).
The marketing mix is defined as the set of controllable marketing variables that
marketers employ to obtain the desired responses from their target markets (Kotler and
Armstrong 1991). Price is one of the key components of the classic “four Ps: product,
price, place, and promotion” grouping of the marketing mix (McCarthy 1960). It has been
cited as the most important component of the mix by marketing executives (1983). Price
has special importance for the marketers due to its inherent flexibility and close
association to profitability. General pricing approaches include cost-based pricing (cost-
plus (mark-up), break-even, and target profit), buyer based pricing (perceived value), and
competition based pricing (going rate, sealed bid, competitive response) (Allvine 1999;
Kotler and Armstrong 1991).
33
The pricing strategy is dependent on the nature of the product in question (e.g.,
innovative versus imitative), and the product mix of the firm. For example, market
skimming and market-penetration are two strategies that can be employed for an
innovative product or service. Products have to be positioned in the marketplace to be
profitable and the following are some of the options that can be employed: product-line
pricing, optional-product pricing, captive-product pricing, by-product pricing, and
product-bundle pricing (Kotler and Armstrong 1991, p.351).
Prices can also be adjusted through the use of discount pricing (quantity,
functional, seasonal discounts, payment terms) and allowances (trade-in, promotional);
discriminatory pricing (customer-segment, product-form, location, time); psychological
pricing; promotional pricing (loss leaders, special-event pricing, cash rebates, low-
interest financing, longer warranties, free maintenance, discounts); and geographical
pricing (FOB-Origin, uniform delivered, zone, basing point, freight absorption).
However, both buyers’ and competitors’ reactions need to be considered before changing
prices (Allvine 1999; Assael 1990).
Price theory in economics defines how the firms should set prices under certain
assumptions to maximize their profits (Pass and Lowes 1994). However, the static nature
of the price theory and its rigid assumptions for cost, price and quantity (product) make it
inapplicable to marketing practitioners. When these assumptions are violated, it becomes
hard to measure the nature of the demand (demand curve) and a profit-maximizing price
cannot be determined. Assumptions that marketers cannot afford to have include, an
unchanging environment, single product firms, and all customers paying the same price
(Allvine 1999).
34
Due to the absence of a single profit maximizing guideline in practice, the
marketers determine their strategy for the product or service before they set the price for
it. There are internal (e.g., marketing objectives, marketing mix strategy, costs,
organization for pricing) and external factors (e.g., nature of the market and demand,
competition, economy, resellers, government) that impact pricing decisions. Marketing
objectives of the firm can include survival, current profit maximization, market-share
leadership, product-quality leadership among others. The marketing mix for a particular
product is closely tied to the marketing mix strategy of the related items that the firm
offers. As most of the dot-com start-ups have painfully discovered, prices must start to
exceed their costs (as effected by their economies of scale and learning curve) at some
point if a company is to survive in the long run. The procedures for pricing and the
flexibility for changing the price also have impact on the pricing decision as an internal
factor (Kotler and Armstrong 1991). It should be noted that the nature of the market and
demand characteristics can be industry specific (e.g., seasonal). Economists have
identified four general types of markets (pure competition, monopolistic competition,
oligopolistic competition, and pure monopoly). The price elasticity of demand is also of
concern. The state of the economy impacts the purchasing power of consumers and has to
be considered. The bargaining power of middlemen and retailers can become important
in making pricing decisions. Finally, laws regarding pricing are also important and
marketers need to make sure that their pricing policies do not violate them. These issues
include price fixing, resale price maintenance, price discrimination, minimum (predatory)
pricing, price increases (ceilings), and deceptive pricing (Kotler and Armstrong 1991).
35
Price fixing refers to price collusion among competitors. It is considered illegal except
when supervised by a government agency (e.g., local milk industry agreements, fruit and
vegetable cooperatives). Resale Price Maintenance problem implies that manufacturers
cannot require that their dealers sell at pre-specified prices. They can only propose
suggested retail prices. They cannot refuse to sell to a dealer or punish the dealer
otherwise because of pricing issues. Regulated price increases refer to certain industries
(i.e., utilities) in a free market economy. Government may also use its influence to
discourage major industry price spikes during shortages or in times of inflation.
Deceptive pricing problem means that the price reduction should not be advertised unless
it is a saving from the usual retail price, not advertise inaccurate factory or wholesale
prices, and not advertise comparable prices for different goods. FTC Guidelines against
deceptive pricing were issued in 1958 (United States 1958). Price discrimination problem
implies that sellers must offer the same price terms for a given type of transaction.
Robinson-Patman Act forbids price discrimination unless the seller can prove that its
costs are different in selling to a particular customer than others (1936). This can usually
be the case when order quantities vary in size. However, the seller needs to prove that
these differences are proportional to the differences of quantities. It can also be justified
if the seller can prove that it is trying to meet competitor’s prices in good faith. Even
then, price discrimination should be temporary, localized, and defensive rather than
offensive (Dalrymple and Parsons 1990). For effective discriminatory pricing, segments
in the market should indicate varying levels of demand. Discrimination does not work if
the low price segment customers can resell the product to the higher price segments.
Similarly, competitors who sell to the higher price segments at lower prices can be very
36
disturbing. This may lead some firms to illegally employ predatory pricing practices to
drive the competition out of the market (Kotler and Armstrong 1991). Thus, last but
certainly not least, minimum (predatory) pricing is an important issue that marketers need
to be aware of.
3.2.1 Predatory Pricing
The best known form of predation is predatory pricing, yet it has currently no
statutory definition (Guiltinan and Gundlach 1996). Areeda and Turner (1975) argued
that predatory pricing occurs when a firm lowers its prices in order to eliminate a current
competitor in the relevant market or to prevent new firms from entering the market:
[P]redation…cannot exist unless there is a temporary sacrifice of net
revenues in the expectation of greater future gains.…Thus, predatory
pricing would make little economic sense to a potential predator unless he
had (1) greater financial staying power than his rivals, and (2) a very
substantial prospect that the losses he incurs in the predatory campaign
will be exceeded by the profits to be earned after his rivals have been
destroyed.
Others such as Professor (then Judge) Posner simply defined predatory pricing as
pricing at a level calculated to exclude from the market an equally or more efficient
competitor (Posner 1976). Viscusi et al. (1995) later commented that, pricing at a level to
exclude a less efficient competitor is naturally what competition is supposed to do. Some
researchers focus on the exclusionary conduct (Grout 2000), and some focus on losses
imposed on others (Sullivan 1977) in their definitions. Professor Baumol’s 1996 article
refined the Areeda-Turner definition:
Indeed, one can, perhaps, define a price to be predatory if and only if it
meets all three of the following conditions. First, the choice of that price
37
must have no legitimate business purpose. Second, that price must
threaten the existence or the entry of rivals that are at least as efficient as
the firm (call it firm F) that has adopted the price at issue (price P). Third,
there must be a reasonable prospect of recoupment of at least whatever
initial costs to firm F were entailed in the company’s adoption of the price
in question, that recoupment taking the form of monopoly profits made
possible by reduction (as a result of price P) in the number of competitors
facing F. (Baumol 1996, p.52)
Hence, Professor Baumol would define a business act as legitimate if the expected
(long-run) net return is positive, and if that return does not depend on the exit of any of
equally or more efficient competitors or the prevention of entry of such firms.
Gregory T. Gundlach, who has published numerous articles on the topic of
predation (e.g., Guiltinan and Gundlach 1996a; Gundlach 1990; Gundlach 1995;
Gundlach and Guiltinan 1998) perceives predatory pricing as a reduction of prices
(usually below cost) with the intention of punishing a competitor or gaining higher profits
in the long run by driving competition out of business (Gundlach 1990). Predatory
pricing can improve a firm’s bottom-line through exclusionary conduct or other anti-
competitive effects in the market. Specific cases of pricing below cost, unlawful price
discrimination and price warring are considered predatory pricing (Grewal and Compeau
1999). The same actions regarding pricing can be interpreted differently, depending on
what the intent of the case is. Pricing below cost with the intent of dumping excess
inventory can be considered legal (Grewal and Compeau 1999).
When the cost concept is built into the descriptions above, a definition which
would generally be accepted by the courts in the U.S. is obtained (i.e., pricing below an
appropriate measure of unit cost (e.g., total or variable) with the intent of driving out
rivals (presumably with shallow pockets), and later raising price above unit cost to recoup
38
losses through supra-competitive pricing. The court definition was shaped through cases
such as Matsushita Electric Industrial Co. v. Zenith Radio Corp.; Cargill Inc. v. Monfort
of Colorado Inc.; and finally in Brooke Group Ltd. v. Brown & Williamson Tobacco
Corp. However, the measure of cost that would distinguish predatory from non-predatory
pricing has not yet been prescribed (McCareins 1996). Figure 3.1 summarizes the current
perception of predatory pricing enforcement in the courts:
Below Cost
Pricing
Recoupment
PREDATORY
PRICING
Intent
Market
Power
Figure 3.1: Perception of Predatory Pricing under Current Law
Barriers
to Entry
Whether a particular conduct is predatory may depend on the jurisdiction in which
a firm does business or whether the suit is brought under state or federal law (McCareins
1996). State and Federal courts use different standards to detect predatory pricing. State
courts are usually more receptive to predation cases and tend to protect the preys (i.e.,
smaller firms) more. On the other hand, the Federal courts have stated that predatory
pricing is “inherently uncertain” (1986b, pp.588-89) and that it is generally implausible
39
(1993b, p.226). Today, intent is at best a secondary concern for the courts in the United
States. Predatory pricing is found only when below-cost pricing exists along with (a
dangerous possibility of) recoupment (1993b). This puts a heavy burden of proof on the
plaintiffs.
Perhaps in one of the most widely quoted statements in predatory pricing history,
the Supreme Court in the Matsushita Case stated that predatory pricing is “rarely tried,
and even more rarely successful” (1986b, p.590). The reason for this sharp conclusion is
discussed later in this chapter under the evolution of predatory pricing section. Similarly,
it will be discussed later that there may be exclusionary and anti-competitive effects of
predatory pricing that are not as obvious as the elimination of a direct competitor. The
following definition however, successfully captures the essence of the diverse nature of
the topic, and hence has been adopted for the purposes of my dissertation:
Predatory Pricing is a price reduction that is profitable only because of
the added market power the predator gains from eliminating, disciplining
or otherwise inhibiting the competitive conduct of a rival or potential rival
(Bolton et al. 2000, p.3).
3.2.2 Predatory Pricing versus Predation
Even though predatory pricing is the most widely recognized form, it is only a
part of the larger conduct of predation. Ordover and Willig (1981) defined predation as
actions that are unprofitable but for their possible contribution to a rival’s exit. There is a
rather wide list of activities that can be considered to be predatory depending on their
impact on societal welfare.
Misleading advertising about a competitor’s product, specifically designing
interfaces for complementary products that are incompatible for a rival, initiating
40
excessive regulatory hearings and/or lawsuits have been noted as predation (Bork 1993).
Cutting supply of an essential output or refusal to provide access to essential resources
for competition (e.g., slots at airports) is also predatory according to the “essential
facility” doctrine (Gundlach and Bloom 1993). Prentice (1996) discussed the predation
liability for fraudulent product announcements that a manufacturer knows will not ever
be launched. In addition, raising rival’s costs through acquisition and sleeping on of
patents, product pre-announcements, useless product modifications, exclusionary market
channel arrangements, market share agreements, bundled discounts, discounts designed
to reduce rivals’ ability to obtain display space (Gundlach 1990), refusals to deal,
mandatory tying arrangements, and other uses of the power in one market to increase the
cost to competitors in other related markets (Meeks 1998) have been reported. Non-price
predation (e.g., reputation, raising rival’s costs) may be used to depress the acquisition
cost of rivals.
Predation is like warfare in that it can only be rational when less expensive
measures fail (Scherer 1980). Indeed, it has been argued that that non-price predation is
more effective than predatory pricing because it is less costly, less risky and more often
successful (Bork 1978). Snyder and Kauper (1991) summarized literature on raising
rivals’ costs, a category of predation. Salop and Scheffman (1987) showed that non-price
predation strategies that increase rival’s costs are more advantageous than predatory
pricing. Not every non-price predation is pre-branded illegal, non-price predation does
not have to force exit to be successful and it does not have market power as a
prerequisite. It may be more cost-efficient for large firms to employ contractual
provisions and other strategies to increase barriers to entry to maintain market power. In
41
the broadest sense, sixteen non-price predation strategies have been reported which
included outright sabotage such as burning down a rival’s plant but also vertical
integration, innovation, and product promotion (1991). The next section discusses the
evolution of the theory and practice of predatory pricing but also reports on non-price
predation where necessary.
3.3. Evolution of the Theory and Practice of Predatory Pricing
3.3.1 A Historical Perspective
Predatory pricing cases have a major role among the anti-competitive practice and
antitrust violation cases (Gundlach 1995). Since the turn of the last century, the courts’
approach to evaluate predatory pricing cases has swung back and forth between being in
favor of the plaintiffs and the defendants. Economic theory and scholarly articles seem to
have played an important role, for better or worse, in shaping the federal court policies.
Especially, the articles by McGee (1958), Koller (1971), and Areeda and Turner (1975)
seem to have influenced the courts to an extent that their impact is still observed today in
their reflection from the 1993 Brooke decision. The early influence of academics ended
the “populist era” of predation enforcement, where plaintiffs won most of the cases. The
adoption of Areeda-Turner rule by the courts as a standard shifted the balance in favor of
the defendants. Predatory Pricing was defined by the courts as irrational business
(1986b), and the neo-classical price theory school continuously presented it as a myth
(DiLorenzo 1992; Koller 1971). Moreover, the 1993 Brooke decision has made it even
harder for the plaintiffs to survive in courts. In this decision, the Supreme Court required
not only evidence of below cost pricing but also of recoupment. In none of predatory
42
pricing cases since the Brooke decision (at least 38 of them), did the plaintiffs prevail
(Carney and Zellner 2000). However, modern view holders such as Guiltinan (1996),
(Bolton et al. 2000), as well as distinguished economists such as Alfred Kahn (1998)
have pointed out that certain assumptions of price theory do not hold in the real world.
Currently, there seems to be a consensus among those who hold the modern Post-Chicago
view of predation that predatory pricing can be a viable and profitable business strategy
especially under monopolistic market conditions (Bolton et al. 2000). This indicates that
a reassessment of the standards that the courts use today may be well justified and
overdue. Milestones of the evolution of predatory pricing and an approximation of
plaintiffs’ success rate in the courts are presented next (Table 3.1; Figure 3.2), followed
by a discussion of the evolution.
Table 3.1: A Summary of the Evolution of Predatory Pricing in the U.S.
Timeline Description
1890
Sherman Act Prohibited contracts or conspiracies that restrain trade
including price fixing, and monopolization.
1910
Clayton Act Defined unlawful uncompetitive behavior and practices
other than monopolization.
1911
Standard Oil Co. v. United
States
Standard Oil found guilty in this classic case of
monopolization. A major component of the case involved
the allegations that Standard Oil had employed predatory
pricing to drive its competitors either out of business or to
force them to sell it to Standard Oil at distressed prices.
Standard Oil was broken up into 33 geographically distinct
companies (Gibb and Knowlton 1965).
1914
Federal Trade Commission
(FTC) Act
Extended Sherman Act in terms of restraints for trade. FTC
authorized to interpret antitrust statutes.
1936
Robinson-Patman Act Price discrimination that lessens competition or promotes
monopoly declared illegal. Protected small business from
price-cutting by large sellers.
1938
Wheeler-Lea Amendment FTC authorized to protect consumers as well as competitors.
1940’s
Strong FTC enforcement Cases are infrequent until 1940’s. More cases observed with
strong FTC enforcement (Bolton et al. 2000).
1936-70’s
The Populist Era Roughly 77% of predatory pricing plaintiffs win their case
(Koller 1971).
1967
In this classical example of the populist era, three producers
in California were accused of charging less for their pies in
43
Utah Pie Co. v.
Continental Baking Co.
Utah than in markets closer to their plants after Utah Pie’s
entry to the market. Court decided in favor of Utah Pie. The
Supreme Court reinstated the jury verdict though this
decision was widely criticized. Justice Stewart argued that
the consumers benefited from lower prices and increased
competition in Utah.
1975
Areeda-Turner AVC Rule From the seminal article proposing a per se AVC standard
for detecting predatory pricing.
1975-80
A Defendant’s Paradise Dramatic change in enforcement –no plaintiff prevailed
during the five years following Areeda-Turner (Bolton et al.
2000).
1982-92
Augmented AVC Era Intent and market structure also considered. Equilibrium was
claimed to be reached at 17% plaintiff success rate (Bolton
et al. 2000).
1986
Matsushita Electric
Industrial Company v.
Zenith Radio Corporation
American television manufacturers sued twenty-one
Japanese corporations that sold televisions in the United
States. Plaintiffs’ argument was that the defendants
conspired to drive them out of the U.S. market with
predatory pricing. The court’s assessment became one of the
most quoted in the following cases to come: “consensus…
that predatory pricing schemes are rarely tried, and even
more rarely successful. . .” (1986b, p.590) Thus, the court
did not find an economic motive for the defendants to
predate. Earlier verdict of the Court of Appeals was reversed
and remanded.
1989
A.A. Poultry Farms, Inc. v.
Rose Acre Farms, Inc.
The defendant’s prices were less than its average variable
costs for a period during (and below ATC throughout) the
price war. Predatory intent of defendant’s executives was
also documented: “We are going to run you out
of…business. Your days are numbered”(1989, p.1398). The
court focused on recoupment and concluded that intent by
itself did not help determine the probability of recoupment,
and in the absence of recoupment, even the most vicious
intent was considered harmless to the system.
1993
Brooke Group v. Brown &
Williamson Tobacco
Brooke Group (formerly known as Liggett) alleged that
Brown & Williamson Tobacco Corporation introduced its
own line of generic cigarettes (that the plaintiff had
pioneered with great success) and used predatory pricing to
stifle price competition in the economy segment of the
national cigarette market. Brooke argued that the defendant
used below cost pricing and offered discriminatory volume
rebates to wholesalers. Brown not only matched the
plaintiff’s retail price but also consistently undercut its
wholesale price. A harsh price and rebate promotion war
took place at the wholesale level that lasted eighteen months.
Ultimately, Brooke gave in and increased its prices. Generic
brand prices increased by 71% and branded cigarettes prices
increased by 39% whereas the costs where roughly constant
(Bolton et al. 2000). Even though below cost pricing and
predatory intent was documented, and the defendant was
found guilty by a jury, the Supreme Court required proof of
below cost pricing, and of recoupment, and subsequently
decided in favor of the defendant.
Table 3.1 (continued)
44
Defendants’Ultimate
Paradise Era Commences
Brooke decision led to excessive summary dismissal by
lower courts. Augmented AVC Era equilibrium destroyed.
Plaintiffs have not prevailed in a case since Brooke.
Summary dismissal is the norm. (Bolton et al. 2000)
1998
DOT Proposed Guidelines Department of Transportation (DOT) realized the strategic
problem and would allow proof of recoupment based on
reputation effects (and would not require proof of below
United States v. Microsoft Microsoft is accused of predatory pricing against Netscape
for bundling its Internet Explorer with its operating system
free of charge, and later settles the case.
1999
United States v. AMR
Corp.
Department of Justice (DOJ) filed complaint against
American Airlines based on strategic and reputation effects
in parallel with DOT guidelines.
2000
Presidential Elections George W. Bush stated during his election campaign that he
would only pursue price fixing antitrust cases (2000d).
DOT switches to case by
case approach
After evaluating thousands of responses from different
parties, DOT dropped proposed guidelines and decided to
take a case by case approach (January).
2001
American Case Dismissed Judge Morton dismissed the case basing on “time honored
rules” (April); DOJ decides to appeal the decision (July).
2002 -
Evolution continues… DOJ appeals the dismissal of American case but does not
pursue the objective aggressively in the aftermath of
September 11 events (the summary dismissal is later verified
by a panel of three judges in July 2003). American Antitrust
Institute observed that this marked “the death of predatory
pricing as a critical antitrust tool in this political climate” (cf.
Foer 2003, p.14). The low cost carriers start to make
headway into bleeding major carriers’ territories. Another
wave of predation against the discounters may be in order
after the major carriers get their act together, however the
discounters (e.g., Southwest, JetBlue, AirTran) now have
deeper pockets. As a preliminary step, the majors launch a
new wave of low-cost versions of themselves (e.g., Song for
Delta, and Ted for United Airlines)…
0
10
20
30
40
50
60
70
80
90
1
9
3
5
1
9
4
5
1
9
5
5
1
9
6
5
1
9
7
5
1
9
8
5
1
9
9
5
P
l
a
i
n
t
i
f
f
S
u
c
e
s
s
%
Figure 3.2: Approximated Predatory Pricing Plaintiff Court Success over Time
Table 3.1 (continued)
45
3.3.2 Regulatory Acts Regarding Predatory Pricing
Predatory pricing is a violation that is monitored by antitrust regulation bodies.
Antitrust regulation is constructed to prevent substandard industry performance
(efficiency) and promote competitive performance and equitable distribution of market
power. The typical targets of antitrust regulation are those industries with oligopolistic
structures with high entry barriers where market power stifles innovation. The following
are renowned Acts that were designed to regulate such behavior:
Sherman Act – 1890: Section 1 of this Act basically prohibited contracts or conspiracies
that restrain trade. The behavior considered illegal (e.g., price fixing, allocating territories
among competitors, tying purchases of one product to another) was defined by courts.
Section 1 makes concerted action illegal regardless of their market impact. If the
conspiracy is proven, the plaintiff does not have to prove actual or potential
monopolization (Hawker and Petty 1996).“Rule of Reason" doctrine, which suggested
analyzing the context of behavior to see if the case displays unreasonable restraint of
trade or legitimate business practice, was used for cases of ambiguous nature (Sullivan
1991). Section 2 declared monopolization and conspiracy to monopolize, as a felony.
Intent was defined further with subsequent legislation (Kovaleff 1994).
Clayton Act – 1914: Defined unlawful anticompetitive behavior and practices other than
monopolization (e.g., price discrimination (Section 2), exclusive dealing and tying
contracts, mergers (Section 7), and inter-locking directorates.) Section 7 was an important
46
antitrust statute in terms of its impact on litigation and structure. Cellar-Kefauver (1950)
made mergers and acquisitions that decrease competition or promote monopoly unlawful
(with respect to Section 7 of Clayton). This Act was intended to prevent unfavorable
market structures. However, labor unions and agricultural organizations (cooperatives)
were exempt (Martin 1959; States 1984).
Federal Trade Commission (FTC) Act – 1914: Extended and overlapped with the
Sherman Act in terms of restraints for trade. It defined certain conduct modes businesses
must compete in (e.g., Unfair Methods of Competition (Section 5)). The FTC was
authorized to interpret antitrust statutes. With the extension of Wheeler-Lea Act of 1938,
FTC was authorized to protect consumers as well as competitors (American Bar
Association. FTC Act Editorial Committee. 1981).
Robinson-Patman Act –1936: Amended Section 2 so that price discrimination that lessens
competition or promotes monopoly was considered illegal. The original Clayton Act had
exempted price discrimination in the form of quantity or volume discounts (1936).
Two main types of price discrimination were recognized. Primary discrimination
included injury to competing sellers, whereas secondary discrimination included injury to
competing buyers. A key case for primary discrimination was the Utah Pie Case (1967).
Pie producers in California, were accused of charging less for their pies in Utah than in
markets closer to their California plants, after Utah pie’s entry to the market. Court
decided in favor of Utah Pie, though Justice Stewart argued that the consumers benefited
from lower prices and increased competition in Utah (Van Cise and McCord 1969). A
47
key case for secondary discrimination was Morton Salt Case (1948). Morton gave
volume discounts to selected buyers. Large chain grocery stores were given the lowest
wholesale prices. Court found against Morton based on Robinson-Patman Act (1948).
It appears as if the Robinson-Patman Act was designed to protect small businesses
against the “chain-store revolution.” Secondary type of price discrimination was enforced
in the courts more, typically in the form of “mom and pop” grocery stores versus the
major food chain. However, today Robinson-Patman Act is interpreted rather strictly, and
is not considered a major factor in food retailing anymore (Dickinson 2003). The
standard for primary-line Robinson-Patman violations has become basically the same as
it is for ordinary predatory pricing under Section 2 of the Sherman Act (Kintner 1979).
Most predatory pricing cases fall under either an attempt or actual monopolization charge
under Section 2 of the Sherman Act or Section 2(a) of the Robinson-Patman Act
(McCareins 1996).
Antitrust legislation is enforced by the Antitrust Division of the Department of
Justice, and by FTC with some overlap. However, most lawsuits are filed by private
parties rather than the government agencies. The cases can be criminal or civil. Most
forms of price discrimination are not illegal themselves, though the market power
exercised to employ price discrimination may be illegal under Section 1 of the Sherman
Act. The Robinson-Patman Act extended antitrust for price discrimination to credit terms,
delivery times, quality, and volume discounts. It was designed to protect small
competitors rather than the competitive process. Robinson-Patman Act is easier to
pursue in courts as it requires proof of a reasonable possibility of substantial injury,
whereas Sherman requires a dangerous probability of actual monopolization. Still,
48
Robinson-Patman Act is not favored by the FTC and Department of Justice (DOJ) which
do not currently enforce the Robinson-Patman Act (Dickinson 2003; MacAvoy 2000).
There was also one Act that specifically regulated marketing behavior within a particular
industry --Agricultural Marketing Agreement Act –1937 (United States 1937).
3.3.3 Early Years and the Populist Era (1890 - 1975)
Very few cases were observed before the 1940’s. This inactive period continued
until the 1936 Robinson-Patman Act. Plaintiffs gained substantial power after Robinson-
Patman which protected smaller firms from price cutting by large ones. FTC initiated a
strong enforcement effort in the early 1940s. More lawsuits started to emerge. Oral and
written statements were used as evidence of intent. It was relatively easy to establish
predatory intent (Koller 1978). Koller (1971) reported that the Federal courts had
identified that predation had occurred in 95 out of the 123 cases. This equals to a 77 per
cent litigated case success rate for plaintiffs. This high rate may be due to the fact that
plaintiffs have won some cases “they probably should have lost. It seems no exaggeration
to call this the populist era of predatory pricing enforcement” (Bolton et al. 2000, p.14). It
was in this context when Justice Sullivan proposed that predatory behavior can be
identified by two non-cost criteria – whether it looks “jarring or unnatural”, and whether
it is aimed toward a particular target rather than an abstraction such as market share
(Sullivan 1977, p.112). Alan Greenspan commented that, “the entire structure of antitrust
statutes in this country is a jumble of economic irrationality and ignorance” (Greenspan
1962).
Despite the cases in which firms were found to employ predatory pricing, there
49
was a lack of economic theory supporting the presence and rationale of it. On the other
hand, Koller’s dissertation titled “The Myth of Predatory Pricing” (1971), and the
literature stemming from this work, was relied upon by the Chicago School and cited by
the courts and influential academics such as Areeda and Turner (1975).
3.3.4 Areeda-Turner Era (1975- 1982)
A new era started to emerge in the 1970’s, in which law’s condemnation of many
forms of predation was criticized. This notion gained significant support among the
Federal judiciary and antitrust scholars (Bernstein 2001), and the legal environment that
favored the plaintiffs came to a halt with the publication of the seminal 1975 Areeda-
Turner article. The Areeda-Turner rule basically stated that a firm should be found guilty
of predatory pricing whenever it sets price less than its marginal cost (i.e., the cost of
material and labor in making the last unit, excluding the startup/fixed costs). It followed
that any price set equal to or above the firm’s marginal cost is non-predatory. According
to Areeda-Turner, competition in an industry would naturally drive prices toward the
marginal costs. Pricing below marginal cost, meant operating at a loss, and was
considered irrational, except for the intention to drive out competitors, which was
considered predatory. After competitors were driven out of business, a predatory
monopolist could potentially recoup its losses by charging supra-competitive prices,
which hurt the welfare of the consumers. Areeda and Turner also recognized that
marginal cost data (i.e., how costs vary with each additional unit of output) were not easy
to compute. In practice, marginal cost was more of a conceptual tool for economists.
Thus, Areeda and Turner suggested the use of a per se standard of Average Variable Cost
50
(AVC) as a substitute for marginal cost. AVC is calculated by identifying those costs that
vary with output, adding them up, and dividing the result by the total number of units
produced (Pass and Lowes 1994). Even though there are differences between MC and
AVC, a firm with prices below AVC is not even covering the variable costs, not to
mention its fixed costs. Deductive reasoning follows that, in the absence of a justification,
a firm with below AVC pricing must be conducting predatory pricing since the only
profits (rationale) would come from rewards through the outcome of predation.
MC
AVC
AC
Price
Output
0
Po
Perf.
Comp .
Demand for
Perfect
Comp etit ion
Q o
Demand for
a monop olist
Po
Mon.
Figure 3.3: Cost Relationships
With the lack of a major alternative theoretical model of predatory pricing to
consider, the courts embraced Areeda-Turner AVC rule and replaced the uncertain
factors adopted previously. As a result, the legal trend was totally reversed against the
plaintiffs. Plaintiffs’ success rate immediately fell down to only eight per cent as opposed
to seventy-seven per cent during the populist era (Hurwitz and Kovacic 1982). Areeda
and Turner also spread the view that predatory pricing is a rare event in practice (Brodley
51
and Hay 1981).
3.3.4.1 Counter Cost and Non-cost Proposals
Despite its increasing popularity in the courts, Areeda-Turner rule was criticized
by economists because it did not capture the strategic factors and long run welfare
effects. Sharp price reductions can also be viewed as market signaling that communicate
threats and sanctions. Alternative cost and non-cost standards were proposed to overcome
the practical drawbacks of Areeda-Turner rule, but none of them could replace the
precedent in court practice (Bolton et al. 2000, p.15):
Alternative Non-cost Standards: Some economists argued that predatory pricing is a
matter of intent, not costs. The price could even be set higher than ATC for a case to be
considered predatory (Shepherd 1986).
Williamson Output Increase Rule. Williamson (1977) suggested that predatory pricing be
evaluated as a long-run strategy, particularly from the perspective of the incumbent firm's
response to entry. Temporary price cuts have negligible benefits and long-term welfare
problems occur when the predator raises prices after competitors exit. Williamson
concluded that his rule would have superior welfare consequences to the Areeda-Turner
cost rule where strategic responses were not taken into consideration. Williamson output
increase rule basically stated that it would be considered predatory conduct if a firm
raised its output significantly within the twelve to eighteen months following a rival’s
entry.
52
Baumol Price Reversal Rule. Baumol (1979) proposed that it is predatory conduct if the
incumbent first decreases the price sharply, but then increases it again after the
competitor exits, and this is not accounted for by a rise in cost or demand.
Alternative Cost Standards: In order to replace Areeda-Turner, other counter cost rules
have also been proposed.
Joskow and Klevorick (Two-stage rule). Under the Areeda-Turner rule, a firm covering
its AVC is considered lawful yet the fixed costs remain unaccounted for in the equation.
Average Total Cost (ATC) includes both fixed and variable costs and it was argued that
pricing below ATC for a considerable period could also be predatory. Thus, even though
the firm may be pricing above AVC, it is still informative to compare the price to ATC.
In one of the more comprehensive alternative proposals to Areeda-Turner, Joskow and
Klevorick (1979) basically argued that a price below AVC is always predatory, and a
price greater than AVC but less than ATC is predatory unless the defendant shows that it
has or had a reasonable justification for the price:
Therefore, the adoption of a strategy of pricing below average variable
cost by a dominant firm confronted with entry is sufficient to demonstrate
predation. A price below average variable cost, and for that matter, a
price below average total cost, could not possibly be sustained in the long
run since, to survive, firms must cover total costs in the long run. A firm
with market power-the ability to control price-would only have an
incentive to impose losses on itself when faced with an entrant if the
promise of future monopoly gains made such a tactic profitable from a
long-run perspective (Joskow and Klevorick 1979, p.252).
In their two-stage test Joskow and Klevorick (1979) would first focus on the
market structure to determine if predation is likely to be successful. This meant a careful
examination of the entry and exit costs. They argued that screening out markets where
53
predatory pricing was unlikely would ensure not discouraging truly competitive price
competition. The assumption was that only firms with market power would have
incentives to exclude rivals through predation. On the second stage, they proposed to test
if price was below cost. They would also determine if Baumol price reversal happened
(price cut, exit of competitor, price rise), and see if corporate predatory intent was
documented.
Average Total Incremental Cost. Incremental costs are defined as the costs that will be
incurred as the result of a decision (e.g., capacity increase). In other words, it is the
difference between the firm’s total cost at the two output levels divided by the change in
output. Average Incremental Cost (AIC) (also known as Average Avoidable Cost
(AAC)), is the unit cost of the added output, or in other words, the cost that would have
been avoided had the additional amount of output not been produced (Baumol and Sidak
1994). As the time frame of the cost measure gets longer, more costs become avoidable,
because there is a greater opportunity for redeploying assets that are fixed in the short
run. Therefore, average long run incremental cost (LAIC) is the incremental cost when all
costs are considered variable, and serves as a natural upper threshold.
Baumol (1996) argued that combinations of the firm’s products had to be
considered for a proper Areeda-Turner test. Accordingly, the price of each product by
itself had to equal or exceed that item’s average avoidable cost. Moreover, any
combination of the firm’s products had to be priced so that the incremental revenue had
to exceed the avoidable cost incurred by that combination of products.
Still, there are a number of instances in which the odor of predation is
strong, as when an entrant airline with its six-plane fleet, operating on
54
almost as many routes, proposes to fly a route coveted by a large
incumbent airline, whereon the latter announces that it will open for
business (for the first time) along each of the most promising of the
entrant’s routes. Analogous examples in which predatory pricing is the
issue are also easily imagined. There is reason to provide the entrant in
such a scenario effective recourse against overaggressive acts by the large
incumbent. Accordingly, the rules for proper execution of an average
variable cost test that are described in this article are designed not to
offer undue protection to the firm suspected of predatory pricing (Baumol
1996, p.52).
Other proposals included that of Posner (1976) (required AVC test, high market
concentration and proof of intent), Scherer (1976) (full rule of reason inquiry with focus
on intent and market structure), and Ordover and Willig (1981) (predatory conduct if
benefits depend on added market power through forced exit).
The following statement summarizes the position of most courts regarding cost
before the decisive Brooke Case (1992b):
If the defendant’s prices were below average total cost but above average
variable cost, the plaintiff bears the burden of showing defendant’s
pricing was predatory. If, however, the plaintiff proves that the
defendant’s prices were below average variable cost, the plaintiff has
established a prima facie case of predatory pricing and the burden shifts
to the defendant to prove that the prices were justified without regard to
any anticipated destructive effect they might have on competitor.
As summarized in the above statement, AVC and ATC were commonly used as
lower and upper thresholds for proving predatory pricing by the courts. However, some
researchers observed that this standard was not appropriate for certain industries (e.g.,
telecommunications) (Grout 2000). Joskow and Klevorick (1979) were the first to
recommend the use of long run average incremental cost as an upper threshold. Grout
(2000) evaluated alternative thresholds and concurred with Joskow and Klevorick in that
LAIC was the most appropriate upper threshold for high fixed cost industries with
55
multiple product catalogs.
3.3.5 Augmented AVC rule era (1982 – 1992)
The lower courts faced with difficulties when they tried to implement the AVC
standard as proposed by Areeda-Turner. Cost was very hard to determine, and there were
many criticisms toward the use of a per se short term cost test. The fact that no single
plaintiff won a case during the five years following Areeda-Turner made the courts “a
defendant’s paradise” (cf. Bolton et al. 2000, p.18).
In the absence of a binding Supreme Court example, lower courts adopted what
was called an augmented AVC rule, which also took intent and market structure into
consideration. High concentration in markets and barriers to entry were recognized as
enablers of recoupment. During the ten year augmented AVC rule period until the Brooke
Decision (1983-1993), the plaintiff success rate increased to seventeen per cent and
probably would have been much higher if out of court settlements were included in that
figure. Bolton et al. (2000) suggested that “a more or less satisfactory equilibrium” was
reached during the augmented AVC era. Dismissal of cases by summary judgment was
common, and cases remained tough to win, yet the occasional excessive jury awards may
have prevented many cases of predatory pricing from taking place.
3.3.6 The Recoupment requirement Era (1993 - )
Recoupment element started to gain importance in the courts during late eighties
as evidenced by three cases leading to the Brooke Decision --Matshushita (1986), Cargill
(1986a), and Rose Acre (1989). The reasoning used in the Matsushita Case required that
56
plaintiffs prove not only below cost pricing, but also that the alleged predator either
recouped its losses or at least had a very high probability of recoupment. Recoupment
would be through succeeding in driving competitors out of the market, restricting output,
and raising prices (Sheffet and Petty 1994). During the same year, the verdict of the
Cargill Case, in which the plaintiff had alleged that a proposed merger had antitrust
implications by enabling the defendants to implement a price-cost squeeze strategy
against the plaintiff, was reversed. The Supreme Court held that the proposed merger did
not constitute antitrust harm and that the Court of Appeals had erred (1986a). In the Rose
Acre Case, the court decided to analyze the possibility of recoupment before analyzing
below cost pricing. The rationale was that the determination of likelihood of recoupment
was easier than undertaking the cost analyses. It would not be necessary to analyze below
cost pricing if recoupment was not considered to be possible (Calvani 1999). The
recoupment element evolved into today’s final and strict form with the critical Brooke
case verdict.
The Brooke decision established a framework in which the plaintiffs were
required not only to prove below-cost pricing, but also recoupment of losses suffered
during predation. No plaintiff has been able prevail in the courts since the Brooke
Decision (Bolton et al. 2000). However, if a company can ever prove that an incumbent
used below-cost pricing and also intended to recoup its losses later through supra-pricing,
than the courts will rule that this was illegal conduct because the incumbent has harmed
the welfare of the society by diminishing the efficiency of the allocation of resources
across the society (Guiltinan 1996). Yet, the Supreme Court did not define what an
appropriate measure of cost is, or elaborate on the issue of recoupment sufficiency.
57
3.3.6.1 Brooke Case
A brief summary of Brooke is provided here and the case is discussed in a later
section. The market in the Brooke Case consisted of six main players with little price
competition, resulting in high profitability. The plaintiff was the first to introduce a
generic and less expensive product. As a result of the introduction of this new brand
category, smokers of both the plaintiff’s and the competition’s brand switched to the less
expensive product, increasing the “generic” sales. The defendant (third largest player
with 12% market share) introduced counter generic cigarettes with wholesale prices even
lower than that of the plaintiff’s. A price war took place during which the plaintiff
claimed that the defendant sold its generic product at a loss. It was claimed that the
defendant was trying to force the plaintiff to increase its generic product prices in an
effort to minimize the cannibalization of the defendant’s premium brand and maintain its
profitability. The jury found the defendant guilty, however that verdict was overruled,
reasoning that recoupment was not probable (1993).
First a plaintiff … must prove that the prices complained of are below an
appropriate measure of its rival’s costs … second … is a demonstration
that the competitor has a reasonable prospect [under Section 2a of the
Robinson-Patman Act], or under Section 2 of the Sherman Act, a
dangerous probability, of recouping its investment in below-cost prices
(1993, pp.4702-703).
Starting with Brooke, recoupment was perceived to be the ultimate objective of
predatory pricing. Unless recoupment was probable, the general view held that price
wars (and predatory pricing attempts that failed) were beneficial for the consumers.
The court admitted that below-cost tests might not always be effective in determining
whether predatory pricing was employed. Nevertheless, it chose to dismiss the
58
possibility in its administration, arguing that it would be beyond the practical ability
of any jurisdiction to control:
As a general rule, the exclusionary effect of prices above a relevant
measure of cost either reflects the lower cost structure of the alleged
predator, and so represents competition on the merits, or is beyond the
practical ability of a judicial tribunal to control without courting
intolerable risks of chilling legitimate price-cutting. (1993 p. 223)
With Brooke, the Supreme Court also welcomed the idea of dismissals by
summary judgment in the absence of proof of below-cost pricing and recoupment. The
lower courts embraced the invitation and have dismissed a vast majority of the predatory
pricing cases since Brooke.
3.3.7 The Emerging School of Thought
It has been noted that Brooke decision was the most important predatory pricing
decision of the modern times. It set a milestone in jurisdiction, and was particularly
crucial, as it seemed to destroy the satisfactory equilibrium that was achieved with the
augmented AVC rule. It is important to note that almost all predatory pricing cases have
been dismissed by summary judgment and that no plaintiff has prevailed in the courts
since Brooke. The reason may be that the Supreme Court encouraged lower courts to
dismiss predatory pricing cases by summary means (at the lack of market concentration,
entry barriers, and capacity to absorb additional market share). Another explanation could
be the Supreme Court’s skepticism that predation can be a rational business strategy and
its disregard of modern theories of predatory pricing. Today, the survival chances of a
predatory pricing case beyond the summary judgment phase are slim (Bolton et al. 2000).
However, the fact that reputation effects were considered plausible in at least one case
59
(1995) was a positive development. The courts may start to integrate the thinking from
modern theories into the jurisdiction in the near future. DOJ’s civil complaint against
American Airlines was also based partially on reputation effects.
The proof requirement for recoupment of losses as well as below-cost pricing put
a heavy burden on plaintiffs to prove their case. The Supreme Court "sent a chilling
message to predatory pricing plaintiffs" with the Brooke decision (Denger and Herfort
1994). Following Brooke, it became very difficult to successfully prove a predatory
pricing case in most markets either with below-cost or recoupment elements (Meeks
1998). Proof of recoupment was a legal element additionally required by the Brooke
Case.
The critical Brooke Decision was not without its critics. Bolton et al. (2000)
argued that the notion that predatory pricing was “rare and implausible” stems from early
and old economic theory and that the new approaches have to be used to assess predatory
pricing cases. They argued that the Supreme Court failed to take a strategic approach to
develop an understanding of the case. The defendant successfully deterred aggressive
pricing for an extended period in the future. Aggressive pricing was also effectively
prevented in other markets (e.g., branded cigarettes).
The limits of the existing enforcement stemming from earlier decisions including
that of the Brooke Case were increasingly criticized. For example, Guiltinan and
Gundlach (1996b, p.88) explained that predatory pricing litigation relied on the narrow
assumptions of neoclassical price theory. They questioned the underlying assumptions of
the theory, which are exclusive motivation for profit maximization, possession of perfect
information, and calculated rationality in decision making. They also questioned whether
60
or not the welfare of the consumers should be judged solely on the basis of allocative
efficiency. Ordover and Willig (1981) suggested that opportunity cost of capacity
decisions should also be considered. The lack of strategic theory approach and lack of
consideration of game theoretic models were also criticized (Bolton et al. 2000).
Predatory conduct in many industries and the inability of neo-classical theory to detect
and deal with it became increasingly apparent. Growing literature on the inability of the
courts to deal with the problem (Bolton et al. 2000), and the observation of anti-
competitive affects of predation on the airline industry (Oster and Strong 2001), led DOT
to draft and propose guidelines for governing the airline industry.
DOT Guidelines: The traditional assumption of the Chicago School does not hold well in
industries where there is a high fixed cost structure. That was the reason why DOT felt
the need to propose a set of guidelines to govern anti-competitive conduct in the airline
industry in 1998. Interestingly, DOT guidelines allowed for reputation effects (expected
gains from deterring future entry by competition) as evidence for predatory pricing, and
did not require proof of below-cost pricing (1998d). In essence, DOT was trying to
update the rules that defined anticompetitive conduct for the airline industry. These
guidelines are detailed in Chapter 6.
DOJ Efforts: DOJ Antitrust Division Section Chief Roger Fones announced that they
would not be using Areeda-Turner (AVC) test, but would be focusing on a measure of
costs that the alleged predator “could have avoided had it not embarked upon the
pricing/capacity strategy under review” (Fones 1997). It was stressed that the division
61
would not look into the very short-term average avoidable cost, but would have a short to
medium run perspective. He made many remarks regarding the airline industry including
“a popular misconception that predation cannot occur in the airline industry because the
avoidable cost of filling an empty seat is low” (Fones 1997). Around the same time as
this announcement, DOJ also took a proactive stance in detecting and preventing antitrust
violations by suppliers. They communicated with general counsels of corporations
offering to make presentations on bid rigging, price fixing, and market allocation, and
distributed a booklet named "What You Should Know About Detecting and Preventing
Antitrust Violations, Antitrust Primer for Procurement Officials" (1997b).
DOJ filed a civil complaint against American Airlines based on reputation effects
in parallel with the DOT guidelines in 1999. American Airlines case was the first DOJ
brought in more than 20 years. It was seen as a tough case because of the very high
standards of proof imposed on the plaintiffs since the Brooke Decision (Carney and
Zellner 2000). The case is detailed in Chapter 6.
3.3.8 Differing Practices among States:
The Federal courts generally embrace the notion arising from the early work of
McGee (1958), Koller (1971) and Areeda and Turner (1975) that predatory pricing is
irrational. Since the Brooke (1993) decision, the Federal courts have required proof of not
only below-cost pricing but also recoupment of losses suffered through such pricing
(1993).
State courts have generally been generally less demanding and more hospitable to
predatory pricing plaintiffs than the federal courts (Sheffet and Petty 1994). For example,
62
Wal-Mart was found guilty in an Arkansas district court even after the Brooke decision,
in a verdict that was later reversed by higher courts (Hawker and Petty 1996). Many
states have laws to treat predatory pricing in form of sales below cost statutes or
minimum mark-up laws (Haynes 1988). These laws can be generic in nature or may
address concerns for competition in specific markets such as gasoline sales. Some states
require proof of below cost sales, and some others still adopt a focus from the populist
era such as of harm to competitors, rather than harm to competition (Calvani 1999). Cost
criterion used by the states is usually Average Total Cost (ATC) or a form of AVC or
ATC. State courts infer intent of predatory pricing action either from the below-cost
behavior or from other means. The concern for bankruptcy of small competitors is higher
at the state level than that at the federal level (Guiltinan and Gundlach 1996). Some
states use a sliding scale requiring differing levels of proof depending on how much the
price was below ATC (Bruckmann et al. 1995). Appendix A includes an unexhaustive
survey of the states regarding their antitrust laws with respect to predatory pricing.
3.3.9 Insights from European Union Regulation:
Even though the outcomes of the U.S. litigation affect the rest of the world and
the U.S. appears seemingly unaffected from other country influences, it is still
worthwhile to take a brief look to how a continent with a long industrial history is dealing
with the predatory pricing issue.
Europe appears to be more aggressive in its antitrust enforcement. Whether above
AVC pricing can be predatory is debated. The European Commission comfortably
employed AIC cost but one with a longer frame in order to not set the floor too low for
63
predatory claims. Proof of recoupment is not required, hence a defendant may be found
guilty even when it had no reasonable prospect of recouping losses made through
predatory pricing (1994c).
The European Commission also recognizes the uniqueness of network industries
due to the high overhead involved. For example, there is a general understanding that the
average variable cost rule does not normally apply to the telecommunications industry
since “the variable costs of providing access to an already existing network are almost
zero” (1998b). Similarly, the 1998 Competition Act acknowledges that AVC tests may
not be relevant for regulated industries since their variable costs may be close to zero.
There is a general agreement among the policy-makers that alternative approaches may
be necessary for such industries. Long run incremental cost (LRIC) is considered to be an
acceptable test for such industries as it includes both capital and operating costs of the
increments. As long as the price is above LRIC, the decision will be profitable, hence
rational and non-predatory. In particular, LRIC is the accepted test for the
telecommunications industry in particular. As for the time frame, neither the very short,
nor the very long time frame is considered appropriate. However, the Commission has
admitted that it may have to examine AIC of longer than a year (1998b).
Moreover, the Commission has employed and is supporting the use of
combinatorial cost tests for the telecommunications sector. Combinatorial tests can be
defined as a sequence of tests where the revenue earned from each service or combination
of services must cover the total incremental cost of adding that collection of services to
the remaining services (2000a). Unfortunately, combinatorial tests can be very hard to
calculate since it requires all combinations to be considered (2000a).
64
3.4 Critical Cases for the Evolution of Case Law
The evolution of predatory pricing can probably be best observed through a study
of important cases through history. The following cases were selected since they
represent a summary of the evolution of jurisdiction over time.
3.4.1 Standard Oil Co. v. United States, 221 U.S. (1911)
Standard Oil of the Rockefellers was found guilty in this classic case of
monopolization. A major component of the case involved the allegations that Standard
Oil had employed predatory pricing to drive its competitors either out of business or to
force them to sell it to Standard Oil at distressed prices (1911). It had aggressively
purchased 223 independent companies prior to 1907 (McGee 1958). As a result of the
case, it was broken up into thirty-three geographically distinct companies (Gibb and
Knowlton 1965).
Standard Oil was accused of engaging in predatory pricing (pricing below ATC)
to drive competitors out of business and then raising price above ATC to recoup losses.
Standard was also accused of buying key input suppliers (i.e., pipeline companies) to
control the market, of using its market power position for negotiating discriminatory rail
freight rates, and of engaging in business espionage. Even though McGee (1958) later
disputed the verdict of the case and argued that no predation actually took place, other
researchers have shown that subtle predatory tactics were employed: United Rail charged
cartel prices to the whole oil industry except for Standard Oil which got rebates. Standard
Oil and the railroads shared the cartel freight profits. By effectively raising the cost of its
65
rivals, Standard Oil was then able acquire them at distressed prices (Granitz and Klein
1996). Predatory pricing coupled with acquisition is more viable because the period
during which the predator incurs losses is shortened and market power is further
enhanced. Celler-Kefauver Amendment (1950) to the Clayton Act section 7 prohibited
the creation of monopoly through horizontal merger.
3.4.2 Utah Pie Co. v. Continental Baking Co, 386 U.S. (1967)
In this classical example of the populist era, Utah Pie, the leading vendor of
frozen pies in its market, brought a suit under Section 2(a) of the Robinson-Patman Act,
against three national bakeries alleging that they had engaged in a geographically focused
predatory pricing campaign to increase their local market share. The court concluded that
the defendants’ charged less for their pies in Utah than elsewhere. The plaintiff’s market
share had diminished from 66.5% to 45% during the forty-four month price war. The
Supreme Court reinstated the jury verdict for the plaintiff (1967; Calvani 1999).
Interestingly, the sales volume of the plaintiff had increased and the firm had
remained profitable throughout the price war. The court suggested that ATC could be
considered an appropriate standard for below-cost. It also “left open the possibility that
prices above cost may be predatory if coupled with evidence of anticompetitive intent
and a deteriorating price structure. The Court did not comment on whether the evidence
in that case would support a predatory pricing claim under the Sherman Act” (McCareins
1996).
The Utah Pie decision was widely criticized. Nevertheless, it marked the high
time of early era predation litigation. The winning predation cases of the time were
66
characterized by the large predators, geographic price discrimination, sales below average
total costs, and predatory intent. Dissenting Justice Stewart then observed about the Utah
Pie Case that “if we assume that the price discrimination proven against the respondents
had any effect on competition, that effect must have been beneficent. The Court has
fallen into the error of reading the [statute] as protecting competitors, instead of
competition” (cf. Calvani 1999, p.5).
3.4.3 Matsushita Electric Industrial Company v. Zenith Radio Corporation 475 U.S
(1986)
In 1986, American television manufacturers sued twenty-one Japanese
corporations that sell televisions in the United States. Plaintiffs’ argument was that the
defendants conspired to drive them out of the U.S. market with predatory pricing. The
court’s logic, and assessment of the Matsushita case became one of the most quoted in
the following cases to come:
Any agreement to price below the competitive level requires the
conspirators to forgo profits that free competition would offer them. The
forgone profits may be considered an investment in the future. For the
investment to be rational, the conspirators must have a reasonable
expectation of recovering, in the form of later monopoly profits, more than
the losses suffered. . . .[T]he success of such schemes is inherently
uncertain: the short-run loss is definite, but the long-run gain depends on
successfully neutralizing the competition. Moreover, it is not simply to
achieve monopoly power, as monopoly pricing may breed quick entry by
new competitors eager to share in the excess profits. The success of any
predatory scheme depends on maintaining monopoly power for long
enough both to recoup the predator's losses and to harvest some
additional gain. Absent some assurance that the hoped-for monopoly will
materialize, and that it can be sustained for a significant period of time,
"[t]he predator must make a substantial investment with no assurance that
it will pay off." . . . For this reason, there is a consensus among
commentators that predatory pricing schemes are rarely tried, and even
more rarely successful. . . . (underline added) These observations apply
67
even to predatory pricing by a single firm seeking monopoly power
(1986b).
Following the above argument, the Supreme Court concluded that the defendants
did not have a significant chance to achieve monopoly power and subsequently raise
prices in the U.S. market. The court also did not find that the industry had high barriers to
entry. Since the court did not find an economic motive for the defendants to predate, it
demanded the plaintiffs to provide more persuasive evidence. Earlier verdict of the Court
of Appeals was reversed and remanded. An appropriate measure of cost was not defined
(McCareins 1996).
3.4.4 A.A. Poultry Farms, Inc. v. Rose Acre Farms, Inc. 881 F.2d (1989)
The case was regarding a price war among egg producers. It was testified by the
plaintiff’s expert that the defendant’s prices were below its average total costs during the
war. The defendant’s prices had also been less than its average variable costs for a period
during the price war. Predatory intent of defendant’s executives was also documented:
“We are going to run you out of…business. Your days are numbered.” (1989, p.1398).
The court stated that it was much easier to conclude from the structure of the
market that recoupment was not probable, than it is to determine the appropriate measure
of cost and measure cost. Thus, the court decided to first analyze the probability of
recoupment before applying a cost test. If recoupment was not possible then it could be
inferred that the low price was not predatory even if it were below cost:
Predatory prices are an investment in a future monopoly, a sacrifice of
today’s profits for tomorrow’s. The investment must be recouped. If a
monopoly price later is impossible, then the sequence is unprofitable and
we may infer that the low price now is not predatory. More importantly, if
there can be no “later” in which recoupment could occur, then the
consumer is an unambiguous beneficiary even if the current price is less
68
than the cost of production. Price less than cost today, followed by the
competitive price tomorrow, bestows a gift on consumers. Because
antitrust laws are designed for the benefit of consumers, not
competitors…, a gift of this kind is not actionable. (1989, p.1401)
The court also concluded that intent by itself did not help determine the
probability of recoupment, and in the absence of recoupment, even the most vicious
intent was harmless to the system: “Entrepreneurs who work hardest to cut their prices
will do the most damage to their rivals….If courts use the vigorous, nasty pursuit of sales
as evidence of a forbidden “intent”, they run the risk of penalizing the motive forces of
competition” (Calvani 1999, p.8). Not surprisingly, the plaintiffs did not prevail in this
case.
3.4.5 Brooke Group v. Brown & Williamson Tobacco 509 U.S. (1993)
The Supreme Court decision regarding the Brooke case has been considered to be
“the most important predatory pricing decision in modern times.” No single plaintiff has
won a case of predatory pricing in the Federal courts since the Brooke Decision was
made (Bolton et al. 2000). The Supreme Court developed a two-stage proof framework
for analyzing predatory pricing claims during the Brooke Case that no plaintiff has been
able to survive since (Bolton et al. 2000).
Brooke Group (formerly known as Liggett) alleged that Brown & Williamson
Tobacco Corporation introduced its own line of cigarettes in the generic segment, which
Brooke Group had pioneered, and used predatory pricing to stifle price competition in the
economy segment of the national cigarette market. Liggett argued that the defendant used
below cost pricing and offered discriminatory volume rebates to wholesalers. The
69
allegation was that the defendant conspired to force Liggett to raise its own generic
cigarette prices so that it could introduce oligopoly pricing in the economy segment.
R.J. Reynolds had 28 per cent and Philip Morris 40 per cent market share at the
time of the trial. The defendant was a distant third with about 12 per cent. The plaintiff’s
share (once 20 per cent at its peak) was just around two per cent in 1980 and around five
per cent in 1984. On the verge of bankruptcy, the plaintiff introduced low-cost generic
(black and white) cigarettes in 1980. This new category of cigarettes was around 30 per
cent less expensive and it was an immediate success.
As the success of the generics became apparent, larger firms responded. The
defendant was the manufacturer that was hit the hardest. Even though they sold around
11% of the branded cigarettes, their customers were the most price-sensitive and 20% of
them had switched to plaintiff’s generic cigarettes. The defendant introduced its own
generic cigarettes in 1984. The defendant was not the first to respond to Brooke’s generic
cigarettes, indeed R.J. Reynolds had already repositioned one of its existing brands as
generic by 1984 (1993).
However, the defendant not only matched the plaintiff’s price at the retail level
but also consistently undercut it at the wholesale level. A harsh price and rebate
promotion war took place at the wholesale level. After the war, the plaintiff gave in and
increased its prices. Generic brand prices increased by 71% and branded cigarettes prices
increased by 39% (Bolton et al. 2000).
The Supreme Court held that Section 2 of the Sherman Act is not violated unless
plaintiff proves not only that defendant's prices were ‘below an appropriate measure of …
costs, but also that defendant had a dangerous probability, of recouping its investment in
70
below-cost prices” (1993). Unable to meet the burden of the increased level of proof, the
plaintiff lost the case despite strong evidence of predation. The defendant had cut its
prices below costs, had predatory intent, had decreased its output and increased prices
following the period of price war even though its costs were pretty much constant. The
parties in Brooke had both agreed that AVC was the appropriate standard therefore the
Court did not discuss the issue further (Watson 1998).
The Brooke case brought clarity to a couple of issues to the disadvantage of the
plaintiffs: below cost pricing would be a prerequisite to predatory pricing (it followed
that a firm can not be held liable of predatory pricing if its prices are above its costs), and
proof of recoupment (in form of a reasonable prospect or a dangerous probability) would
be required. Condemning above-cost price cuts would be “beyond the practical ability of
a judicial tribunal to control without courting intolerable risks of chilling legitimate price-
cutting” (1993). Unfortunately, this message enables the incumbents in network
industries (which have very low AVC structure) to employ predatory pricing legitimately
with above (variable) cost prices.
The logic for the recoupment aspect was as follows: the predator needs to recoup
its losses of predation to be profitable in the long run (i.e., rational), if predatory pricing
does not result in elimination of the rival, then recoupment cannot occur. If there is no
recoupment and the competitor is not eliminated, there is no harm to competition and
consumers have simply benefited from the low prices. If recoupment is not possible then
summary dismissal of a case should be appropriate. This view also shared by the Chicago
School of thought, has its flaws, discussed in the Post-Chicago School of thought section
of this chapter.
71
3.4.6 United States v. AMR Corp. et al. (American Airlines) (1999)
--dismissed (2001)
The 1999 DOJ suit against American Airlines was the first predatory pricing suit
brought forward in more than twenty years (Carney and Zellner 2000). Following the
DOT guidelines in spirit, the prosecutors wanted to set an example for the industry. DOJ
basically alleged that American Airlines used a temporary capacity expansion and fare
reductions to drive new entrants out of its Dallas/Ft. Worth (DFW) hub (1999c). DOJ
also had evidence of intent:
“If you are not going to get them [LCCs] out then no point to diminish
profit.” Don Carty, Chairman and CEO of American Airlines, 1996
DOJ’s complaint alleged that American Airlines dominated many of the routes
from its Dallas hub and charged monopoly prices. American controlled seventy percent
of the flight capacity from DFW. Low-cost carriers (LCCs) were proven to have positive
impact on consumer welfare through lower prices. DOJ complained that American
Airlines cut its prices to a level that would not make business sense except if it could
drive LCCs out of DFW before they could get a foothold in the market. Vanguard, Sun
Jet, and Western Pacific were victims of the alleged predatory strategy (1999c). All of
them were successfully driven out of the DFW hub and subsequently out of business. The
DOJ built its case incorporating elements of game theory and strategic economic theory.
These allegations, if proven, should have met the burden of predation (Piraino 2000).
The case was to go on trial in Wichita, Kansas on May 22, 2001. However, the
Federal court dismissed the case by summary judgment before the trial date. The
reasoning was derived from the Brooke Decision. The government had failed to show
72
that American’s pricing was below cost and that recoupment was highly probable. DOJ’s
modern approach to the case was perceived as an attempt to change antitrust law (Priest
2001), and time-honored rules (case memorandum and order 2001). With the new
administration in place, it is not likely that DOJ will be aggressive to bring a new case
forward. However, the decision to simply apply the Brooke philosophy to a network
industry case basically made parts of the new economy (e.g., airlines, software,
semiconductors, bio-technology) exempt from antitrust law (Carney 2001). Even so,
Judge Morten concluded:
The government's claims in the present case fail because American did not
price below an appropriate measure of cost, because it at most matched
the prices of its competitors, and because there is no dangerous
probability (even assuming below-cost pricing) of recoupment of
American's supposed profits by means of supra-competitive pricing. With
respect to costs, the evidence shows that American priced its fares
consistently above its average variable costs. Alternative, creative
measures of costs proposed by the government are inconsistent with
existing law, and inconsistent with an antitrust regime which seeks to
nurture rather than throttle vigorous price competition. With
respect to the question of recoupment, the government's claims suffer from
a pervasive failure of proof…Actual or likely recoupment by supra-
competitive pricing finds no basis in the evidence…The government's
theory of liability by reputation is not the law, and should not be. A
fundamental principle of antitrust law is that it be capable of effective and
accurate administration, and not chill the competition it seeks to foster.
The government's reputational liability approach would violate this
principle, permitting claims of predation based solely upon the subjective
and unverifiable complaints of a defendant's competitors. The low fare
carriers in question entered the core markets seeking to play a new sort of
ball game. The government's theory — that an established competitor
should not, and indeed, cannot deviate from its existing market strategy in
the face of aggressive price cutting by a new entrant — represents a whole
new mid-game spin on time-honored rules. Here American played by the
traditional rules. It competed with the low fare carriers on their own
terms. It did not price its fares below cost; it did not undercut the other
carriers' fares... Summary judgment is appropriate (U.S. v. AMR Corp et
al., Summary Judgment, 2001, pp.136-37).
73
3.5 Legal Elements of Proof
A look into the legal elements to prove for winning a case is insightful for an
understanding of the mechanics of the case trials. This section is based on McCareins
(1996).
Appropriate Measure of Cost: The courts have not agreed on a definition of an
appropriate measure for the cost tests. Appendix A: State Law Survey demonstrates that
the states employ many different standards. Similarly, Federal circuits do not have a
general standard. AVC and ATC are generally used as thresholds for burdens of proof.
The reason for the lack of a generally accepted measure is due to the understandable
hesitation of the Supreme Court to define it (McCareins 1996). However, twelve Federal
Appellate Courts have typically interpreted the appropriate measure as selling below
average variable cost.
Aggregation of Products: Whether cost calculations should be done with respect to a
single product, a product line, or a particular outlet or production facility has not been
defined. Some circuits have favored an assessment of predatory pricing based on the
costs and revenues associated with a full product line, as long as the competition was not
limited to a narrower scope (McCareins 1996).
Classification of Costs. There involve heavily debated issues such as determining what is
an appropriate measure of cost (average variable cost, average total cost, average
avoidable/incremental cost etc.), and classification of the incurred costs as variable, fixed
or avoidable depending on the definition of relevant market(s) and choice of time
horizon. Accounting and inventory systems vary; the distinction between costs and
investments is not clear cut, and it is hard to attribute costs to a specific product line or a
74
product in the often encountered case of multiple product lines. More recently, the
complexity of these issues has been further elevated where whether and how above-cost
pricing can be predatory is being examined (e.g., Edlin 2002; Meeks 1998).
Feasibility of Recoupment: This element refers to the proof that the predator will actually
profit from predation. This would require that once the rival is eliminated, the predator
would be able charge monopoly prices. The predator should be able to exert the
monopoly prices for a sufficient period of time so that he can recover his losses during
predation and also make a fair return on its investment in predation. Evidence of
recoupment was not a required element of predatory pricing in all courts before the
Brooke decision, which announced that the predator must be able to recoup losses
through supra-competitive prices for the case to be found predatory in nature. It was also
inferred that recoupment would not be possible if the markets are competitive, barriers to
entry are low, and the alleged predator either does not have the capacity or lacks the
resources to create capacity to capture the market share of an eliminated rival (Watson
1998).
There are two general ways that a plaintiff can attempt to prove recoupment. The
first path is to show that recoupment occurred through actual market data (supra-
competitive pricing for sufficient time to recoup). The second path would be trying to
prove that the predatory behavior was likely to bring tacit coordination and oligopoly
pricing. Tacit coordination would depend on similarities between the goals of the
competing firms, product variety and differentiation in the industry, and likelihood of
firms successfully signaling each other about price and output (1993).
75
Role of Defendant's Subjective Intent: Whether or not the intent of the defendant will
play a role in the case depends on the circuit. Predatory intent may be admissible to
reverse the assumption that above AVC is legal. Some courts have argued that proof of
intent leads to illegal scrutiny through corporate documents and hence is
counterproductive. Some have argued that intent is not important as long as the rival
could not be eliminated through predation. Intent to monopolize may also be derived
from attempts to create artificial barriers to entry, by entry deterring pricing, restrictive
marketing practices, acquisition of key inputs, or aggressive expansion of capacity
(McCareins 1996).
Role of Barriers to Entry: The Supreme Court view holds that high barriers to entry
enable the firm to recoup profits over a longer term. However, with low barriers,
predatory pricing may have limited effect on competition since as soon as the incumbent
increases prices new rivals will see the opportunity and enter.
Relevant Market: Last but not least, the relevant market refers to the product, service or
geographical areas that are involved in the case. Monopoly power of the predator applies
to the relevant market through its exclusion of rivals in the market and the subsequent
supra-competitive pricing. A product/service market should have its own elasticity of
demand, whereas geographic markets can be local, state, national, and even international.
“Evaluation of actual, probable, or presumed anticompetitive effects can be done only in
the context of a relevant market. Therefore, one of the first steps in an antitrust analysis is
determining the boundaries of the relevant market(s)…” (Enders 1986, p.24).
The determination of the relevant markets becomes harder when the multi-product
nature of the markets is taken into account. Most businesses have at least one product
76
line, and most large companies have multiple product lines. A static analysis of predatory
pricing may have indicated that it is irrational in a single market, however it can be
rational in a multiple market environment where a reputation effect of predation can deter
potential entrants in related markets (Trujillo 1994). From a legal standpoint, the relevant
market refers to the product, service, or geographical areas that are involved in the case.
Market Power: Monopoly power has been defined as “the power to control prices or
exclude competition” (1956, p.391; 1979, p.272). The existence of market power can be
simply inferred when a defendant has pre-dominant market share (1966, p.571).
The structure of a market may be derived from the study of the number and size
of firms, their cost and demand conditions, product differentiation, the nature of any entry
barriers, and degree of regulation. Industry Concentration Ratio (percentage of total sales
of the n (usually four) largest firms in an industry) and the Herfindahl-Hirschmann Index
((HHI) –sum of the squared market shares of all firms) are two commonly used market
structure indicators. However, the calculations for both measures need to be based on the
relevant market figures to be meaningful.
Application of performance based market power tests include practical or
conceptual difficulties, hence most studies use market structure tests. Previous studies
have generally shown a positive correlation between market concentration and
profitability (Shepherd 1970) (though the actual relationship may be more complex due
to omitted third variables (Bharadwaj and Varadarajan 2004)).
The U.S. Department of Justice's merger guidelines suggest that an HHI over
1,000 may raise antitrust concerns (1997a). According to the 1992 DOJ/FTC Horizontal
Merger Guidelines, markets with an HHI greater than 1800 are considered "highly
77
concentrated," and the DOJ/FTC will not approve mergers and acquisitions in highly
concentrated markets if there is even a slight increase in the HHI (1997a). Courts will not
predation credible unless the incumbent has market power.
3.6 The State of the Debate on Predatory Pricing
The purpose of the Antitrust Act and its application in the courts is a continuously
debated topic. As mentioned previously, there are two main schools of thought. The
Chicago School of thought based on the neo-classical theories argues that predatory
pricing is “rarely tried and even more rarely successful” (1986b). On the other side, the
emerging Post-Chicago School of thought argues that there is plenty of evidence that new
approaches such as strategic theory and game theoretic models provide that predatory
pricing can be rational and profitable for the predator (Bolton et al. 2000). The two
schools are compared next:
3.6.1 The Chicago School
In an early work that continues to influence the courts today, McGee (1958)
studied the 1911 Standard Oil Case, which was considered to be the classic case of
predation. He found no evidence in the case trial records that Standard Oil had indeed had
cut its prices below cost to drive out smaller competition and later intended to increase
prices. He argued that predatory pricing by Standard Oil would have been irrational
because of the relatively larger losses it would have had to suffer due to its higher market
share. He also argued that the prey would not be inclined to leave the market since it
knows that the predator cannot afford such large losses infinitely. Funding for the prey
78
would not be a problem either since capital markets would effectively step in as long as it
is an efficient producer. He claimed that the predator would not have gained anything
even if it could drive competition out of the marketplace because the prey, a purchaser of
prey’s assets or new competition could enter the market as soon as the predator increases
its prices.
With the lack of a rival theory, McGee’s (1958) analysis was considered to be the
only coherent economic theory of predatory pricing (Bolton et al. 2000). Other scholars
such as Koller (1971; 1978); Areeda and Turner (1975; 1996), Harold Demsetz (1973),
Demsetz and Weiss (1975); George Stigler (1987); and Wesley Liebeler (1986) expanded
upon this work and built the foundation of the Chicago School of thought. During the
nineties however, despite the late effort by John Lott (1999; 1996) the Chicago School
seems to have lost its dominance in academic journals.
2
Free enterprise institutes carry on the mission of the Chicago School. Institutes
such as the American Enterprise Institute, Heritage Foundation, and Cato Institute
(www.cato.org) promote free markets and limited government. They sponsor Chicago
School stream of research through their funding which comes from foundations,
corporations, and individuals. They maintain numerous web sites and publish journals,
magazines, and newsletters to keep the Chicago torch burning. Foer and Lande (1999)
reported that each of the above institutions spends around up to $30 million and argued
that their lobbying has resulted in considerable decline in Federal antitrust funding over
time.
The courts continue to empathize with the Chicago School and have been
2
I am joined by Bolton et al. (2000), and Sappington and Sidak (2000) in my critical view against
the validity of Lott’s findings.
79
exercising it as though there are no opposing theories. Judges who have openly sided
with the Chicago School include judges Bork (1978) and Easterbrook (1984; 1981).
There seems to be a prejudice that most cases are brought forward by the small
companies to create heavy litigation costs for large corporations. For example, Judge
Easterbrook estimated the average cost of a predation case for a major corporation to be
around $30M and even went as far to argue that the antitrust offense of predation should
be forgotten (Easterbrook 1981, p.337).
The general argument of the Chicago School has been summarized here based on
the work of DiLorenzo (1992). The counter argument of the emerging school of thought
is presented in the next section.
7.1.2 Chicago School Claim: Predatory pricing is irrational and very rarely (if ever)
occurs because:
1. Predatory practices are more costly for the large firm due to its larger market
share (it has to incur losses on a larger number of units) (McGee 1980).
2. It is not possible to continuously charge supra-competitive prices. Potential
entrants will be lured one after another. The incumbent cannot recoup losses
(Gomez et al. 1999).
3. Price wars are inherently uncertain, thus recoupment possibility cannot be
calculated accurately. A price war could also spread to surrounding markets,
making predation even more risky. McGee (1980) argued that even though
predatory pricing is usually irrational, it is generally rational for the victim to hold
on because predatory pricing strategies represent temporary cuts in prices.
Furthermore, Easterbrook (1981) argued that the victim has the same incentive as
80
the predator to outlast its rival and possibly collect eventual monopoly rents. In
expectation of supra-normal price levels and profits that will follow the price war,
capital markets can step in and help the prey. Thus, deep pockets argument will
not hold (Gattuso and Boudreaux 1999).
4. The prey firm can shut down and wait for supra-normal prices. In the meantime, if
they go out of business, someone else can take over.
5. There are opportunity costs associated with the funds allegedly used for predation
by the large firm.
6. Consumers could potentially stock up during predation; limiting quantity per
consumer also would not work as competition would step in and supply the
unserved demand.
7. Victims can arrange for long term contracts above predation prices with
customers if they also realize that monopoly prices will follow. Since the
customers will benefit from the prey’s continued existence as a supplier, they may
agree to such long term agreements to buy at a truly competitive price.
8. Anticipated monopoly profits have to be discounted to present value, diminishing
any value predation may have had. Indeed, the future recoupments must be
discounted by the probability that monopoly power will not be achieved, and then
discounted again to present value (Easterbrook 1981, p.272). Acquisition is a
much more profitable way of eliminating a competitor (Gomez et al. 1999).
According to the Chicago School, economic efficiency of the market should be
the goal of antitrust (Bork 1978). This interpretation evolves around maximizing
81
productive and allocative efficiency. The productive efficiency is simply defined as the
ratio of the outputs in relation to the inputs the companies employed. Allocative
efficiency refers to the general efficiency of the markets and involves how the limited
societal resources should be allocated across industries. Allocative efficiency in general is
referred to as the consumer welfare (Gundlach 1995).
The stance regarding merger and acquisitions is also a very important antitrust
issue. It is also related to predation since mergers and acquisitions typically lead to
dominant market power for the firms involved, the prerequisite for predatory pricing. The
Chicago School view of mergers can be summarized as follows: as long as benefits
(productive efficiency benefits to merging firms, and allocative efficiency benefits to
consumers) of a proposed merger is not lower than the potential loss to the competitors, it
would not be seen as a violation of antitrust. Similarly, the Chicago School claims that in
an industry with two levels where the production is monopolized and distribution is
competitive (i.e., gasoline), the monopolist cannot increase its profits by acquiring the
distributors. Increasing the retail markup will mean decreasing the producer markup by
the same amount. The monopolist cannot maximize its profits beyond the monopoly
profits (Posner 1976).
The intent and drastic price cuts that the populist era emphasized are not
important for Chicago School: “intent plays no useful role… Firms “intend” to do all the
business they can, to crush the rivals if they can…[A] desire to extinguish one’s rival’s is
entirely consistent with, [and] often is the motive behind, competition…[P]rice
reductions are carried out in a pursuit of sales, at others’ expense. Entrepreneurs who
work hardest to cut their prices will do the most damage to their rivals…” (1989,
82
pp.1401-1402). It follows that the intention to compete is not different than the intention
to exclude a rival.
Some Chicago School extremists take the position even further by claiming that
below cost pricing should also be legal. It has been argued that meeting price cuts,
discounting for introducing new products and excess capacity of perishable products are
already acceptable reasons for below cost pricing (Boudreaux and Kleit 1996). It was
proposed that all government monopolies including the postal service and public schools
be deregulated (Boudreaux and Kleit 1996b) and that all direct competitor suits be
prohibited (Boudreaux and Kleit 1996a).
The Federal Courts attempt to detect cases of predatory pricing through the
assumptions and the logic of Chicago School which is derived from the neo-classical
price theory view (Guiltinan and Gundlach 1996). Figure 3.4 illustrates the decision-
making criteria that the Federal courts use today.
83
Is the defendant sacrificing current profits?
1. Cost-based Tests
Are the Relevant Competitors Affected?
1. Exit?
2. Reduce Output?
3. Decide not to Enter?
Will There be Injury to Consumer Welfare?
1. Allocative Efficiency
2. Productive Efficiency
Is Recoupment Probable?
1. Market Power
2. Barriers to Entry
3. Industry Characteristics
Case Dismissal
by Summary Judgment
Verdict:
Defendant Prevails…
Verdict:
Plaintiff prevails…
Harm to consumer and
competitive process
Adopted from Guiltinan and Gundlach (1996, p.89)
chasm
Yes
No
No
No
Yes
Yes
Yes chasm
No
Figure 3.4: Decision-Making Framework in the U.S. Federal Courts
3.6.2 Post-Chicago School of Thought
Even though Koller (1971), Areeda and Turner (1975) and other earlier literature
provided seemingly counter evidence for a rationale of the existence of predatory pricing,
the claim for a myth of predatory pricing remained unjustified (Bolton et al. 2000). Koller
had reported that out of 23 cases he studied, actual predation was attempted in seven
(30%) and achieved in four cases (17%). Zerbe and Mumford (1996) studied the same
cases since 1940 and updated them until 1982. They detected predatory pricing in 27 out
of 40 (68%) of the cases.
Both of these studies probably under-reported the cases of predation because they
did not include settlements, predatory disciplining where no suit is filed, forced
acquisitions, and cases that were not filed because supporting theory was not yet
84
discovered or known (Bolton et al. 2000). Granitz and Klein (1996) conducted a re-
assessment of the Standard Oil case, and on contrary to McGee's (1958) work found
evidence that predation had occurred. Perhaps, actual predatory pricing cases are not as
rare as the courts have concluded (Adams and Brock 1996). “Exclusionary strategies are
frequent, not exceptional business practices”(Brodley and Hay 1981, p.1045). Bolton et
al. (2000) argued that there is now a consensus in modern economics that predatory
pricing can be a rational and successful business practice. They indeed made the claim
that no major article has had a counter statement during the last three decades. Many
weaknesses of the Areeda-Turner rule and the Chicago School arguments have been
reported (McCall 1987).
Many of the early critics of Areeda-Turner focused on its short run focus. For
example, Joskow and Klevorick (1979) argued that “...to dismiss entirely an assessment
of long-run effects, as for example Areeda and Turner seem to do, is to dismiss the
essence of the predatory pricing problem.” Williamson (1977) argued that temporary
price cuts have negligible benefits and resulted in long-term social welfare problems.
The solely cost based approach of Areeda-Turner was also criticized for being
impractical. Accounting and inventory systems vary; the distinction between costs and
investments is not clear cut; and it is hard to attribute costs to a specific product line or
product in the often seen case of multiple product lines. Areeda-Turner cost-based tests
are difficult to apply due to the multi-product nature of most businesses (Gomez et al.
1999). The emerging view’s counter arguments of the Chicago School are summarized
below, and additional points follow.
85
7.2.2 Post-Chicago School Response: Predatory pricing can be rational and may
frequently occur because:
1. Predatory practices are not necessarily costly for the large firm due to its larger
market share because predation does not occur in every segment of the market. It is
usually geographically or otherwise localized to a segment. Instead, the predator
may subsidize its local losses by its profits in other segments and wage war much
longer.
2. It may be possible to charge supra-competitive prices due to high barriers to entry,
reputation and other signaling effects. Potentially more efficient entrants see what
happened to the previous entrant, and may decide not to commit the high level of
resources needed to fight the incumbent. A valid purpose of predation may be to
develop a reputation as a tough competitor (Comanor and Frech 1993), (Kreps and
Wilson 1982). Bad lock-ins in network industries may result in loss of welfare.
3. Price wars often result in the destruction of the small competitors due to the deeper
pockets of the predators. Capital markets will not necessarily step in and help the
prey because of reputation effects. “Each time a start-up is driven out of a market,
the difficulty of obtaining funding ratches up a notch.” The fact that there is no
readily available financing for a start-up that has faced predatory practices is a direct
contradiction of Chicago theory. Lenders normally require substantial security, a
demonstrated ability to repay. If the loan is perceived to be risky, higher interest
rates will result. It is not likely for prudent lenders to give credit at all to a small
prey facing predation by a large and experienced firm (Atwood 1998).
86
4. It is not realistic to argue that the prey firm can shut down and wait for supra-normal
prices. Brand equity is hard to build up and customers value continuity of products
and services.
5. There are opportunity costs associated with the funds allegedly used for predation by
the large firm, however there are opportunity costs associated to the capital of
potentially more effective entrants as well. They will not enter the market if they
have other options where they are not likely to meet with predation. Moreover, the
opportunity cost argument presents a rationale for predatory pricing. Foregone
profits, that would have been earned if the monopolist had employed its assets in
alternative options rather than using them to predate in the subject market, also
represent costs (Ordover and Willig 1981). Comanor and Frech (1993) argued that
the predator incurs losses from an economic sense, but not necessarily from an
accounting sense. The loss is the lower profits than otherwise could have been
earned.
6. The argument that consumers could potentially stock up during predation, simply
does not hold for network industries (e.g., transportation, software,
telecommunications) and services and is not practical for many other industries.
Customers would not engage in long-term contracts above predation prices unless
they realize the full implications of the ongoing predatory pricing. They normally
would try to maximize their short run returns as long as they believe they have
alternate suppliers. Moreover, long term contract argument does not hold for
consumer markets (e.g., airlines).
87
7. Anticipated monopoly profits may have to be discounted, however the strategic
long-term gains also need to be built into the equation. What would the incumbent’s
losses be if it had remained inactive? How many potential entries were discouraged
by the predatory behavior? These are hard to estimate but the long term impact on
business is not. Recouping the investment may be much easier than thought. Case
studies have suggested that predators have occasionally succeeded in recouping their
losses (Adams and Brock 1996).
8. Contrary to what Easterbrook (1981) argued, litigation can be much more costly for
smaller companies which throughout the case have to incur legal fees they cannot
afford. Typically, large firms departmentalize the case as they have the personnel
and the resources to do so; however, the prey’s management can easily get drawn
into this side effort and drown in the increasing details of the case.
The arguments above and the emerging Post-Chicago School of thought could be
observed in the works of Klevorick (1993), Ordover and Saloner (1989), Craswell and
Ratrik (1985). The Post-Chicago School raised numerous issues that the Chicago School
theory had not captured before (1998a). Guiltinan and Gundlach (1996) criticized the old
school of thought because it relies on the assumptions of neo-classical theory such as
singular motivation of managers for profit maximization, and firms’ possession of
complete information.
Sole Profit Maximization: There may be practical objectives other than sole profit
maximization. Managers may settle for satisfactory rather than optimal levels (Baumol
1967; Simon 1979). Predatory pricing generally leads to a drastic increase in market
88
share, which if aligned to performance evaluations, may be a motivational factor for
managers. Maximization of career opportunities has also been offered as an alternative
for sole profit maximization (Stelzer 1987). Urbany and Peter (1994) have shown
empirical evidence for a preference for volume rather than profit orientation. Their
experiment indicated that manufacturing firms held a long-term perspective for customer
acquisition and in estimating their worth.
Rational Decision Making: Decision making under uncertainty is not rational. Guiltinan
(1996) argued that managers do not estimate probabilities accurately when they are
dealing with risky situations. Simon’s (1957) highly regarded concept of bounded
rationality implies that: a) buyers/managers are not aware of their comprehensive set of
alternatives, b) they do not know the outcome of taking a specific alternative with
certainty, and c) they do not have the mental capacity to rationally process all perceived
alternatives. Individuals tend to develop noncompensatory preferences (e.g., elimination
by aspects) when faced with complex decision scenarios (Bettman et al. 1998).
Kahneman and Tversky (1979) have long demonstrated that individuals do not
necessarily maximize their economic utility in the way predicted and assumed by neo-
classical economics. Their Nobel prize winning prospect theory essentially claims that
individuals have differing risk preferences and display risk-averse characteristics for
gains and risk seeking characteristics for losses (i.e., a sure $50 bonus would be preferred
over a coin toss for $100 or nothing, whereas a coin flip for a $40 parking ticket or
nothing would be preferred over a sure $20 parking ticket). That is, the utility curve for
gains is concave whereas the utility curve for losses is convex and also with a steeper
slope. Thus, managers are risk-averse for gains but risk-affinitive for losses.
89
Post-Chicago Proposals:
Meeks (1998) criticized the static nature of the traditional analysis and argued that
it may lead to unsatisfactory results if cost is used as the main criterion. He argued that
for anti-competitive conduct to occur, the price need not even be below cost, as long as it
was low enough to deter entry, and supra-competitive prices would occur once the
potential competition is eliminated. He proposed a non-cost based approach for detecting
predatory pricing in transition markets (e.g., telecommunications). It was important to
consider if the market was recently deregulated or associated with one that is regulated.
He suggested that if the lower price was offered to a large segment of the market, it was
probably not a static move aimed at preventing new competition. The predating firm had
to possess market power (at least 50%). Strategic harm to potential competition had to be
the most likely explanation for the predatory action and the barriers to entry had to be
high (Meeks 1998).
Bolton et al. (2000) in their award winning effort argued that modern economics
principles should be employed to submit proof of predatory pricing. In particular, they
suggested that the pro-competitive dynamic gains such as reputation effects should be
considered, and short and long run incremental costs should be employed for proof of
below-cost pricing. Essentially, their proposal required the following elements to be
incorporated:
1. Facilitating market structure: “Short-run pricing power” had to be present in the
market. This would be typically observed by the existence of one or more
dominant firms and high entry and re-entry barriers.
90
2. Scheme of predation and supporting evidence: A persuasive evidence of
recoupment would have to be presented at the absence of a plausible scheme of
predatory pricing.
3. Probable recoupment: Probable recoupment as opposed to actual should be
sufficient. Moreover, intangible benefits from injury to competition and
exclusionary effects would be acceptable.
4. Price below cost: They proposed that AVC should be substituted with Average
Avoidable Cost
3
, and ATC should be substituted with long run average
incremental cost (LRAIC).
5. Absence of an efficiencies or business defense: A plausible efficiencies gain, no
less restrictive alternative, and efficiency-enhancing recoupment would have to be
demonstrated.
This proposal was a novel attempt to incorporate the dynamic strategic
perspective into the current antitrust policy. Guiltinan and Gundlach (1996) argued that a
dynamic strategic approach underlines the insufficiency of the current law to cope with
aggressive and predatory pricing.
In summary, the Post-Chicago School contends that it may be rational and
plausible for firms to employ predatory pricing (Hazlett 1995). It examines the relevant
markets with a strategic perspective, considers impact of opportunity costs, and imperfect
information (Bolton Brodley, and Riordan 2000; Brodley and Hay 1981; see Bloom and
3
Average Avoidable cost: Average per unit cost that predator would have avoided during the
period of below cost pricing had it not produced the predatory increment of sales.
91
Gundlach (2001a) for a discussion of differences between Chicago and Post-Chicago
Schools of thought). Simply put, the Post-Chicago view is that markets are generally not
perfect, and they are not necessarily self-correcting (Gundlach et al. 2002). Some of the
prominent names that have supported this stream include Lawrence Sullivan, Paul
Joskow, Alvine Klevorick, Januzs Ordover, Robert Willig and Joseph Stiglitz (Elhauge
2003).
Marketing perspective is well aligned with the Post-Chicago School of thought in
its basic position. Additional assumptions of marketing to that of the Post-Chicago
School are alternatives for assessing consumer welfare (e.g., variety, innovation,
satisfaction), and the possibility of irrational decision making and non-profit maximizing
goals (Gundlach 2001). Marketing could well assume a complimentary role due to its
cross-disciplinary nature which has not yet been realized (Gundlach et al. 2002).
3.7 Synthesis and a Marketing Perspective
3.7.1 The Need for a Marketing Perspective:
The unique position of marketing to offer public policy insights has been
observed by several scholars working on the subject (Grewal and Compeau 1999;
Guiltinan and Gundlach 1996b). It was suggested that a marketing focus on predatory
pricing is overdue (Grewal and Compeau 1999). The many aspects of consumer welfare,
which is of key concern to public policy, can be captured with the comprehensive
marketing measurement and modeling tools that enable the study of the benefits of
quality, service, variety, and innovation to consumers (Guiltinan and Gundlach 1996).
The marketing discipline has the potential to further the understanding needed for the
92
development of a more suitable antitrust policy (Gundlach 1995). This potential has not
been utilized by the courts so far. A review of the marketing mix variables could reveal a
better understanding of predation than a focus on pricing alone.
The issue facing legislators of predation is indeed similar to one researchers face
when designing a scientific study. Defendant’s conduct can be presumed legal (i.e., non-
predatory) for the null hypothesis (stemming from the notion of being innocent until
proven guilty). Alternate hypothesis would be that the defendant’s conduct is illegal
(predatory). The risk that an incorrect conclusion may be reached always exists. A
balance needs to be sought through the manipulation of Type I (rejecting the null
hypothesis when it is true), and Type II (failing to reject when null hypothesis is false)
errors, and the respective power (probability of successfully rejecting the null hypothesis
when it should be rejected) of the test. Legislators set the evaluation criteria (e.g., Areeda
–Turner AVC rule), imposing rigid parameters that apply to everyone in the name of
being just. The set criteria, hence the parameters can drastically shift over time as
differing success rates of plaintiffs for similar cases show. In some cases, however, the
judges may have set unfortunate examples when they interpreted these criteria too
strictly. The influence of these (Supreme Court) decisions may have resulted in the self-
fulfilling prophecy that predatory pricing is a “myth.” Yet, the courts seem to be content
with this situation due to their comforting assumption that predatory pricing “rarely”
succeeds and that consumers benefit from failed predation attempts. A minimal standard
(i.e., higher alpha level) would have been costly for the competition, because a number of
wrong inferences (i.e., rejections through Type I error) would have a negative impact on
desirable price cutting behavior (1993, p.226).
93
3.7.2 A Focus on Deregulated Markets:
It was argued that price discrimination has to accompany strategic predatory
pricing since predation is usually targeted by geography or other segmentation (Meeks
1998). Dempsey (1989) discussed how deregulation of an industry facilitates
discrimination. Brennan (1995) studied deregulation in the telecommunications industry
and detected potential for predatory pricing. The same hazard was detected for electric
power, natural gas (Meeks 1998) and airline industries (Kahn 1987). Meeks (1998)
argued that recently deregulated industries are much more likely to have examples of
predation than those that are not. This interesting notion is reflected in the following
framework (Figure 3.5).
Figure 3.5: Framework for Deregulation and Predatory Pricing Relationship
Naturally, industry characteristics influence all factors in this framework
including deregulation. The number and size distribution of firms, cost and demand
conditions, and barriers to entry impact firm strategies. However, it is also true that
deregulation often transforms a monopolistic industry to an oligopoly with a number of
Deregulation Price
Discrimination
Predatory
Pricing
New / Potential
Inter Entry
Non-price Predation
94
firms with high market power. Since no precedent has been set, it can be easy to employ
price discrimination. It is particularly easy to discriminate in previously regulated service
industries because there are few arbitrage opportunities for the consumers (e.g., leisure
traveler cannot transfer plane pre-purchased plane tickets to businessmen on the run).
Robinson-Patman Act limits price discrimination for commodities and does not generally
apply to service or lease industries (1936). This enables such industries to set supra-
competitive prices. Kotler and Armstrong (1991, p.343-44) described an airline industry
that has taken full advantage from the possibilities of price discrimination:
The passengers on a plane bound from Raleigh to Los Angeles may pay as
many as ten different round-trip fares for the same flight—first class; first
class-night; first class-night, child; first class-youth; coach; coach-night;
coach-night, child; Super-Saver, nonrefundable fare; Super-Saver, 25
percent cancellation penalty; and military personnel. These fares vary
from $238 to $1512!
The existence of discrimination could provide rationale for predation because
discounters with their low cost structure can be very damaging to the incumbent’s
established price discrimination strategy based on market segmentation. For example,
revenue management is essential to the success of the hospitality and airline industries,
and is closely related to marketing through price. Major carriers’ predatory behavior
against the low cost carriers has been observed in the airline industry (1998c; 1999b;
1999c; 1999d). Incumbent firms may also employ non-predatory tactics to prevent new
entry to their bases of market power.
95
3.7.3 Epilogue
By all accounts, the 1975 Harvard Law Review article “Predatory Pricing and
Related Practices under Section 2 of the Sherman Act” by Phillip Areeda and Donald F.
Turner was the seminal article in predatory pricing. All U.S. courts have been influenced
by this article in one way or another. It has been noted that the widespread acceptance of
Areeda-Turner and the conservatism in the U.S. courts, especially since the eighties, may
be explained by judiciary appointments by conservative Presidents, the intellectual
influence of the Chicago School and the extensive public relations campaigns to
communicate its aspects to the judges, and the contrasting lack of an organized antitrust
effort (Foer and Lande 1999). Following Areeda and Turner (1975), that predatory
pricing is rare and AVC is the measure to test the claims of predatory pricing have been
generally presumed by the courts with some exceptions mostly on the State level.
However, Areeda and Turner and the Chicago School have had their share of criticisms:
Scherer has demonstrated that the Areeda-Turner rule would not promote
long-run economic welfare, would not ensure an efficient allocation of
resources, and would encourage firms to maintain excess capacity. Greer
has shown that the Areeda-Turner rule relying on either an average
variable or marginal cost floor would be overly lenient in that it would
allow the destruction of equally (or more) efficient rivals. Dirlam has also
rejected cost-based rules, noting that they are too rigid and would require
difficult and ambiguous short-run cost measurements. Finally,
Beckenstein and Gabel have argued that succinct per se rules are unable
to deal with some anticompetitive practices and other subtle forms of
business behavior like predatory investment, economies of scope, vertical
integration, and experience curve learning (McCall 1987).
The criticisms of the Post-Chicago School are summarized in Table 3.2.
96
Table 3.2: Criticisms regarding the Areeda-Turner rule and the Chicago School
Theoretical
• Uses unrealistic assumptions of neo-classical price theory (e.g., Lande 1993;
Guiltinan and Gundlach 1996)
• Describes the manufacturing era of the past versus the current knowledge/network
society (e.g., Bolton, Brodley, and Riordan 2000)
• Previous empirical findings were flawed (e.g., Granitz and Klein 1996; Zerbe and
Mumford 1996).
• Predatory pricing can be rational (e.g., Adams and Brock 1996; Sullivan and Grimes
(2000)
• Lacks a strategic perspective (e.g., Williamson 1977; Joskow and Klevorick 1979)
• Ignores reputation, signaling effects and information asymmetries (e.g., Comanor and
Frech 1993; Jung, Kagel, and Levin 1994; Milgrom and Roberts 1982)
• Ignores the nature of network/service industries (e.g., European Commission 1998)
• Ignores opportunity costs of actions (e.g., Ordover and Willig 1981)
• May exclude (more) efficient rivals (e.g., Greer 1979)
• Promotes inefficient allocation of resources and excess capacity (e.g., Sherer 1976)
Legal
• Cost rules are hard to calculate, impractical (e.g., Sievers and Albery 1991)
• Short-run cost focus is irrelevant (e.g., Dirlam 1981; Meeks 1998)
• Per se rule is inapplicable to all cases (e.g., Beckenstein and Gabel 1986)
• Litigation can be costly for the prey (e.g., Atwood 1998)
Managerial • Promotes selective and geographic predation (e.g., Allvine 1996)
• Narrow view of market power (e.g., Trujillo 1994)
• Ignores deep pockets of predators (e.g., Atwood 1998)
Buyer • Does not promote long run economic welfare (e.g., Scherer 1976)
Despite these criticisms, the Chicago School has keenly used Areeda-Turner as
the basis for their static, neo-classical price theory arguments for more than two decades.
However, not only the contexts of predation and the nature of the problem have changed
over the years, but also the opposing Post-Chicago views have been growing stronger
(1998b). Hence, it is worthwhile to revisit the Areeda and Turner (1975) article and
comment on its key limitations in today’s economy.
The context in which Areeda and Turner published their article was radically
different than that of today. In the Areeda and Turner world, the economy was still driven
by production of oil, steel, and commodities. Sheer output was what mattered most. In the
manufacturing era, it was presumed that the bigger was better, higher capacity would lead
to better economies of scale and a better cost structure. Being large enabled them to gain
97
dominant market share and hence market power. Areeda and Turner projected of a
dominantly manufacturing world where there is either perfect competition (i.e., many
firms take going price) or monopoly (i.e., one firm sets the price). “The view that
predatory pricing is rare or even fanciful is based on theoretical constructs that ignore the
realities of markets in which oligopolistic structure, incomplete information and strategic
behavior are commonplace” (Sullivan and Grimes 2000, p. 145).
The “real world” is almost always somewhere in between a monopoly and perfect
competition, with differentiated products of oligopolies instead of commodity outputs
and identical firms. In the real world, the firms do not compete only on quantity and
price. It is a service-based economy that is pre-dominant in the U.S. society today. The
small firm can be more efficient and welfare enhancing than the corporate giant. Small
service companies from catering to transportation, and entertainment to management
consulting can be as or more efficient than the big players. Southwest in the airline
industry is one clear example of an efficient firm. The nature of the airline industry
enables clear illustrations of the new issues facing the policy makers.
Areeda and Turner perceive two pre-requisites for predatory pricing: a greater
financial staying power by the predator and a very substantial prospect of recoupment.
The use of Chicago School doctrine, made it really easy for Judge Marten to decide that
American had indeed not violated the law and to dismiss the case by summary judgment.
American had not priced below AVC, and the recoupment of losses incurred during the
alleged predation was not probable anyway (memorandum and order) (1999c). Yet,
looking at the same facts, DOJ and the Post-Chicago School concur otherwise. Modern
thinking in antitrust, dubbed here as the Post-Chicago School deals with reputation
98
effects, game theory, conspiracies to raise rival’s costs or reduce rival’s revenues and
even scanner data input analyses where applicable. The Post Chicago School studies the
relevant markets with a strategic perspective and considers the firm’s opportunity costs of
its actions.
The current antitrust enforcement does not consider the differences in competitive
reaction to entry by intra- and inter-type companies. This purposeful ignorance extends
from the Pre-Areeda-Turner enforcement era heavily criticized by the Chicago School.
The Robinson Patman Act (1936) was passed with the general intent of protecting the
mom-and-pop stores from the chain-store revolution. Many small stores won cases of
predatory pricing against the large (and generally more efficient) chains until Areeda-
Turner set the tone for the rest of the century, and the objective of antitrust became
guarding the competitive process and not the (small) competitors. For example, entrant
Wall-Mart grew as a discounter by undercutting the high-service/high cost department
stores and established itself as one of the largest retailers in the world. The incumbent
local mom-and-pop stores did not have much other than their older but loyal clientele to
defend themselves against the chains. The stores that were less efficient simply faded
away.
Areeda and Turner (1975) utilized a number of purely theoretical examples to
illustrate their points, and they readily admitted that some of these scenarios posed threat
to long-term competition. Areeda and Turner did not believe that a long-term dynamic
test of predation could be practically developed. They reasoned that their short-term static
cost focus would be practical. However, the very notion of recoupment represents a
strategic assessment of a long-term return (supra-normal profits) on a short-term
99
investment (predatory pricing). Thus, the nature of the problem imposes the need to use a
strategic-dynamic test (rather than a short-run static rule) to assess predatory pricing.
Today, technology enables market segmentation at the micro-level. Companies
offer product/service lines with different levels of utility at varying prices (e.g., premium,
standard, discount). Specialization and focusing on different market segments or niches is
also common. Customer Relationship Management and other sophisticated marketing
tools permit customizing the product/service offer even to the individual consumer’s
needs. However, this targeting capability becomes destructive if predatory pricing is
pinpointed at a local rival. For example, American Airlines did not cut prices on all or
even most of its routes on its DFW hub when it faced entry by discounters. It cut its
prices only on those routes in that it competed head to head with them. This geographic
segmentation enabled them to subsidize losses incurred during the time of predation on
targeted markets. The revenue and losses at stake were insignificant for American but
vital for the survival of Vanguard (and SunJet and Western Pacific which are now out of
business). As the American Airlines case justified “devices other than a general price-cut
may, however, be the subject of suits for predation” (Areeda and Turner 1975). Such
selective or geographic price cuts provide the major airline next to infinite staying power
through cross-subsidization.
Moreover, American increased capacity on these routes by shifting its aircraft
from more profitable routes. Any business expanding capacity entails opportunity costs.
“[A]ny cost calculation that totally ignores the opportunity cost component is likely to be
illegitimate…it is essential to include all opportunity costs of ownership inputs…”
(Baumol 1996). Foregone profits that would have been earned if the monopolist had
100
employed its assets in alternative options rather than using them for predation represent
costs (Ordover and Willig 1981). Ironically, Areeda and Turner rule does not take the
opportunity costs into account. Predatory pricing analyses appear to be plagued by cost
issues.
Areeda and Turner have advocated the per se AVC rule, however, in their seminal
work they also asserted that “virtually all costs are variable when a firm, operating at
capacity, plans to double its output by constructing new plants and purchasing new
equipment” (1975, p.701). Deep discounts coupled with drastic volume increase can
indeed be the typical response to market entry depending on industry characteristics (e.g.,
air and truck transportation). Thus, the use of an ATC test (i.e., a less rigorous test than
that of AVC), may be practically justified under special circumstances.
Price discrimination today is a typical practice in service industries and is not
regulated by the Robinson-Patman Act. It has been shown that economies of scale can
motivate a price-discriminating monopolist to engage in unprofitable conduct and sell
even when the average cost is higher than the price. The incurred losses can be
effectively subsidized with profits from other markets. This can have harmful effects on
social welfare (Park 2000).
Areeda and Turner (1975) argue that “a demonstrated willingness to indulge in
predatory pricing might itself deter some smaller potential entrants… Repeated predation
in the same market, moreover, is not only costly but is likely to be easily detectable and
thus the occasion for severe antitrust sanctions.” Indeed, this argument is the essence of
the theory of reputation effects. Areeda-Turner is practically the standard rule in courts
today. Yet, it does not equip us to detect predatory pricing in many cases in a service-
101
based economy and makes it almost impossible to detect it in network industries (Carney
2001). Even the most deliberate cases of repeated predation may not be detected today
because of over-reliance on the Areeda-Turner rule.
According to the Federal courts, (predatory) intent is at best a secondary criterion
to consider (A.A. Poultry Farms, Inc. v. Rose Acre Farms 1989). This may have to do
with the fact that both the Chicago School and the Post-Chicago School agree that the
firm’s objective is profit maximization. Interestingly, a decision-making perspective
adopted by marketing, and one that goes beyond both schools is an understanding that
alternative and multiple managerial objectives such as survival, satisficing (e.g., simply
meeting the Wall Street numbers), and enhancing sales, personal welfare (i.e., self-
compensation, career advancement), reputation effects, and social welfare do exist
(Gundlach 2001). Sales management researchers have long studied the optimal employee
compensation mix for different business objectives such as building, holding, harvesting,
or divesting market share (Strahle and Spiro 1986). Similarly, top executive
compensation seems to correlate positively with sales but not necessarily with
shareholder return (McKnight and Tomkins 2004). The use of a sales/market share
growth objective typically enhances not only personal welfare but also corporate
reputation effects (i.e., warranted aggressive/predatory response by incumbent).
Therefore, the courts should recognize that alternative objectives can be legitimate goals.
For example, maximizing social welfare can be prevalent in the case of non-profit
organizations. Often an overlooked research dimension, the nature of competition
between organizations with contrasting objectives presents challenging issues for policy
makers. Manufacturers may aggressively cut prices (i.e., market share growth objective)
102
to secure contracts in their main business market (e.g., copiers/high speed printers) to
make premiums on complementary products (e.g., ink cartridges/services). Not-for-profit
hospitals have been accused of abusing market power in concentrated markets (Simpson
and Shin 1998). Competitive bidding/pricing by public universities for government
contracts may also be considered as unfair competition by the private industry
members/universities. This interesting phenomenon underlines the necessity of
undertaking a broader examination of the consumer welfare construct which is discussed
at a later section.
Moreover, the strategic marketing orientation of the aggressor business should be
examined in the courts. The orientation of the business can directly impact its
profitability and selection of performance objectives (Gatignon and Xuereb 1997). For
example, competitor-oriented businesses would likely retaliate more aggressively than
customer-oriented businesses in the face of competitive entry (Narver and Slater 1990).
Predatory sacrifice perceived by managers in competitor-oriented businesses is likely to
be much smaller than that perceived by their customer-oriented counterparts. The notion
of perception of sacrifice requires a discussion of rationality and risk tolerance as well.
The state of marketing and consumer research has progressed beyond the basic
risk-averse/prone notion and explored the role of multiple and conflicting personas and
non-financial (e.g., social) risks and moderating roles of risk perceptions on attitude
formation and decision making (Campbell and Goodstein 2001). Choices can also be
heavily influenced by the framing/perception of the proposition: a teenager might drive
through town to save $5 on a single DVD that would have cost $20 but would not do the
same to save $5 on a pair of athletic shoes that cost $150 (Thaler 1985). Similarly, a
103
purchasing agent/jobber may fly across the world to resolve a $50K dispute on an office-
supply deal worth $1 million, but may be reluctant to do so to save $50K in a $15 million
worth of automation equipment. Admittedly, the reluctance of the agent/jobber in the
latter example can also be explained by other factors (e.g., complexity of the deal,
purchase category etc.) but the point is that there are implications not only for consumer
and small business contexts but also for million dollar decisions. Organizational buying
committees consist of individuals with different backgrounds, departments, levels of
expertise, interest, motivations, power, and so on. The study of the consistencies in
irrational buyer behavior is currently a hot topic for behavioral economists,
psychologists, and marketers. It would be fruitful to bring new decision-making insights
to the courts (Korobkin and Ulen 2000). It was shown that decision-makers may “over-
compete” on price to maximize the profit difference and perform significantly inferior
than rational price models that maximize the profits (Griffith and Rust 1997). Irrational
competitive responses such as incumbents’ herding to new markets have also been
observed (Debruyne and Reibstein 2005). Effects such as information availability (also
information primacy and recency), biases in representativeness, categorization, and
optimistic overconfidence can change inferences drastically (Tor 2002). As such,
temporal, internal and external strategic reference points (historical precedents/cognitive
constructs) complement prospect theory to explain managerial decision-making and
predict competitive reactions (Shoham and Fiegenbaum 1999). Accordingly, managers of
underperforming businesses would be expected to be more risk-prone (increasing
likelihood of predatory conduct) under pressure than those of businesses that are doing
104
well. Therefore, predatory pricing (e.g., dumping by distressed steel producers in Russia
and Brazil) may be attributed to high debt and deteriorating performance.
3.7.4 A Reaction Framework for Competitive Price Reductions
In Figure 3.6, I integrate the learning from the literature review and present a
reaction framework for competitive price reductions from a cost-based perspective. The
framework illustrates how the current law might be interpreted by executives, and has
implications for business marketers, managers, and policy makers.
Resource base obviously plays a role but I do not want to convey the message that
the business with the deeper pockets has the final word in predatory pricing engagements.
Many small suppliers/manufacturers are able to defend their markets effectively. An
important reason is that marketers typically develop ongoing relationships with their
customers. Obviously, developments in the CRM arena have enhanced the effectiveness
of this key phenomenon. The stronger the relationship, the lower is the price elasticity for
the product, hence the lower the threat from predatory pricing. A very strong form of pre-
emptive market defense against predatory pricing occurs when these relationships are
converted into binding commitments in the form of legal contracts (i.e., business cannot
lose the buyer for a pre-specified time period regardless of predator’s pricing).
Moreover, many bids are not granted to the lowest cost provider but to the value
package solution provider. Therefore, bundling becomes a strategic tool and is especially
effective when the business has differentiated itself. However, careful consideration is
advised to marketers when employing strategic bundling since the same tool can be used
for aggression. For example, 3M recently got in trouble for bundling rebates for its office
105
supplies anti-competitively (2003b).
In essence, the Post-Chicago School does not yearn for a return to the Pre-Areeda-
Turner era, it rather incorporates the learning from deregulation, modern strategic
thinking, advanced technology, the new economy and globalization. It concludes that
Areeda –Turner is inadequate to cope with today’s dynamic and complex environment.
It is my contention that the nature of the network industries is such that, when
considered in an integrated paradigm, the use of pricing strategies for pre- and post-entry
defense is not only feasible but also rational. I discuss the nature of the network
industries, develop a price competition framework for network industries, and present my
hypotheses next.
Yes
No
No Yes
Yes No
Yes
No
Price above ATC
Price below ATC
above AVC
No Yes
Price
below AVC
Figure 3.6: Incumbent’s Reaction Framework for Competitive Price Reductions
Read:
The Case of Accommodation: The incumbent businesses tend not to compete on price when challenged by
a company of the same type. Rather, they tend to emphasize quality, design and differentiate their
products/services. Alternative reactions include temporary price cuts to distract buyer attention on
competitor or try to cross-sell existing customers.
The Cases of React and Defend and Predation: These scenarios typically take place between a major
company and a discounter. Discounter cuts the price to a level that is disturbing for the incumbent. If the
incumbent chooses to accommodate, the discounter can establish a foothold of the market which it can use
to drive the prices further down. The incumbents typically choose the other route of aggressive
competition.
Has the entrant
cut the price?
Accommodate
(Intra-type competition)
-Non-price competition
-Differentiate product
-Run temporary price cuts
-Cross sell existing
customers
React and
Defend/Counter-attack
(Inter-type competition)
-Intensive competition
-Non-price predation
-Match price
- Selective discounting
-Increase capacity
How much has the
price been cut?
Recoup losses and
resume pre-entry
pricing strategy
Did the rival fold?
-Increase price?
-Exit?
-Reduce activity to non-
significant levels?
Retreat and
Restructure
-Restructure and
decrease AVC/ATC
-Cut product variety
and focus on
competency
-Acknowledge
defeat and accept
new market share
Predate
-Predatory pricing
-Non-price predation
-Flood the market with
capacity
-Cannibalize rival at
all fronts
-Subsidize losses
-Sue
Maintain price;
monitor (potential)
competitors
Will the price cut
considerably deter the
bottom-line?
Does the entrant have
deep pockets?
Is the price cut
introductory (non-
permanent)?
106
CHAPTER 4
NETWORK PRICE COMPETITION FRAMEWORK
The Network Pricing Strategy framework presented is in Figure 4.1. The
framework has three features that should be noted. First, it represents a dynamic process a
la Resource Advantage Theory. That is, two-way knowledge flows (competitive signals)
and competitive responses are embedded in the framework although not explicitly drawn.
It follows the logic from empirical industrial organization literature and the PIMS
paradigm in that the structure of the market, the competitive position, and competitive
strategies determine a business unit’s performance (Bharadwaj and Varadarajan 2004).
Incumbents perceive market entry to strategic markets with discounts as acts of
aggression and will consider retaliation. Similarly, the entrant choice for market entry
depends on the choice of pricing strategy of the incumbent and strategic assessment of
the market. The state equilibrium of the model relies on the incumbent’s and the potential
entrant’s behavior. For example, the optimal pricing strategy for the incumbent may
change even without new entry, it is only necessary to increase the incremental potential
likelihood of entry (i.e., aggregate strategic assessment) sufficiently. Equilibrium points
of dynamic models have been shown to be more realistic than those of static models that
assume perfect information (Coughlan and Mantrala 1992).
Second, the framework is focused on a special but important type of market entry,
inter-type entry (see section 2.4.2 for a discussion of inter-type competition). Inter-type
competition is based on price, whereas firms with similar cost structures compete on non-
price factors such as promotion and service (Allvine 1999; Palamountain 1955). The
107
108
Department of Transportation and independent based analyses have observed this
phenomenon (Oster and Strong 2001; DOT reports 1997-2000). “Entry by major carrier
on a point-to-point basis into another carrier’s hub has become very much the exception”
(Nannes 1999, p.4). The rules of competitive conduct between major (i.e., intra-type)
firms has been the focus of research in marketing strategy literature. It has been laid out
that conduct in these markets can be explained by leader-follower pricing system (Roy et
al. 1994).
The framework is domain specific and focuses on price competition among firms
with inter-type cost structures. In particular, it is assumed that the incumbents are major
firms with deep pockets and the new entrants are discounters (that are more efficient than
and have cost advantages over the incumbents but have shallow pockets). Therefore,
intra-type competitors and inter-type competitors form two strategic groups of
competition (Hunt 1972). There are “mobility barriers” that prevent changing group
membership as suggested by the Strategic Groups literature (Bharadwaj and Varadarajan
2004).
Third, the model is designed for network industries (or industries with network
characteristics –discussed next) which are typically characterized by high barriers to
entry. Generic (industrial organization school) and sustainable capabilities (R-A theory)
are also incorporated in the form of market and firm specific barriers. I posit that
developing a framework by integrating these theories reveal a scenario in which
predatory strategies can be rational.
Three main factors, market power of the incumbent, barriers to entry and strategic
assessment of the incumbent, act as antecedents for choice of the pricing strategy of the
109
incumbent. This is consistent with the classical industrial organization literature. It is also
suggested that firm specific factors should be more influential than market and industry
specific factors, which is consistent with the Resource Advantage theory.
The price levels in markets and strategic assessment of the incumbents lead to the
decision on whether or not to enter any given market. This is consistent with classical
economics and signaling theory literature.
Incumbents are expected to react sharply in markets that they consider to be
strategic with the intention to drive entrants out of their markets. This is consistent with
findings from industrial organization and strategy literature. If the inter-entrant is driven
out of the market as a result of retaliation and or lower strategic prospects, it is expected
that consumer welfare will be negatively affected. This is consistent with views from
marketing and Post-Chicago school of thought (see Figure 4.1).
Next, I discuss the network industries and then develop my hypotheses following
the five phases in the model: pre-emptive defense, entry decision, post-entry defense, exit
decision, and policy consequences.
Pricing
Strategy:
•Supra-competitive
•Competitive
•Limit
Incumbent’s
Response:
•Predate
•React
•Accommodate
Consumer
Welfare
BEFORE DURING AFTER
NETWORK PRICE COMPETITION FRAMEWORK
Incumbent’s
Market
Power
Incumbent’s
Strategic Assessment (t
0
)
-
Barriers to Entry
-
Inter-
Entry
Inter-
Exit
Entrant’s
Strategic Assessment (t
exit
)
-
Potential Entrants’
Strategic Assessment (t
0
)
Figure 4.1: Pre-entry and Post Entry Network Price Competition Framework
1
1
0
4.1 Network Industries
Network industries have a unique and dynamic nature of competition. There are
four main characteristics that differentiate network industries from others:
complementarity, compatibility, and standards; consumption externalities; switching
costs and lock-in; and significant economies of scale in production (Shy 2001).
Complementary products (e.g., PC and mouse) require compatibility for industry-wide
appeal and this leads to the development and adoption of standards. This notion is
especially emphasized by firms in information markets. Consumers’ perceived value of a
network increases as others use or adopt complementary or compatible products or
services which leads to positive consumption externalities (Katz and Shapiro 1985;
Lemley and McGowan 1998). High switching costs in three categories (e.g., procedural
(economic risk, evaluation costs, learning costs, and set-up costs), financial (benefit loss
costs, monetary loss costs), and relational (personal relationship loss costs, brand
relationship loss costs)) cause many consumers to continue with their existing providers
(Burnham et al. 2003). Network industries are also characterized by high start-up (fixed)
costs, and low unit production (marginal) costs (Watson 1998). Heil and Robertson
(1991) argued that incumbents’ reaction propensity increases with high fixed costs and
economies of scale. Network learning dynamics (within firm, between firm and end-user
based dynamics), surveillance, and shared resources also enhance the capabilities of
network firms (Dickson et al. 2001). Users in network markets may derive benefits from
the user network (e.g., e-mail), the complements network (i.e., Windows-based software),
and the producer network (level of competition) (Frels et al. 2003).
111
112
Industries such as airlines, telecommunications, and software have been at the core
of the growth of this society. Network industries are often associated with capital-
intensive industries serving mass markets. Historically, these industries have been heavily
regulated due to the neo-classical economics argument that the social welfare is better-off
when these industries are preserved as natural monopolies. However, the positive
consumer experience after deregulation of several of them (e.g., telecommunications) has
demonstrated otherwise. Other industries that possess network industry characteristics
include: broadcasting, cable television, electricity, water, pipelines, sewage systems, oil
pipelines, natural gas pipelines, road and highway systems, bus transport, truck transport,
inland water transport, ocean shipping, postal service, package delivery systems, refuse
pickup systems, airline computer reservation systems, bank automated teller machine
systems, bank and non-bank credit card systems, bank debit card systems, bank check
and payment clearance systems, local real estate broker multiple listing services, and the
Internet (White 1999).
An important feature of network externalities is the Positive Feedback Cycle (or
Snowball effect) –the greater the potential for growth, the higher the number of future
participants. Watson (1998) argued that as in a natural monopoly, a network industry
would usually converge towards a single product (e.g., Windows computer operating
system). A lock-in happens when the costs of switching from a network become higher
than gains due to externalities. A bad lock-in occurs if inferior products win over others.
Positive feedback cycles can cause bad lock-ins. That is why the Microsoft Case has been
paid high attention and linked to the future of antitrust (2000b).
Positive feedback cycles can enable a firm to achieve monopolization very
113
quickly. Once achieved it may be even harder to resolve the monopolization due to
network effects (Rajiv 1998). Indeed winner-take-all (or most) competition is commonly
observed in network industries with snowballing --the more the customers, the more
valuable the product/service, the more the attractiveness of the product/service to even
more customers (Valente 1995). Customer switching costs become higher over time.
Also, due to the high-fixed costs of network firms, increasing sales decrease average
costs substantially. These factors can easily motivate an Internet company to give its
products for free (e.g., Adobe Acrobat Reader) to penetrate the market and set the
standard (Evans and Schmalensee 2001). The antitrust standards become irrelevant
because of the unique dynamic competition in the network industries to capture the
dominant position in the markets.
Network industries are very visible and include some that are key to the welfare
of the society such as utilities, software, telecommunications, cable services, credit cards,
and transportation (e.g., airlines). Thus, they are of uppermost concern for antitrust
enforcement bodies. The current antitrust approach is characterized by heavy use of cost
analysis and is not appropriate for all cases in dynamic markets, especially where high
level of fixed costs (e.g., networks) are involved (Sievers and Albery 1991).
Evans and Schmalensee (2001) argued that there is no cost-based test to
distinguish predatory innovation from non-predatory innovation in a winner-take-all
setting. The marginal cost of a network firm (with excess capacity) is often negligible.
Average Variable Cost (AVC) test provides a safe harbor for network firms enabling
them great flexibility to drop prices for predatory reasons. Network effects can enhance
the effects of market power particularly in dynamic industries. Stifled innovation may
114
cause bad lock-ins to obsolete technologies, resulting in additional loss of welfare. At
the same time, it is common for such companies to price their products/services below
AVC for non-predatory reasons (e.g., introductory low prices to penetrate the market).
The interconnections in network markets and the high tendency of market power in one
market to spill-over to other connected markets make the defense of strategic markets
even more critical.
4.2 Hypotheses
4.2.1.1 Market Power
Typically, the existence of market power has been simply inferred when an
incumbent has pre-dominant market share (1966, p.571). Firm(s) may use a monopolistic
market structure or conduct to achieve market power, which it then uses to achieve better
performance at the expense of its competitors and the competitive process. I define
market power as the ability to influence (pricing) conduct in the marketplace for the
purposes of the current research.
Structure-conduct-performance paradigm implies that higher the concentration in
a marketplace the further it is from the case of perfect competition, the higher the prices,
and the lower the societal welfare (Lopez 2001). Previous studies have generally shown a
positive correlation between market concentration and industry profitability (Shepherd
1970) among other factors.
Establishing the relationship between market power and pricing strategy is also
important from a public policy perspective because, in order to win a case, one of the
elements that the plaintiff must show is that the incumbent had market power to act upon.
It must establish that the incumbent used or pursued market power by anti-competitive
115
means.
Demonstration of market power is typically sought through high market share.
The structure of a market may be derived from the study of the number and size of firms,
their cost and demand conditions, product differentiation, the nature of any entry barriers,
and degree of regulation. Industry Concentration Ratio (percentage of total sales of the n
(usually four) largest firms in an industry) and the Herfindahl-Hirschmann Index ((HHI)
–sum of the squared market shares of all firms) are two commonly used market structure
indicators.
Pre-entry:
The incumbent’s pricing strategy for a specific market is heavily influenced by its
market power for the market in question. Market price premiums are conceptualized at
the three levels of strategic pricing options: supra-competitive, competitive, and limit
pricing, which is intuitive and consistent with the microeconomics view of market
performance.
1. Supra-competitive pricing: This refers to pricing at levels that yield premium margins
for the incumbent. This type of pricing is typical of a monopolist. The incumbent is able
to reap profits at levels that would not otherwise be possible because of a lack of
competition in the market place. Neo-classical economic theory suggests that supra-
competitive prices cannot exist in the long run because of the threat of potential entrants
(a.k.a. theory of market contestability). However, industry observations have shown that
firms can and do differentiate themselves and as a result, they are able exert consistent
116
price premiums in the marketplace especially when barriers to entry are high (Baker and
Pratt 1989).
2. Competitive Pricing: This refers to the going price levels in the marketplace in which
the incumbent makes reasonable profits (i.e., industry average for comparable markets).
Neo-classical theory would predict that competitive profits would converge towards zero
in the case of perfect competition. However, since markets are not perfectly competitive,
we observe modest levels of profits in competitively priced markets.
3. Limit Pricing: As discussed in the introduction, this strategy is also called entry
deterring pricing (Porter 1980, p.14). With limit pricing, the incumbent prices its services
low so as to prevent competition. Thus, the price is intentionally set low so that entry is
discouraged and (lower) profits are secured for a longer time. The apparent signal is low
current and future profits (Bain 1956). It should be noted that an entry-deterring price
does not necessarily have to be below (variable) cost to be anticompetitive. It could
negatively influence the entry decision of a more efficient firm at above cost levels.
Potential small scale entrants may assume that the incumbent enjoys economies of scale,
and potential large scale entrants may assume that the total demand is inelastic (i.e.,
increased supply will lead to even lower prices) (Gruca and Sudharshan 1995).
Limit pricing can also be justified by interests in protecting market leadership or
to prevent competitor growth in strategic territory. In doing so, the incumbent may utilize
signal-jamming to influence to potential entrants’ to believe that their costs are lower
than they actually are. The result would be little or no profits in markets where limit
pricing is employed. Similar to predatory pricing, the evidence so far is that the use of
limit pricing is rare (Smiley 1988). Low prices do not deter entry (especially against
117
innovative entrants, e.g., the cases of disruptive technologies or leaps in production
efficiencies (Christensen 1997; Han et al. 2001)), but they do diminish the chances of the
entrant’s survival in a given market (Lieberman 1989). Furthermore, limit pricing has
been documented to be a rational strategy in game-theoretic experiments (LeBlanc 1992;
Milgrom and Roberts 1982a). Thus, network industries may prove to be the ideal setting
to show that there may be more limit pricing in practice than meets the eye. If limit
pricing can stimulate and absorb the demand it stimulates, the residual demand upon new
entry may be too low to justify the entry effort (Eliashberg and Jeuland 1986).
Furthermore, both industrial organization literature and PIMS based studies
suggest that it pays off to be the dominant firm in an industry (Buzzel and Gale 1987;
Sudharshan and Kumar 1988). “The superior financial performance of businesses with
large market shares is attributable to their ability to obtain inputs at lower costs, extract
concessions from channel members, and set prices rather than be price takers” (cf.
Bharadwaj and Varadarajan 2004, p.223). Aggregate market concentration and
performance relationship has been found to be positive and significant in two meta-
analyses (Capon et al. 1990; Dutta and Narayan 1989). There is a well established link
between market power and price levels in economics and strategy literature (Abunassar
1994; Borenstein 1989) and it is expected that this will hold true for the network
industries as well.
4
Therefore,
H1: Market power of the incumbent and the market price premium will be
positively associated.
4
It should be noted that some scholars have argued that the market share-performance
relationship does not exist and may be attributed to third factors (e.g., Symnanski et al. 1993).
118
4.2.1.2 Barriers to Entry
“Entry conditions are central to antitrust analysis”
John M. Nannes, Deputy Assistant Attorney General, Antitrust Division (1999).
Barriers to entry can be defined as factors that prevent an entrant from competing
on an equal footing with the incumbent(s) (Dolan 1986, p.602). For a particular industry,
if the barriers to entry are high and retaliation from the incumbents is expected, potential
entrants are not likely to be enthusiastic about entry (Mintzberg and Quinn 1996).
Karakaya and Stahl (1989) used a broad categorization of the concept and identified
nineteen barriers to entry. Their list included cost advantage of incumbents, product
differentiation of incumbents, capital requirements, customer switching costs, access to
distribution channels, government policy, advertising, number of competitors, research
and development, price, technology and technological change, market concentration,
seller concentration, divisionalization, brand name or trademark, sunk cost, selling
expenses, incumbent’s expected reaction to entry, and possession of strategic raw
materials (essential facility). They later advanced this list to 25 items for consumer goods
markets (Karakaya and Stahl 1992).
Porter (1980) identified six major categories for barriers to entry. These six
barriers were cost advantages of incumbents, product advantages of incumbents, capital
requirements, customer switching costs, access to distribution channels, and government
policy. Barriers to entry may potentially enable the firm to recoup profits over a longer
term and make predatory pricing rational. Multiple industry data also indicate that firms
use entry deterrence less frequently when other barriers exist (Bunch and Smiley 1992).
The model posits that the market power of the firm is a primary determinant of its
119
pricing strategy. However, this link should hold true and probably be even stronger with
the inclusion of the comprehensive barrier to entry measures that include firm specific
(i.e., brand equity) and those determined by outside forces (e.g., regulation --tobacco
settlement that requires new entrants to pay additional State taxes along with the industry
giants that originally caused the case (Forbes 2005)). The industrial organization
literature argues for the positive effect of barriers to entry whereas the efficiency school
suggests that barriers are not necessary for superior performance (Bharadwaj and
Varadarajan 2004). This leads to the following hypothesis:
H2: The positive relationship between the market power of the incumbent and the
market price premiums will be positively moderated by barriers to entry.
It is also possible that barriers to entry could have a direct effect on the pricing
strategy of the incumbent, thus forming a quasi-moderation effect. This should be
investigated as an alternate specification. Therefore,
H3: Barriers to entry and the market price premium will be positively associated.
Of particular interest here are the firm specific barriers that can be enhanced as
opposed to market specific barriers (e.g., regulation) that are generally beyond the control
of the incumbent. Neoclassical price theory traditionally emphasized the choice for the
industry as the strategic decision whereas the Resource Advantage theory predicts that
firm specific barriers would be more important than market specific barriers (Viscusi et
al. 1995). Varadarajan and Jayachandran (1999) observed that the earlier focus in
industrial organization literature (Bain 1956) on “why some firms are more profitable”
later shifted to “why some firms are more profitable” (Demsetz 1973), and ultimately, the
resource-based view of the firm (Barney 1992) where profitability is determined by
120
competitive advantage. Hunt (2002, p.287) observed the state of the theoretical debate on
the firm specific versus industry (i.e., market) specific factors:
By the time of McGahan and Porter (1997), the entire nature of the debate
over firm performance had changed dramatically. Originally, advocates of
industry-based strategy (e.g., Montgomery and Porter 1991) were citing
Schmalensee (1985) to justify their focusing on “choosing industry” as the
key strategic decision. After Rumelt’s (1991) replication and extension of
Schmalensee found firm factors to account for almost six times the
variance of industry factors (46 percent vs. 8 percent), the debate shifted
toward whether industry choice at all, Thus, McGahan and Porter’s
(1997) study, which finds that firm effects dominate industry effects by
only 36 percent to 19 percent, is interpreted by its authors as confronting
the challenge from Rumelt and others that industry, far from being key,
doesn’t seem to matter at all. The point to be emphasized here is that no
one now claims empirical support for the neoclassical position. That is,
after Rumelt’s (1991) and other studies, no one argues seriously the
neoclassical position that either industry is everything or industry effects
dominate firm effects.
McGahan and Porter (2002) found even lower influence of industry factors on
performance than their earlier findings (48 percent business-unit factors and 10 percent
industry factors). Still, market-specific barriers (e.g., institutional environment) can be
important from a public policy perspective since all types of barriers can be antecedents
to monopoly conditions in a market (Yip 1982). Karakaya (2002) reported four major
underlying dimensions of barriers (i.e., firm specific, product differentiation, cost of
market entry (financial requirements), and profit expectation of entering firms) with firm
specific barriers being the most important. His study was in an industrial setting. I inquire
if this finding will hold true in network industry setting as well. The relative effectiveness
of these two categories is also interesting to verify good, and potentially bad and ugly
effects that barriers to entry may have (Han et al. 2001). Therefore,
H4: The positive moderating effect of firm specific barriers on the incumbent’s
pricing strategy will be higher than that of market specific barriers.
121
Firm specific barrier influence would also be expected as direct effects.
Therefore,
H5: The positive effect of firm specific barriers on the incumbent’s
pricing strategy will be higher than that of market specific barriers.
These hypotheses, if supported would also have managerial implications since
building firm specific barriers is a prerogative that managers typically do not have with
market specific barriers.
4.2.1.3 Strategic Assessment
The model also alludes to the role of potential entrants’ strategic assessment for
entry for the market in question. Game theoretical analyses have indicated that
competitors’ moves are calculated by investigating its resources and entry patterns (e.g.,
Milgrom and Roberts 1982b). The resource-advantage theory advanced by Hunt (2000)
suggests that resources play a critical role in the long term prosperity of the firm. Firm
growth is constrained by internal management resources (Penrose 1959). Available
resources (e.g., human resources, capital) constrain the choice of markets for entry
(Wernerfelt 1984b). For example, employees (relations), type and capabilities of the
aircraft, and management leadership that potential entrants have can also be potential
resources in the airline industry context. The fit of the market with potential entrants’
existing portfolio would also be important. It is not likely that an existing airline would
start flying a route that is not connected to its existing routes on either end (Dixit 2000).
The attractiveness of the market is also an important part of the strategic assessment
(Baldwin 1995) as growing markets are more likely to be under the threat of new entrants
122
both in the form of existing or start-up competitors (Gatignon et al. 1989). The successful
analysis and perception of these issues have been linked to higher performance (Clark
and Montgomery 1996). The incumbent is less likely to employ supra-competitive
pricing and make the market more attractive for entry if it deems that the potential for
entry for a given market by inter-competition is high after strategic assessment.
Therefore,
H6: The positive relationship between the market power of the incumbent
and the market price premiums will be negatively moderated by the
incumbent’s strategic assessment (i.e., resources, strategic fit,
market growth) of the potential entrants.
H7: Incumbent’s strategic assessment and the market price premium will be
negatively associated.
4.2.2 Entry Decision
The questions that an incumbent should normally answer before reacting to a
competitor’s price cut are: “is the price cut likely to have significant impact on our
sales?” and “is it likely to be a permanent price cut?” If the answer to any of these two is
negative, then there actually is no need to react to a price cut (Kotler and Armstrong
1991) (also see Figure 3.8). However, in practice (i.e., with information asymmetry), it is
not very easy to answer these questions accurately, especially with a strategic perspective
and that encompasses uncertainty. A price cut insignificant today could reshape the
industry landscape tomorrow (e.g., the case of Dell Computers). Similarly, price levels
that are below cost and non-sustainable today could become sustainable with economies
of scale tomorrow (e.g., eBay.com, Amazon.com). Incumbent firms in network
industries, especially when challenged by low-cost start-ups with drastically lower cost
structures, tend to fear the worst (e.g., Microsoft versus Linux; American versus
123
Southwest Airlines). In high–tech industries, this fear is partially transformed into
defensive acquisitions by leading firms (e.g., acquisitions by Ticketmaster; Microsoft;
Cisco). However, in recently deregulated industries where merger and acquisitions have
to be approved after a detailed antitrust scrutiny, it can be quite rational to employ non-
price and predatory pricing schemes. Thus, managers in these firms may answer
affirmatively when they strategically analyze the two important questions. Recoupment
may not be feasible in the short run and not even in the foreseeable future. But perhaps
these firms are sacrificing gains and incurring losses in exchange for lower future losses
or even survival.
The entrant can fairly accurately observe the pricing strategy of the incumbent.
Upon analysis of its resources, strategic fit, the market growth, and the current pricing
strategy of the incumbent, the entrant decides whether or not to pursue entry to the
market. The framework suggests that the response of the incumbent will be heavily
influenced by its pricing strategy for that market. If the incumbent were using supra-
competitive pricing, it is anticipated that it will predate (sharp price and capacity
reactions) to protect its market power, if the incumbent were using competitive pricing, it
is anticipated that it will react competitively (reasonable price and/or capacity reactions),
and finally if the incumbent was using limit pricing, it is anticipated it will accommodate
(insignificant price and/or capacity reactions). This structure has not been previously
empirically tested, however, content analyses of reports from the popular press support
these assertions. In the predation case, the odds are against the entrant to establish itself
and succeed in the marketplace (exceptions such as Southwest Airlines can occur
depending on the strategy, efficiency and the funds available). In the competitive reaction
124
and accommodation cases the entrant is increasingly more likely to survive given that it is
more efficient than the incumbent and managed well.
The following set of hypotheses are also supported by the signaling literature
which suggests that price levels are indicators of market potential and the cost structure
of the incumbent (Heil and Walters 1993, Prabhu and Stewart 2000). Heil and Robertson
(1991) argued that that the major benefits from signaling are preemption and
development of competitive norms of conduct, and proposed that market power and
antitrust action due to price signaling would be positively related. This verifies the
before-during-after price modifications pattern previously mentioned.
Adam Smith (hence, the classical industrial organization literature) would also
predict that capital flows into markets with above average returns on investment. This
capital flow could be in the form of new market entries. Still, this relationship should not
be taken for granted. Dixit (2000) hypothesized that the higher prices would lead to
higher probability of entry to markets (in the airline industry) but was perplexed by a
negative relationship. Therefore,
H8: Incumbents’ pre-entry market price premiums for markets with inter-category
entry will be higher than those of markets without inter-category entry.
The rationale for the consideration of strategic assessment for Hypothesis
7, and the Resource Advantage theory also applies to the entry decision phase.
Therefore,
H9: Potential inter-category entrants’ strategic assessment of the markets will be
higher for the markets that they enter.
125
Pre-entry defense summary: The model posits that the main positive effect of market
power on pricing strategy is moderated by the incumbent’s strategic assessment of
potential entrants’ and the barriers to entry. The moderating effect of barriers to entry is
positive. However, the moderating effect of strategic assessment is expected to be
negative. Direct linkages from both factors to pricing strategy are also plausible. Ceteris
paribus, the trade-off between these countervailing forces determines the pricing strategy
of the firm. If the positive effects are dominant, the firm is likely to employ a supra-
competitive pricing strategy resulting in large fare premiums (and loss of consumer
welfare). If there is a balance between these effects, the firm is likely to employ
competitive pricing. Finally, if the negative effects are dominant, then the firm is likely to
engage in limit (entry-deterring) pricing which results in lower prices in the short run but
has negative welfare effects in the long run due to lessened competition (and decreased
efficiency due to stifled competition and potential innovation).
Potential entrants’ own strategic assessment and the pricing strategy of the
incumbent lead to entry analysis. The framework flow is interrupted, and equilibrium is
reached if there is no entry. However, if the challenger decides to enter the market, a
response by the incumbent is triggered as measured by a price change. This response may
be categorized at three levels, which are discussed next. The incumbent may choose to
predate, to react or to accommodate depending on its pre-entry strategy. If the entrant is
able to endure the retaliatory response of the incumbent, a new market equilibrium will
be observed.
126
4.2.3 Post-entry Defense
Trend analysis indicates that the firms react sharply when an entry by a discounter
(i.e., inter-type entry) occurs in their markets (Oster and Strong 2001). Previous work on
competitive interaction also suggests that retaliatory actions to entry depend on the
perception of threats (Kuester et al. 1999). Namely, it is expected that the incumbent
would react most sharply in cases where its supra-competitive profits are threatened. As
discussed in Chapter 3, the current tests of predatory pricing are based on average
variable cost, the courts are not effective in distinguishing predatory action from vigorous
competition in network industries. Therefore, the extent of paired price and capacity
reaction could easily reach the scale of predation without being detected by the courts.
The evidence for that would be consistent with observations of market exits by the inter-
category entrants (presumably more efficient than the incumbents) which were likely to
survive had predatory tactics not been utilized. Supra-competitively priced markets
improve the profit margins and are more valuable and strategic from the perspective of
the incumbents. Hence, the expected incumbent retaliation in the case of inter-entry to
supra-competitively priced markets is predation. The reaction of the incumbent will likely
not be as drastic in the case of entry to competitively priced markets. Finally, the
incumbent is expected to react little or not at all in the case of entry to markets where
limit pricing was employed (i.e., accommodation).
The logic of accommodation is supported by the widely established Defender
model (Hauser and Shugan 1983) which predicts that an opposite reaction such as cutting
back on advertising or increasing price can be the optimal strategy against market entry.
Moreover, there are several PIMS based studies that reported that no or limited reaction
127
to market entry is the norm (Biggadike 1979; Robinson 1988; Yip 1982). Retaliation was
associated simply with high-growth markets.
On the other hand, price reduction was found to be the optimal response against
competitive entries (Gruca et al. 1992). It is well documented that hostile acts trigger
stronger competitive actions (Heil and Walters 1993). Due to the interconnected nature of
the networks, the number of markets that are considered strategic could be even higher.
MacMillan and colleagues (1985) have reported that such strategic challenges accelerated
competitive response. Swift and more aggressive responses are expected if the focal
market of entry is viewed important by the incumbent (Chen and MacMillan 1992; Chen
et al. 1991). Assuming that firms derive their market power from strategic emphasis to a
given market, and that market power and price are positively correlated, it is only natural
that the entries to markets that are supra-competitively priced would attract more
intensive retaliation to entry than others. Bowman and Gatignon (1995) also concur that
the retaliation is delayed when the incumbent has low market share. An incumbent that
does not react to entry to markets where its profits lie may send signals of weakness to
potential competitors and invite further entry. “If the strategy fails and entry occurs,
consequences for the incumbent firm can vary depending on which strategy was chosen
[prior to entry]. Certain strategies may leave the incumbent in a worse competitive
position, whereas others may lead to a stronger posture after entry ” (Gruca and
Sudharshan 1995, p.44). For example, “the airline industry has certain characteristics that
make a predatory theory more than plausible” (Nannes 1999). To match a move is a
strong signal by itself, indicating unwillingness to give up a position without escalating
the war to mutually destructive levels (Chen and MacMillan 1992). Robinson (1988)
128
observed that the current strategy may influence future as well as the competitive
conduct. Therefore,
H10: The magnitude of the incumbent’s response to inter-entry will be
positively associated with its pre-entry pricing strategy.
4.2.4 Exit Decision
As detailed in previous sections, the inter-type entrants are typically subject to
aggressive price cuts when they enter strategic markets. The incumbents generally target
their responses so that the inter-entrants are driven out of their key markets before they
can establish their structure and inflict considerable damage to the incumbents. However,
since the inter-entrants are presumably more efficient than the incumbents due to their
operation and cost structures, their exit patterns would also be affected by their financial
resources and the price elasticity of the market. When discounters enter markets, the
passenger volume typically increases significantly more than the percentage decrease in
average price. Therefore, the decision to exit the market will be effected by the
incumbent’s response and the post-entry strategic assessment of the entrant.
Incumbent’s choice for reaction strategy sends a signal to the entrant as to how
determined the incumbent is to deter entry (Heil and Walters 1993). A price matching
move is a strong signal in its own right (Shelling 1960), however this move becomes
even more powerful given the context that the entrant is a discounter with less a
comprehensive value proposition. Accurately perceiving the intent of competitive
reactions enhances firm performance (Clark and Montgomery 1996; Day and Nedungadi
1994). This has further implications than just sheer financial impact of price reductions.
A sharp price cut may cause the entrant to leave the markets upon strategic assessment
129
even though it might have made more economic sense to fight back due to a more
efficient structure of the entrant (Gundlach 1995). Successful retaliation has been
associated with holding a competitive advantage (e.g., brand equity), low scale of new
entry, and low to medium access to resources by the new entrants (Robertson and
Gatignon 1991), and all these conditions are met in the inter-type competition framework.
In essence, the entrant weighs the strategic benefits against the threats posed by the
incumbent’s retaliation and makes the decision to exit or stay. Therefore,
H11: The likelihood of inter-category exit from a given market and the magnitude
of the incumbent’s competitive response will be positively associated.
H12: The likelihood of inter-category exit from a given market and the entrant’s
strategic assessment will be negatively associated.
4.2.5 Policy Consequences
Increased long-term social welfare is the ultimate goal of antitrust. Lower prices
and competition increase the welfare of the consumers and the society in general (Grewal
and Compeau 1999, p.3). However, if the consumers feel that the price they have to pay
is unfair, then social harm may occur (Guiltinan and Gundlach 1996a). The courts
consider economic harm to consumers as the best way of assessing harm to society (Baer
1996). However, a long-term assessment of social welfare should not be simplified to
allocative efficiency (Guiltinan and Gundlach 1996a). For example, Bloom and
Gundlach (2001a) identified thirteen paths through which marketing affects consumer
welfare.
130
Economic Efficiency
Perspective:
Chicago:
• Allocative Efficiency
• Productive Efficiency
Post-Chicago
• Dynamic Efficiency
Marketing Oriented
Perspective:
• Affordability
• Variety
• Satisfaction
• Convenience
• Price Stability
• Perceived Value
Innovation
Competition
Figure 4.2: Marketing Extension for Consumer Welfare Assessment
Consumer
Welfare
Figure 4.2: Marketing Extension for consumer Welfare Assessment
The long-term consequences of predatory pricing (or any anti-competitive action
for that matter) on consumer welfare are central to antitrust analysis. Marketing insights
for a more informed consumer welfare analysis are presented in Figure 4.2 and Table 4.1.
The shortcomings of just focusing on allocative efficiency for assessing consumer
welfare, and marketing’s potential contributions for its operationalization and conception
have been observed by many scholars in marketing and elsewhere (cf. Gundlach et al.
2002).
131
Table 4.1: Dimensions of Consumer Welfare and Marketing Insights
Dimensions of
Consumer
Welfare
Marketing Insights Public Policy
Implications
Efficiency Local inefficiencies
may disrupt the general
efficiency of the
markets
An antitrust analysis that
goes beyond economic
efficiency
Affordability Focus on net income is
too narrow
Affordability is a better
measure of welfare than
income
Satisfaction Established scales such
as SERVQUAL exist
Track satisfaction from
market as well as market
concentration
Variety Broader definition of
variety
Discount for seemingly
variant offerings. Let the
consumer define variety.
Price Stability Fluctuating prices are a
great nuisance to
consumers. Price
confusion is
detrimental to social
welfare
Active public
communication/advertising
and legislation to prevent
price confusion
Convenience Non-cost based tests
may be relevant; the
level of convenience
can serve as such a
measure
Monitor the level of
convenience before/after
predatory action as well as
price levels. (side note:
investigate unjustified
convenience charges)
Innovation Product Life Cycle
concept and impact of
line/brand extensions
versus radical
innovations
Emphasis to prevent bad-
lock-ins to obsolete
technology. Support
superior technology by
adoption/promotion
Competition Track market share and
sales but also
signaling, reputation
effects, and strategic
decision making
Fund and utilize research
that quantifies the impact
of signaling, reputation,
and strategic decision
making
132
4.2.5.1 Consumer Welfare
Affordability. There is more to measuring social welfare than efficiency. Purchasing
power is a classic variable considered by the Chicago School/courts in terms of price
levels. From a marketing perspective, purchasing power goes beyond the notion of the
basic price itself, but also encompasses the notion of affordability which includes credits
terms, payment options, etc.
Variety. Variety (choice) is a variable that is of paramount importance for consumer
welfare (Guiltinan 2002; Lande 2001). A high number of players in an industry does not
necessarily imply variety in commoditized markets. Therefore, perceived (buyer
defined/unobservable) variety mix is more important to capture than simply supply
variety. Marketing implications for a choice-centered antitrust policy have been identified
from both demand and supply perspectives (see Guiltinan 2002).
Satisfaction. Buyer satisfaction is a key construct in marketing, yet considered to be too
difficult to deal with and ignored in welfare analyses by economists. Marketing has a
significant history of research with the satisfaction construct for both products and
services and has established scales to measure it. Historical data can be used to estimate
the impact of a predatory strategy on buyer satisfaction in the long run.
Convenience. Overall convenience associated with the use of a product or consumption
of service can also be an important dimension of satisfaction and welfare. Related
dimensions include decision-making, access, transaction, benefit and post-benefit
convenience (Berry et al. 2002). Levels of convenience provided to the buyers before the
alleged predatory action should be compared to those of the period following it.
Price stability. Predatory pricing typically results in drastic price fluctuations especially
133
in network industries. For example, in one DOJ alleged case of predatory pricing,
American Airlines cut its prices by 26 per cent, but then raised it by 84 per cent after
driving out the competition (Carney and Zellner 2000). Fluctuating demand and prices
can prove fatal especially for small manufacturers that do not hedge their risks
effectively. Similarly, unstable prices can be a source of great inconvenience and
confusion for the consumers. Buyers may feel the price they paid is unfair (Smith and
Nagle 1995; Zeithaml 1988) or they may even feel betrayed. At the same time, it is
common for price-discriminating businesses to fuel price confusion among buyers in
order to avoid competing on price (Grewal and Compaeau 1999). Since each of these
options diminishes consumer welfare, the social implication is to communicate/advertise
and sometimes even intervene to prevent price confusion and unnecessary price
fluctuations.
The positive impact of lower prices on consumer welfare is generally accepted.
However, predatory pricing is detrimental to consumer welfare in the long run because
once the competitors exit the market, the predator raises prices with the intention of
collecting supra-normal profits and recouping its predatory investment. Moreover, the
problem with predatory pricing is not limited to harm to consumers through the increase
of prices back to monopoly levels. Predatory pricing, successful or not, can potentially
reduce incentives for investment and innovation, and prevent new entry or expansion by
more efficient firms. There are special implications for network industries such as
telecommunication and software where the value of the product/service increases along
with the number of users. Innovation can be stifled when predatory prices induce
134
consumers to continue to use an old technology as opposed to a superior alternative
offered by a new entrant (Guiltinan and Gundlach 1996a).
Grewal and Compeau (1999) suggested that marketing researchers have not
engaged in public policy implications of pricing until recently and that a focus on this
issue is long overdue. They argued that developments such as the internet, global
markets, mega-corporations, and cooperative marketing arrangements created the
necessity of taking a closer look at the pricing and public policy interaction with
consumer welfare in mind. After all, courts consider economic harm to consumers as the
best way of assessing harm to society (Baer 1996). Guiltinan and Gundlach (1996a)
argued that marketing was in a unique position to help form public policy guidelines with
comprehensive measurement and modeling procedures, and that predation and predatory
pricing have not been addressed by marketers until recently. Gundlach (1995) suggested
that the marketing discipline had the potential to further the understanding needed for the
development of a more suitable antitrust policy.
In all cases of exit by inter-entrant there will be a loss to consumer welfare
stemming from the assumption that the inter-entrant was presumably more efficient. The
loss to consumer welfare (mainly in terms of purchasing power and price stability) occurs
because the incumbents are typically able to increase the prices to pre-entry levels and
sometimes even higher to recoup losses incurred during predation. The long run effect of
predation is higher prices for the consumers.
It appears that there is an increasing gap between the insights from the modern
economic theory and the enforcement of current judicial policy. Government
enforcement concern has never been higher in many years. The new economy requires
135
new rules for the assessment of predation because of the growing importance of
intellectual property (e.g., Microsoft Case). Increasing market concentration in many
industries and number of mergers, are also of concern (Bolton et al. 2000). Obviously,
measuring social welfare with comprehensive measures itself constitutes a dissertation
topic and is well beyond the scope of this dissertation. However, the market power of the
incumbent can be practically used as a proxy of long term social welfare and the
following hypotheses can be tested. Therefore,
H13: The inter-entrant’s exit will result in loss to consumer welfare (as
measured by the increase in the incumbent’s post-exit market power).
If supported, the above hypothesis would undoubtedly generate public policy
implications. The overall premises of the proposed framework are summarized in Table
4.2.
136
Table 4.2: Summary of the Framework Premises
Context Theoretical Base
H1-
H7
Antecedents of Pricing Strategy:
• Market power leads to higher price
premiums
• Incumbents consider threat of potential
entrants when deciding strategy
• High barriers to entry prevent market
contestability
• Firm specific barriers more influential
than market specific barriers
Industrial Organization, Strategy,
Resource Advantage, Market
Segmentation
H8-
H9,
H12
Antecedents of Market Entry:
• Entrants will consider price levels as
indicators of efficiency and potential
• Entrants will conduct strategic analyses
before entry and exit
Industrial Organization,
Signaling, Game Theory, Entry
Deterring Prices, Reputation
Effects, Multi-market competition
H10
Consequences of Entry:
• Incumbents will respond differently
depending on the market of entry
Competitive Interaction, Inter-
Intra-type Competition, Multi-
market competition, Game
Theory
H11
Determinants of Market Exit:
• Entrant is likely to exit in the face of
drastic response by the incumbent
Competitive Interaction,
Signaling, Resource Advantage,
Game Theory
H13
Consequences of Market Exit:
• Exit by will have negative
consequences on long and short-term
consumer welfare.
Post-Chicago Economics, Public
Policy and Marketing Interface
137
CHAPTER 5
CONTEXT: AN INQUIRY OF THE DYNAMICS OF COMPETITION IN THE
U.S. AIRLINE INDUSTRY
It was a love of the air and sky and flying, the lure of adventure, the
appreciation of beauty. It lay beyond the descriptive words of men –where
immortality is touched through danger, where life meets death on an equal
plane; where man is more than man.
Charles Lindbergh, The Spirit of St. Louis. 1953
This is a nasty, rotten business.
Robert Crandall, American Airlines, 1994 (Petzinger 1995).
The airline industry has certain characteristics that make a predatory
theory more than plausible (Nannes 1999).
An inquiry of price competition in a key network industry –airlines, provides a
unique perspective into the dynamics of competition. The main contributions of this
chapter are two-fold: it introduces the deregulated passenger airlines as an ideal context
for empirical research, and describes the history and competitive landscape of the
industry.
Several unexpected consequences were observed as a result of the airline
deregulation in 1978. These included increased range of services at the expense of overall
service quality, the dominance of hub-and-spoke systems, and the subsequent fare
structure that penalized passengers flying out of major hubs in the form of hub premiums.
Contrary to popular belief, the effect of deregulation on decreasing fares was not robust.
In comparison to pre-deregulation era, tickets today cost less on longer haul routes
138
regardless of the nature of competition, but not on shorter routes where low cost carrier
(LCC –such as Southwest or Air-Tran) competition is absent. Waves of LCC entrants
were driven out because of cut-throat competition and/or predatory practices of the major
carriers. The existence of such practices is suggested by the observations and actions of
the DOT and DOJ.
The eighties saw the emergence of yield management systems, frequent flier
programs and the widespread use of computer reservation systems. The primary
distribution and sales channel was the travel agencies, which lost much of their power
after the explosion of the Internet, and the subsequent movement to cut their commissions
throughout the industry.
After September 11, the LCCs have performed better than the major carriers,
which could not respond to increased pressure to cut costs. Anticipated near future trends
in the airline industry include consolidation attempts among the major carriers,
subsequent antitrust action, revision of the hub-and-spoke system and re-organization,
restructuring of corporate travel policies, and increased market access by LCCs aided
either by federal or local governments.
5.1 Historical Perspective
5
As the phrase “America on wheels” indicates, cars are often referred to as the
foundation of American culture. However, by the 1990’s the number of adult individuals
who owned cars were less then those who had flown (Petzinger 1995, p.i13). Convenient
access to air service is one of the key considerations for choosing business locations. This
5
The historic facts in sections 5.1.1, 5.1.2 and intro of 5.1.3 have been adopted from the Air Transport
Association web site (www.air-transport.org) except where cited.
139
chapter will present the rich history of the U.S. Commercial Aviation with its ups and
downs and bring the reader to date. An overview is provided in Table 5.1.
Table 5.1: The U.S. Airline Industry Timetable
The Early Years
1903 First flight by Wright Brothers at Kitty Hawk, NC.
1914 The first scheduled air service starts between St. Petersburg and Tampa,
FL.
1918 First Airmail delivery
1925-26 The Contract Air Mail Act and The Air Commerce Act of 1926 passes
1933 The first modern passenger airliner the Boeing 247 can carry 10
passengers
1934 Air Mail Act of 1934 is enacted due to scandal regarding the mail bidding
process
1938 The Civil Aeronautics Act of 1938 is passed. CA Board commences
1940 The first plane with pressurized cabin, the Stratoliner is introduced by
Boeing
1943 The U.S. builds its first jet plane Bell P-59
The Jet Age
1958 Boeing remodels and introduces its KC-135 jet tanker as the first U.S.
passenger jet, the 707 (capacity 181 passengers)
The Federal Aviation Act is passed to monitor the growing airline industry
1963 Sabre Computer Reservation System introduced by American
1967 Department of Transportation (DOT) is created
1969 Boeing launched its first widebody jet the 747 in 1969. It could carry 450
passengers. The first supersonic plane the Concorde is introduced
1978 The Airline Deregulation Act is passed. A new era begins…
New Entrant Low Cost Carriers (LCCs) start to compete for markets
1981 American offers the first frequent flier program: AAdvantage
1982 Braniff goes bankrupt, the first case for a major airline since 1938
1984 Deregulation is considered to be successfully completed, and the CAB is
abolished
Most remaining functions are transferred to the DOT
Accusations of predatory pricing occur
Antitrust ruling enforces equal listing on Computer Reservation Systems
1986 People Express goes bankrupt mainly as a result of American’s revenue
management system
1986-89 Mega-Mergers era: Concentration increases and fortress hubs are formed
1992 American ignites a price war with its “value pricing” campaign and causes
many bankruptcies and the worst performance in the industry
1993 The peak year for Travel Agency share in bookings
1995 Explosion of the internet
140
ValueJet Crash causes misfortune for discounters
Majors employ aggressive/predatory pricing to contain/destroy LCCs
1997-98 Competition stagnation: No new entrants in any markets
1998 DOT proposes guidelines to prevent anticompetitive practices in the
industry
1999 DOJ reviews complaints and files a lawsuit against American
May 2001 American case dismissed by summary judgment. DOT appeals
September
2001
Terrorist attacks cause turmoil in the industry that was already sending
alarming signals
2002- Majors carriers drop commissions to travel agents
LCCs gain market share and force majors to become more efficient
5.1.1 The Propeller (Early) Era
The first flight was accomplished by Orville and Wilbur Wright on December 17,
1903 at Kitty Hawk, North Carolina. All lasted 12 seconds for 120 feet distance but it
was the first flight with the exception of balloons and gliders. The first U.S. passenger
flew with Orville Wright five years later in 1908 and the transportation industry would
never be the same.
The first scheduled air service started in 1914 between St. Petersburg and Tampa,
Florida. The seaplane had a capacity of one passenger and made two flights a day, and
the service cost $5 one-way. The company ceased operations after only four months.
Early flights were always great publicity but commercial aviation did not become
widespread until much later.
World War I necessitated more powerful engines and resulted in planes that could
fly at 130 mph. Increased power also enabled larger aircraft. However, the end of war left
a huge surplus of planes (i.e., no demand for new planes) and many aircraft builders went
out of business. The railroad was still the primary means of transportation for Americans.
Table 5.1: (continued)
141
The U.S. government decided to put the surplus planes to use. In 1917, The Congress
approved $100,000 for an experimental service and the first airmail was delivered from
New York to President Wilson in Washington on May 14, 1918. By 1920, the Post Office
was able to save almost a day on transcontinental deliveries through airmail.
The fleet, able to fly at night by mid-1920s, flew an average of 2.5 million miles
and delivered 14 million letters annually. However, the Post Office typically contracted
private companies for mail transportation, and had no intention of staying in the airmail
business. The Contract Air Mail Act (also referred to as the Kelly Act after its main
supporter, Rep. Clyde Kelly of Pennsylvania) was passed in 1925.
The Air Commerce Act of 1926 gave the Secretary of Commerce the authority to
designate air routes, and license pilots and aircraft. National Air Transport (owned by the
Curtiss Aeroplane Co.), Varney Air Lines, Western Air Express, Colonial Air Transport
and Robertson Aircraft Corporation were the initial contract winners. They would form
the core of the U.S. private airlines: “National and Varney would eventually become
important parts of United Airlines. Western would merge with Transcontinental Air
Transport (TAT), another Curtiss subsidiary, to form Transcontinental and Western Air
(TWA). Robertson would become part of the Universal Aviation Corporation, which in
turn would merge with Colonial, Southern Air Transport and others, to form American
Airways, predecessor of American Airlines. Juan Trippe, one of the original partners in
Colonial, later pioneered international air travel with Pan Am - a carrier he founded in
1927 to transport mail between Key West, Florida, and Havana, Cuba. Pitcairn Aviation,
yet another Curtiss subsidiary that got its start transporting mail, would become Eastern
Air Transport, predecessor of Eastern Air Lines.”
142
Henry Ford was also among the early contact winners. He carried mail from
Detroit to Chicago on the same planes that transported spare parts for auto
manufacturing. In 1927, Ford introduced the Trimotor (also referred to as the Tin Goose).
The “Tin Goose” was the first plane designed specifically to carry people and had a
capacity of 12 passengers. The Ford brand was assuring for the public but it took Charles
Lindbergh’s historic flight across the Atlantic to bring flying to unprecedented public
attention. Lindbergh became an instant hero when he completed his non-stop trip from
New York to Paris on May 21, 1927. The Air Age had commenced.
The 1930 Watres Act (after one of its main supporters, Rep. Laurance H. Watres
of Pennsylvania) authorized the Post Office to sign long-term airmail contracts based on
space or volume as opposed to weight, and to consolidate airmail routes where necessary.
The idea was to promote larger, hence stronger airlines enabling more frequent and faster
mail delivery. The Post Office held a number of meetings attended by select larger
airline executives (thus, these meetings were later referred to as the Spoils Conference).
The idea was to have one company operating on each of three transcontinental mail
routes as opposed to mail changing hands between several smaller airlines until it reached
its destination.
Following the victory of the Democrats in 1932, smaller airlines complained that
there had been unfair practices in the bidding process. In one case, it was discovered that
a big airline was favored over a smaller airline even though its bid was three times
higher. Congressional hearings followed, and by 1934 the scandal had reached such
proportions that President Roosevelt cancelled all mail contracts and gave the job to the
Army. Unfortunately, the Army pilots were not familiar with the routes and the
143
corresponding weather conditions. A number of accidents took place during practice runs
and President Roosevelt had to retreat from his plan just one month later. The result was
the Air Mail Act of 1934. The airmail services would be run by the private sector but the
former contractors were not allowed to bid, and the bidding process was re-structured to
be more competitive. The resulting lower margins mandated that the airlines pay more
attention to the passenger business. The government also put a halt to vertical integration
in the industry and manufacturers and operators were separated (e.g., Boeing, Pratt &
Whitney, and United Airlines) resulting in a more focused and re-organized industry.
Arguably, the 1930s were the most innovative period in aviation history. Safer,
larger, and faster planes were necessary to have a feasible airline business. Air-cooled
engines, better cockpit instruments (e.g., improved altimeters, airspeed indicators, rate-
of-climb indicators, compasses) and the introduction of artificial horizon were among the
innovations of the period. Naturally, the wide spread use of radio was also of prime
importance. Eighty-three radio beacons across the country were fully operational in 1932.
The first air traffic control tower was established at Newark, New Jersey in 1935.
Launched in 1933, the first modern passenger airliner is considered to be Boeing
247. The 247 had a capacity of 10 passengers and could fly 155 miles per hour. United
Air Lines purchased sixty 247s. TWA wanted to outdo United and the search for a better
alternative led them to the Douglas Aircraft Company. Douglas’ prototype DC-1
incorporated and improved upon many of Boeing’s innovations. The longer version the
DC-2 could accommodate 14 passengers and was a big hit. However, DC-3, later called
the plane to change the world, had a capacity of 21 passengers, and was considered cost-
efficient, safer, more comfortable, with more powerful engines. It could complete a coast-
144
to-coast trip in 16 hours. Following the marketing concept of product development,
Douglas got American Airlines heavily involved in the design process of DC-3. The
result was the first plane that enabled airlines the make money from passenger business.
DC-3 became very popular and introduced many new travelers to the joys of flying.
Despite the success of DC-3, a technical difficulty remained: The airlines wanted to fly
higher so that they could avoid air turbulence and storms at lower altitudes. Motion
sickness was also a problem for many passengers. However, they could not fly higher
than 10,000 feet because the lack of oxygen made passengers dizzy or even unconscious.
In 1940 the solution was called Stratoliner by Boeing, the first plane with pressurized
cabin. First embraced by TWA, the Stratoliner had a capacity of 33 passengers, could fly
at 20,000 feet at 200 miles per hour.
There were also political problems resulting in economic problems for the
airlines. Before the Civil Aeronautics Act of 1938, many government agencies were
involved with the airlines resulting in bureaucracy and no long-term policy for the
industry. All the airlines were losing money due to the reduced mail revenues since the
1934 Airmail reform. The wish of the airlines for a rational government regulation
through an independent agency was granted through the Civil Aeronautics Act (Rhyne
1939):
Sec. 2. In the exercise and performance of its powers and duties under this act, the
Authority shall consider the following, among other things, as being in the public interest,
and in accordance with the public convenience and necessity:
a. The encouragement and development of an air transportation system properly
adapted to the present and future needs of the foreign and domestic commerce of
the United States, of the Postal Service and of the national defense;
b. The regulation of air transportation in such manner as to recognize and
145
preserve the inherent advantages of, assure the highest degree of safety in and
foster sound economic conditions in, such transportation and improve the
relations between, and coordinate transportation by air carriers;
c. The promotion of adequate, economical and efficient service by air carriers at
reasonable charges, without unjust discrimination, undue preferences or
advantages, or unfair or destructive competitive practices;
d. Competition to the extent necessary to assure the sound development of an air
transportation system properly adapted to the needs of foreign and domestic
commerce of the United States, of the Postal Service, and of national defense;
e. The regulation of air commerce in such manner as the best promote its
development and safety; and
f. The encouragement and development of civil aeronautics.
The Civil Aeronautics Authority (CAA) was founded with the Civil Aeronautics Act.
CAA was empowered to regulate fares, mail rates, inter-line agreements, mergers, and
routes. CAA had a mission with a double edge. It needed to hold rates at reasonable
levels for the public but also strengthen the financially weak airline business to help
develop the commercial air transportation business. First, Air Safety Board was created to
investigate accidents. However, that function was also transferred to CAA in 1940. CAA
was then renamed Civil Aeronautics Board (CAB). The CAB was modeled after the
Interstate Commerce Commission so that public utility type of regulation could be
imposed and the airline industry would not be harmed because of “cut-throat”, wasteful,
destructive, excessive, unrestrained competition (Dempsey 1989, p.18).
There was one remaining factor before the commercial aviation could take off, the
World War II. Interestingly, just as aircraft helped efforts of war, warfare helped the
aircraft industry. In 1939, there were less than 300 air transport planes in the U.S., but by
146
1945, 50,000 planes were produced annually. While the U.S. focused on mass
production, the breakthrough innovations –radar and jet engines- took place in Europe.
5.1.2 The Jet (Post-war) Era
The first jet engine was designed by a British pilot in 1930; but the first to build
and test a jet plane were the Germans in 1939. However, it would take five more years
for them to perfect the design –too late to change the outcome of the war. Nevertheless,
the jet age had arrived. The U.S. built its first jet plane, the Bell P-59 in 1943. The
breakthroughs in military aircraft were eventually applied to commercial aircraft. For
example, Boeing remodeled and introduced its KC-135 jet tanker as the first U.S.
passenger jet, the 707 in 1958. The 707 had a capacity of 181 passengers and could fly
550 miles per hour. It burned kerosene, which cost half as much as the gasoline the
traditional planes were using at the time. Pan Am was the first customer of the legendary
Boeing 707. The same year the Federal Aviation Act was passed to monitor the growing
airline industry. With this Act, Federal Aviation Agency was founded to establish and run
the air traffic control system, and to monitor the safety of the overall flights (its name was
later changed to Federal Aviation Administration when the Department of Transportation
(DOT) was created in 1967). CAB still had authority over economic issues such as routes
and fares.
Boeing launched its first widebody jet the 747 in 1969. The 747 had a capacity of
450 passengers. Pan Am was the first customer. Douglas and Lockheed jumped on the
widebody jet bandwagon with the DC-10 and L1011 respectively. However, these planes
147
were smaller in size, seating about 250 passengers. The supersonic plane, the Concorde
flew the same year.
5.1.3 Deregulation Era
Deregulation will be the greatest thing to happen to the airlines since the
jet engine.
Richard Ferris, President, United Airlines (Peterson and Glab 1994, p.49)
Before deregulation the U.S. airline industry was run much like a public utility.
Civil Aeronautics Board (CAB) was in charge determining who would fly which routes
and how much they would charge for the service. The CAB would not permit most new
companies to fly. For example, it delayed its decision on a proposal to fly coast-to-coast
for less than half the going rate for eight years and then dismissed it (Peterson and Glab
1994). As a matter of fact, some of the practices in the regulated airline industry would be
considered illegal in other industries. For example, the CAB approved a war chest fund in
which the airlines agreed to support any airline that was suffering from a union strike. By
mid-70s the mutual aid had amounted to $350 million (Peterson and Glab 1994). In 1970,
American, TWA, and United jointly decided to cut the capacity on their coast-coast
services so that they could achieve higher load rates. There were no other competition on
coast-to-coast routes and this would constitute a deliberate antitrust case for any other
industry except the airlines. Instead, the CAB approved it and granted antitrust immunity
for one year (Peterson and Glab 1994, p.30). Similar agreements followed. Regulatory
structure enabled airlines to fly at half capacity to capture market share. Since carriers
148
could not compete on price, they were competing on non-price terms such as “sandwich
wars.” For example, Delta was alleging that Northeast airlines did not have their steaks
“cooked to order” as their advertisements stated (Peterson and Glab 1994, p.30). This
dispute continued until Delta’s acquisition of Northeast (Kuttner 1996). This kind of
subtle competition did not create winners but an industry with low profitability.
The notion of deregulation was becoming stronger as studies showed that
unregulated interstate fares in Texas and California were significantly lower than
intrastate flights after controlling for distance (Levine 1987). For example, Southwest
was able to avoid Federal regulations by servicing the Dallas-San Antonio-Houston
triangle, which enabled it to offer fares much cheaper than that of the incumbent airlines.
It was asserted that regulation gave consumers excessive service, but a lack of price
competition, and inflated fares (Dempsey 1989). Moreover, the increased use of the
widebody jets coincided with the OPEC oil embargo in 1973. The fuel prices and
inflation increased drastically. The result was increased capacity and cost but falling
airline traffic. In order to assure a reasonable rate of return for the airlines, the CAB
allowed airlines to increase fares. It also decided to not approve any new service on any
route for four years and limited the overall capacity on major routes. However, the poor
performance of the airline industry continued despite the fact that it cost more for the
public to fly. The industry became a target for the Ford Administration, which was after
regulatory reforms. Hearings of the Senate Subcommittee on Administrative Practice and
Procedure, chaired by Senator Edward Kennedy, concluded that airline prices would fall
automatically at the absence of government-imposed limits (1977). The major airlines
had not taken the hearings seriously and had sent their junior executives. On the other
149
hand, Kennedy had enlisted elite academics (e.g., Alfred Kahn of Cornell) and discounter
heroes such as Freddie Laker (of U.K.’s Laker Airways) (Peterson and Glab 1994).
Kennedy recognized the political importance of bringing the airfares down and it
economically seemed to make sense. Alfred Kahn had written two volumes entitled the
Economics of Regulation and had concluded that even imperfect competition was
preferable to inherently imperfect regulation (Kahn 1970). Deregulation would allow for
new and innovative services, increased productivity and efficiency resulting in higher
consumer welfare. The CAB admitted to similar conclusions (1975). The airline industry
“was naturally competitive, not monopolistic” therefore entry and price constraints could
no longer be justified. Led by John E. Robson, the CAB started to loosen its grip on its
own. Cornell University Economics Professor Alfred E. Kahn became the chairman of
the CAB in 1977. Kahn was not happy with the existing CAB regulation. He argued that
the system inflated fares, and caused misallocation of resources, carrier inefficiency,
excess capacity, and acute range of services and prices (Dempsey 1989, p.20).
Kahn actively participated in an effort for reform by effectively using the
preliminary results of flexible regulation under his leadership for convincing political
figures and media. Behind the scenes, United Airlines
6
, and Federal Express were also
supporting the idea of deregulation. Each trusted that the size of their fleet would give
them an edge over competition in an unregulated industry (Petzinger 1995).
The theory of contestable markets in which potential market entry would prevent
monopolies was going to be tested with deregulation. Kahn was confident that the results
6
Worldwide, only Aeroflot had a larger fleet than United Airlines at that time.
150
would benefit all constituents including consumers, all communities, and the airlines --its
employees, stockholders, and creditors (1978, p.8).
Kahn would be dubbed the father of the U.S. airline deregulation a year later in
1978. Deregulation first commenced on the cargo side of the business. The express
package delivery service experienced tremendous growth since deregulation. Fed-Ex,
UPS, DHL are all considered owners of major airlines today.
The Airline Deregulation Act:
The Airline Deregulation Act, a milestone for the airline industry, was approved
by Congress on October 24, 1978 and signed soon after by President Jimmy Carter.
Government Controls on domestic routes and schedules were discarded and the free
market economics took over. Congress had scheduled that route and rate regulations to be
phased out in four years. However, CAB moved very liberally with Kahn on board, and
all restrictions on routes were practically abolished within one year.
There was also much deregulation in the international arena. The U.S. had signed
45 “Open Skies” agreements by mid-2000. “Open Skies” agreements abolish limitations
on routes including those on capacity, frequency, and provide flexibility for pricing,
charters, cooperative marketing agreements, and other joint operations. The DOT
continues to monitor the remaining regulated international routes.
After deregulation was successfully completed, the CAB was abolished at the end
of 1984 and most of its remaining functions were transferred to the DOT. Among the
most important of these inherited functions was to review and grant antitrust immunity to
merger and acquisition activities. However, the role of the government on safety has not
151
been abandoned, and the FAA continues to regulate safety issues. Deregulation sparked a
wave of change in many aspects of the airline business including many that were not
forecasted. For example, the rise and fortification of the hub-and-spoke networks, more
complex fare structures, and the survival of major firms at the expense of entrants were
all unforeseen (2001g). These changes are discussed next.
5.1.4 Consequences
Transportation was the first industry in the nation to be regulated, and the first to
enjoy significant deregulation (Dempsey 1989). While the evaluation regarding the
success of deregulation has been mixed, the evidence has been encouraging. Yet the
experience was far from perfect, and there were criticisms:
Of the six intellectual assumptions behind the airline deregulation, four
have been proven completely false. Deregulators believed that airline size
was not critical to efficient operations. The marketplace, to the contrary,
has ruled that bigger is better. Deregulators believed that barriers to entry
are low in the airline business. Experience has demonstrated that they are
very high. Deregulators believed that increased competition would
produce low unrestricted fares. In fact, it has produced a bewildering
array of discriminatory prices. Deregulators believed that travel agencies
were obsolete as well as potentially misleading channels of information
and distribution. But travel agencies became more powerful than ever. A
fifth assumption, that antitrust laws would restrain competitive abuses,
has been negated by the policy default of two administrations…(Kuttner
1996).
Despite the above criticisms “Deregulation has been one of the most successful
regulatory and economic policy reforms in the late 20
th
century” (Kasper 1998). Kahn
himself admitted that he had surprises in the deregulation process including the
turbulence and painfulness of the process, the reconcentration of the industry, the
intensification of price discrimination and the deterioration in quality of airline service
152
(Kahn 1988). He asserted that deregulation resulted in lower fares, increased range of
price-quality options and improvements in efficiency but that the competition was
unevenly distributed across markets and that congestion and delays were causing
problems (Kahn 1988). Several consequences of deregulation are investigated next.
5.1.4.1 Service and Quality
Serious concern about the decreasing level of safety and service quality and
increasing concentration levels after deregulation has been voiced (Dempsey 1989):
The industry rapidly became an oligopoly, with an unprecedented wave of
mergers consolidations, and bankruptcies. Today, the top 8 airlines
dominate more than 94% of the domestic passenger market… [F]unneling
of aircraft into “hub-and-choke” bottlenecks… have significantly
narrowed the margin of safety and sent the number of near misses
skyrocketing. Airline service has gone to hell in the 1980s. We are headed
aboard aerial slums, served cardboard food, overbooked, bumped, and
misconnected. Our luggage is routed through the Twilight Zone, never to
be seen during our natural lives… We can either spend an arm and a leg
or sleep in a strange city on a Saturday night.
The number of departures have increased by 50% for small community, 57% for
medium, and 68% for large community airports since deregulation (1996a). The DOT
indicated that the number of domestic passengers tripled since deregulation and that more
than 80% of domestic passengers enjoy two or more carriers alternatives (2001g).
However, Dempsey (2000a) points put that 61% of the non-hub communities suffered
from decreases in service. 28% of them lost all the service they had and only 6% enjoyed
new services (Dempsey 2000a). Meanwhile customer satisfaction was also deteriorating
(Dempsey 1989). The following were reported to be the top customer complaints ranked
in order:
153
Flight Problems: Cancellations, delays or other deviations,
Baggage handling: Claims for lost, damaged, or delayed baggage; charges for
excess baggage; carry-on problems; difficulties with claim procedures,
Refunds: Problems in obtaining refunds for unused or lost tickets or fare
adjustments,
Customer Service: Rude or unhelpful employees, unpleasant meals or cabin
service, and treatment of delayed passengers
Reservations, ticketing and boarding: Airline or travel agent mistakes in
reservations and ticketing; problems in making reservations and obtaining tickets
due to busy phone lines or waiting in line; delays in mailing tickets; and problems
boarding the aircraft,
Oversales: Bumping problems, whether or not the airline complied with DOT
oversale regulations,
Other: Cargo problems, security, airport facilities, claims for bodily injury, and
other miscellaneous problems,
Fares: Incorrect or incomplete information about fares, discount fare conditions
and availability, overcharges, fare increases, and the level of fares in general,
Smoking: Inadequate segregation of smokers, failure of the airline to enforce no-
smoking rules,
Advertising: Ads that are unfair, misleading, or offensive.
Source: (Coleman 1987) in (Dempsey 1989)
One of the major concerns about deregulation was that smaller communities
would lose air service after deregulation. Despite the ongoing program, (Dempsey 1989)
notes that many smaller communities have either lost all air services or the services were
downgraded to commuter carriers and suffered from a sharp decrease in quality. Forty
percent of small communities have suffered from decreasing service and increasing
prices after deregulation (Moore 1986).
154
As a potential remedy, the DOT administers the Essential Air Service Program,
which was launched with deregulation. Essential Air Service Program was designed to
provide subsidies to carriers to fly to certain smaller communities to which service would
be not feasible otherwise. Section 419 of the Federal Aviation Act ensures that smaller
communities remain linked to the national aviation system. This program was initially
approved by the Congress for a period of ten years (expiring in 1988) but was later
extended for another ten years (until 1998). Seventy-eight communities were still being
subsidized under this program as of May 1998 (1998d). In 1998, the end date for the
program was also abolished and its annual budget was increased and set to $50 million
with Rural Air Service Survival Act. The program continues to this day. Some 113
communities were being subsidized as of October 2001 (2001b; 2001j). The carriers
under the Essential Air Service program are usually assigned for a period of two years.
The following is expected from the basic essential air service:
(a) service to a hub airport, defined as an FAA-designated medium- or large-hub
airport,
(b) service with no more than one intermediate stop to the hub,
(c) service with aircraft having at least 15 passenger seats at communities that
averaged more than 11 passenger enplanements a day in any calendar year from 1976-
1986,
(d) under certain circumstances, service with pressurized aircraft, and
(e) flights at reasonable times taking into account the needs of passengers with
connecting flights (1998d).
Currently, communities are eligible to be a part of this program if they are further
than 70 driving miles of an FAA-designated Large or Medium Hub airport, unless the
subsidy per passenger exceeds $200 (with certain exception for communities that are
further than 210 highway miles from the nearest Medium or Large Hub) (2001f).
155
In conclusion, it can be said that while service in hub markets increased in
frequency, some community markets suffered from a loss of service as a result of
deregulation. While the range of the quality of services (e.g., first class, business, coach)
increased, it can be argued that the average service level also decreased. However, this
issue should be taken into consideration along with the fare levels, which generally were
reduced due to heavy competition following deregulation.
5.1.4.2 Fare Levels
Numerous low cost carriers (LCCs) challenged the major carriers following
deregulation. Major carriers
7
responded by price cuts, and airfares (adjusted for inflation)
fell by 33% from 1976 (the dawn of deregulation) to 1993 (Morrison and Winston 1995).
Thus, benefits of new entrants included lower fares and increased frequency. Dropping
fares ignited demand: “For example, when AirTran entered the Atlanta-Buffalo market,
average fares declined by 36%, from $185 to $119, and the number of the passengers in
the market increased by 65%, from 23,000 per month to 38,000 per month. Similarly,
when Vanguard re-entered the Kansas City-Minneapolis market in late 1996, average
fares declined by 59%, from $246 to $101, and the traffic more than doubled, increasing
from about 12,000 passengers per month to 25,000 per month” (2001g, p.6). The DOT
estimated that consumers saved $6.3 billion per year due to low fare carriers (1996c).
“Most upstarts have average seat mile costs in the 7 cent range, compared to the 10 cent
7
A major airline is defined by the DOT as airlines with annual operating revenues of over $ 1,000,000,000.
(Oster and Strong 2001) There were 12 major U.S. passenger airlines in 2000: Alaska, America West,
American, American Eagle, American Trans Air, Continental, Delta, Northwest, Southwest, Trans World,
United and US Airways. In addition, three all-cargo airlines were classified as majors: DHL Airways,
FedEx and United Parcel Service.
156
range of traditional carriers. Upstarts maintain their low costs by a number of strategies,
including lower labor costs, direct marketing and “no frills” service” (Fones 1997). Thus,
discount airlines have approximately 30% cost advantage over network carriers.
However, the increase of concentration in hub markets limited the extent of
benefits that discount carriers could offer. The average number of carriers per route was
2.2 (most routes are served by major carriers only) and those routes that Southwest flies
were 47.2% cheaper than comparable routes (Kuttner 2000). DOT reported that
passengers paid $54 less, on average, if a low cost carrier served the same market
(Dempsey 2000a). On the other hand, while the public has been getting better deals, it
was reported that unrestricted fares (i.e., fares that mainly business class passengers pay)
have increased by 73% after deregulation (Dempsey 2000a, p.485).
157
Figure 5.1: Demonstration of fare level differences 1998Q2
Note: SIFL stands for Standard Industry Fare Level. It is the pre-deregulation fare level (adjusted
for inflation). Source: Domestic Airline Fares Consumer Report 1998 Q2 (1998a)
158
Figure 5.1 powerfully demonstrates the impact of low fares carriers and
deregulation has had on fare levels. The 100% base line represents the prices in 1978
(pre-deregulation) adjusted for inflation.
The implication is that the overall fare levels are indeed lower than they would
have been under regulated pricing for distance blocks longer than 750 miles. The prices
are lower than the regulation era for low-fare carrier markets for all distance blocks.
Finally, the major carriers’ fares are lower than regulation era only for distance blocks
longer than 1500 miles. The figure applies to the second quarter of 1998 but the
implications can be generalized to other time frames. For example, Figure 5.2 presents
the same notion for the third quarter of 2001.
Figure 5.2: Demonstration of fare level differences: Top 1000 markets, 2001 Q3
Note: SIFL stands Standard Industry Fare Level. It is the pre-deregulation fare level
(adjusted for inflation). Source: Domestic Airline Fares Consumer Report 2001 Q3
(2001d)
159
5.1.4.3 Market Concentration and Power
Despite the early success, ten years into deregulation, the increasing levels of
concentration were causing concern. The DOT observed that the number of carriers had
increased from 39 to 131 from 1978 to 1987 (1987a). However, this observation was
misleading since “nearly two-thirds of city-pairs were airline monopolies and another
20% were duopolies” (Dempsey 1989, p.87). “The 11 major airlines have shrunk to eight;
the eight local former local service carriers are now two and they are trying to merge; the
eight original low-cost charter airlines have been reduced to one, through bankruptcy and
abandonment; 14 former regional airlines have shrunk to only four; over 100 new upstart
airlines were certified by the CAB and about 32 got off the ground and most of these
crashed, leaving only a handful still operating; of the 50 top commuters in existence in
1978, 29 have disappeared… Today, the top 50 carriers who constitute 90 percent of that
industry are captives of the major airlines and relegated to serving the big airlines at their
hubs” (1987b, p.61-62).
A major reason for increased market concentration was the DOT’s relaxed merger
evaluations in the 80’s. Particularly in the second half of the 80s, the DOT approved a
number of mergers based on the notion that there were no barriers to entry even tough the
mergers resulted in dominant markets shares. The analysis overlooked the issue of
whether or not those potential entries would be economically feasible (Nannes 1999).
Despite the warnings in the Airline Deregulation Act of 1978 against “unreasonable
industry concentration, excessive market domination”, the DOT “never met a merger it
didn’t like” (Dempsey 1989, p.87):
DOT approved them all. It approved Texas Air’s (i.e. Continental and
New York Air) acquisition of both People Express (which included
160
Frontier) and Eastern Airlines (which included Braniff’s Latin American
routes); United acquisition of Pan Am’s transpacific routes; American’s
acquisition of AirCal; Delta’s acquisition of Western; Northwest’s
acquisition of Republic: TWA’s acquisition of Ozark; and US Air’s
acquisition of PSA and Piedmont, to mention only a few. This has sharply
increased national levels of concentration to the point that the eight
largest carriers control over 94% of the domestic passenger
market…Under deregulation, [charter flights] virtually vanished
(Dempsey 1989, p.88)
A critical review can indeed establish that the mergers in the Reagan era
started an irreversible process and were an important reason of the increase in
market concentration in the airline industry. Table 5.2 demonstrates that twelve
mergers took place in a matter of two years:
Table 5.2: Airline Mergers in the 1985-87 Period
Acquiring
Airline
Passengers
(thousands)
Acquired Airline Passengers
(thousands)
Final Bid
Southwest 10698 Muse 1980 March 11, 1985
Piedmont 14274 Empire 1084 October 3, 1985
People 9100 Frontier 7068 October 9, 1985
Northwest 14539 Republic 17465 January 24, 1986
Texas 19640 Eastern 41662 February 24, 1986
TWA 20876 Ozark 5541 February 28, 1986
Alaska 3132 Jet America 774 September 8, 1986
Delta 39804 Western 9062 September 10, 1986
Texas 19640 People 11907 September 16, 1986
American 41165 Air Cal 4451 November18, 1986
Alaska 3132 Horizon 942 November20, 1986
US Air 19278 Pacific Southwest 9049 December 9, 1986
Us Air 21725 Piedmont 22800 March, 1987
Braniff 2557 Florida Express 1415 December 15, 1987
Source: DOT Air Carrier Traffic Statistics (Oster and Strong 2001, p.7).
Other major mergers since deregulation have included (1993a, p.459):
American: TWA, Air Cal, Eastern (Latin America), Reno
United: Pan Am (transpacific), Pan Am (Latin America), Pan Am (Heathrow)
161
Delta: Pan Am (Europe)
Continental: Texas International, Frontier, New York Air, Rocky Mountain, Britt, PBA
Pan Am: Pan American World, National, Ransome
Eastern: Braniff (Latin America)
Republic: North Central, Southern, Hughes Airwest
US Airways: US Air (Allegheny), PSA, Empire, Henson
Southwest: Morris Air
TWA: TWA, Ozark
Former CAB chairman Alfred Kahn heavily criticized DOT for their hasty
approval of the mergers: “It is absurd to blame deregulation for this abysmal
dereliction.”(Kahn 1988). Empirical research has also indicated that the airline
consolidation has led to market power (Kim and Singal 1993). The ten largest airlines
accumulated 88% of the revenue passenger miles flown; that figure had risen to 94% by
1990 (Sheehan 1993). General Accounting Office reported that consumers flying from
small to major airports had to pay 34% more if the major airport was concentrated and
42% more if both airports were concentrated (Dempsey 2000a, p.485). However,
mergers were probably not the primary reason for the increased concentration in the 80’s.
Market concentration and the resulting power was a consequence of the rise of the hub-
and-spoke system.
5.1.4.4 Hub-and-Spoke Networks
I never heard a word spoken about hub and spoke in the entire debate
leading up to deregulation.
Congressman James Oberstar, Ranking Minority Member, U.S. House of
Representatives Aviation Subcommittee (Dempsey 2000a)
One of the unforeseen results of airline deregulation was the emergence of the
hub-and-spoke networks. The hub and spoke concept lies at the heart of any network
162
carrier. Securing hubs enables the dominant carrier to charge “hub premiums” (Hecker
2001). The following statistics should clarify their importance: U.S. General Accounting
Office (GAO) compared prices at concentrated hub airports and relatively unconcentrated
airports, and found that prices were 27% higher in the concentrated hubs (Dempsey
2000b). Adjusting for the average trip distance and the size of the markets, the
concentrated hub fares were on average 18.7% higher than for similar markets in other
airports. In the absence of LCC competition, the major carrier was able to charge fares
that exceeded its fares in non-hub markets of comparable distance and density by
upwards of 40% (Dempsey 2000b).
The incumbent hub carrier has an advantage over their competitors because they
are able offer a wider range of flights and services. In effect, they are able to attract a
larger portion of business travelers and obtain a higher yield than their rivals (2001g).
There is also higher brand recognition and the advertising costs are spread across more
markets than that of the rivals (Levine 1987). Robert Crandall, CEO of American airlines
summarized the benefits of the system as follows:
While a hub and-spoke system is admittedly more expensive to operate
than a comparably sized system of point-to-point routes, the system’s
incremental costs are more than offset by its enormous revenue benefits.
For example, we estimate that there are fewer than 500 city pair markets
in the United States big enough to adequately support point-to-point jet
service. However, our hub-and-spoke system makes it possible for
American to effectively serve over 10,000 markets –and realize a large
revenue per available seat mile premium relative to point-to-point
carriers. (1993a, p.3)
However, the efficiency of using hub and spoke networks for short-haul flights is
questionable. For example, Southwest has an average of 20.4 minutes ground time as
163
opposed to American’s 50.3 minutes (Dempsey 2000a). The result is a 22% better
utilization of aircraft for Southwest plus additional gains from personnel productivity
(Dempsey 2000a). The hub-and-spoke system also led to the use of smaller aircraft,
which meant relatively poor seat-mile cost efficiency. The same pattern of increasing
concentration of fortress hubs has been observed over and over again (Dempsey 1990;
1997). Table 5.3 demonstrates the increase of market share by hub carriers particularly
during the decade following deregulation. These figures continued to be alarmingly high
in the 90’s.
Table 5.3: Single Carrier Market Shares at Major Airports Pre and Post Deregulation
Airport 1977 1987 1997*
Baltimore/Washington 24.5% US Air 60.0% US Air 33.7% US Air
Cincinnati 35.0% Delta 67.6% Delta 79.4% Delta
Detroit Metropolitan 21.2% Delta 64.9% Northwest 63.4% NW
Houston Intercontinental 20.4% Continental 71.5% Continental 39.1% CO
Memphis 40.2% Delta 86.7% Northwest 52.2% NW
Minneapolis/St. Paul 45.9% Northwest 81.6% Northwest 69.7% NW
Nashville Metropolitan 28.2% American 60.2%American 29.3%SW
Pittsburgh 43.7% US Air 82.8% US Air 73.3% US
St. Louis – Lambert 39.1% TWA 82.3% TWA 48.7% TWA
Salt Lake City 39.6% Western 74.5% Delta 55.5% Delta
Average 33.8% 73.2% 54.4%
Source: Extended from Consumer Reports (June 1988), at 362-367 (Dempsey 1989, p.89)
* 1997 figures were retrieved from the DBP database and may not be fully compatible
with earlier figures. The loss of market share at the hubs were typically due to Southwest
competition.
Hubs enable the incumbent to have more frequent service on the spokes than
would otherwise be economically feasible (Levine 1987). Therefore, they are generally
beneficial for the public, however, their disadvantages include increased barriers to entry
(2001g; Anderson 1997). The resulting market power may also lead to supra-competitive
164
pricing (i.e., higher prices than would be considered normal for the market) and decrease
the welfare of consumers especially for those living in hub markets (Brown 1991). A
major concern regarding competition was that the major carriers stifled competition in
their hub airports. The incumbents carry the majority passengers in and out of their hub
markets (2001g). “The hub carrier dominates city pairs it serves directly from its hub,
except to other cities that are also hubs for other carriers, in which case the two carriers
providing hub service dominate. Entry by a major carrier on a point-to-point basis into
another carrier’s hub has become very much the exception” (Nannes 1999, p. 4). One
2001 DOT report concluded that remaining passengers pay 41% more than those flying
in hubs with low-cost carrier presence (2001e). The situation in short-haul hub markets
was even worse, with its 54% price premium over comparable routes with low fare
competition.
5.1.4.5 Hub Premiums
Federal Aviation Administration categorizes airports into four categories: large
hubs, medium hubs, small hubs and non-hubs (Morrison and Winston 1997):
Categorization Total U.S. Traffic Accounted by Hub
Large Hubs 1% or more
Medium Hubs 0.25% to 0.99%
Small Hubs 0.05% to 0.24%
Non-hubs less than 0.05%
When the fare levels are compared to non-hubs, fare premium effects are robust
and present. A hub premium is the increased fares that the consumers have to pay for
flying from a certain hub and is defined in percentages. Premiums have been observed to
165
be 62% for Charlotte, 51% for Cincinnati and Pittsburgh and 41% for Minneapolis/St.
Paul (Oster and Strong 1996). These premium effects were “persistent” over time
(2001g).
Interestingly, hub premiums are either mild or non-existing for hubs that are
served by Southwest (hence, the phrase “Southwest effect”) (2001g). For example, one
year after Southwest began serving Providence, RI markets, fares fell by almost 50% and
traffic more than tripled (Slater 2001). In summary, hub-and spoke networks enable the
use of market power and act as barriers to entry (Leigh 1990). Table 5.4 demonstrates
the consistent increase of hub premiums for the leading hub airports in the nation on an
airline basis. For example, it shows that consumers in Atlanta have had to pay 20-40%
higher fares for the privilege of flying out of Hartsfield airport in comparison to the rest
of the nation for the same distances.
Table 5.4: Changes in Hub Premiums over Time
Source: (Oster and Strong 2001, p.33)
166
The trends in Tables 5.4 and 5.5 indicate that the hub premium effect is persistent
over time. It should be noted that the occasional negative premiums may be either due to
presence of discounters or due to low entry barriers and limit pricing (e.g., St. Louis,
Baltimore).
Examining the HHI index at the national level can be misleading. A more
accurate assessment would be possible through an examination at the hub airport level. It
was found that the concentration for the airline industry had increased to an alarming
3877 (above 1800 is considered highly concentrated by the DOJ) and thirty-three airports
were assessed to be highly concentrated (1997a). It should be noted that only La Guardia
and Los Angeles International airports are below the 1800 threshold for high
concentration. There has been a general trend for more concentration fueled by the
mergers as explained previously.
Table 5.5: Sample HHI Index for the 10 largest US Airports 1985-1996
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
Chicago O’Hare 2449 2836 3263 3491 3526 3591 3617 3427 3557 3474 3505 3445
Atlanta 4190 4468 4569 4619 5420 5489 7586 7604 7058 6223 5585 6233
Dallas/Fort Worth 4143 4500 4519 4759 4906 4927 4833 5008 4541 4451 4581 4684
Detroit 2189 2151 3903 3793 4655 5124 5557 5973 6188 5949 6240 6465
Los Angeles 852 911 1025 1066 1131 1156 1297 1376 1393 1379 1528 1524
Denver 2356 3015 3767 3674 3504 3562 3587 3661 3787 4307 4980 4892
Phoenix 1601 2000 2395 2459 2507 2700 2849 2679 2645 2556 2544 2451
San Francisco 1437 1723 1847 1715 1903 2134 2552 2685 3073 3448 3760 3831
Newark 2821 2395 2154 2400 2815 2933 3092 3143 3222 3287 2929 3060
NewYork
LaGuardia
1368 1384 1320 1175 990 1102 1293 1662 1763 1733 1688 1777
Source: DOT Form 11 Data (1997a)
Professor Allvine observed that the major airlines widely employ predatory
pricing in the name of defending their “fortress hubs”:
167
Government study after study shows that airline industry is not perfectly
competitive… These studies show that the major airlines employ many
monopolistic practices that contribute to the market power to raise and
maintain prices above the competitive level. In imperfect competitive
markets, it makes perfectly good sense for large firms to use predatory
pricing to destroy competition that threatens the monopoly prices charged
(cf. Dempsey 2000a, p.474).
5.1.5 Marketing
Deregulation has resulted in stiff competition for the airlines and marketing
efforts became key as airlines scrambled to differentiate themselves from competition.
The airlines spent in excess of $10 billion for domestic sales and promotion activities in
2000 which was almost twice their operating profits (2001c). The following section
investigates the impact deregulation had on important elements of airline marketing.
5.1.5.1 Yield Management
“It was a typical flight in the era of deregulation: United Airlines flight
815 from Chicago to Los Angeles, with 204 tickets sold at almost as many
prices.”
(the range of prices on that flight was from $87.21 to $1258.51)
NY Times Reporter Matthew Wald (Wald 1998)
Bob Crandall of American Airlines allegedly is the innovator of the airline yield
(a.k.a. revenue) management. Based on the concept that the cost of additional seats on
scheduled flight was negligible, he ordered his staff to study the demand and price
elasticity patterns from its computer reservations system “Sabre”. The result came in the
form of 35% “Super-Saver” discounts in 1977 all around the nation. Unable to compete
without a price advantage many charter services were soon bankrupt (Petzinger 1995).
168
American training manual stated the new objective “to sell the highest priced product that
the customer is willing to buy” (Peterson and Glab 1994, p.59). The minds behind
regulation and economists had anticipated that deregulation would lead to a simpler fare
structure (2001g). However, the yield management systems led the industry in the
opposite direction.
Yield management essentially uses the power of price to communicate with the
customer and to lure them from one service (own or competitors’) to another. The
demand for different products/services is monitored continuously and price and
promotional adjustments are made to maximize revenues, thus the profits of a company
through segmentation of the market. Many companies in other industries have also
realized the benefits of revenue management. Examples include the hotel industry,
railroads, telecommunications, and broadcasting. Certain conditions need to be satisfied
for revenue management to be useful (Daudel and Vialle 1994). The product should be
perishable, it should be possible to price target different customers. It should be possible
to sell the product in advance. The variable costs should be low. The demand should be
cyclical or it should vary so that it can be smoothed by revenue management. Typically,
segmentation by time is key. Needless to say, the airline industry is ideal for the
application of revenue management. Today, virtually every successful airline is using
yield management in some form.
5.1.5.2 Frequent Flyer Programs
Among the innovations that deregulation inspired, a most important marketing
tool became the frequent flier programs. In its purest form, the frequent flier programs
169
attempt to create brand loyalty by inducing a passenger to fly on only the owner of the
program or one of its affiliates (through code-sharing). As with yield management, the
Frequent Flier Miles concept was also innovated by American Airlines (Peterson and
Glab 1994, p.60) in May 1, 1981. Five days later United Airlines responded by
introducing Mileage Plus (Woodyard 2001). All major airlines developed their own
programs shortly thereafter.
Table 5.6 demonstrates the scope of the frequent flier programs in the U.S.. For
example, American’s AAdvantage program has induced some 35 million members
resulting in 2.3 million travel awards redeemed during 1998.
Table 5.6: A Comparison of Major Frequent Flier Programs
Airline Members
(million)
Number of
Awards
Accumulated
Number of travel
awards redeemed in
1998
American Aadvantage 35 NA 2.3 million
US Airways Dividend Miles 20 4.4 million 900K
Continental OnePass 16 NA 1 million
Northwest WorldPerks 18.5 6.1 million 1.2 million
United MileagePlus 27 6.1 million 2.1 million
America West FlightFund 2.9 NA NA
Alaska Mileage Plan 3 812K 191K
TWA Aviators 12 1.1 million NA
Delta Skymiles 24 9.6 million 1.9 million
Southwest Rapid Rewards 1.2 NA 927K
Source: InsideFlyer (Stoller 1999)
Frequent flier programs have advanced into more sophisticated forms and
consumers can earn mileage points through several means including credit card purchases
or long-distance phone calls. As perfectly acceptable as they are as a marketing tool, a
potential anticompetitive problem with the frequent flier programs is that they can be
effectively used to target new entrants. Additional bonus miles on certain routes can
170
effectively bring down the fare without being detected and then be revoked (Oster and
Strong 2001). It has been argued that having a reward program tied in with an excellent
customer service will help take the customer’s eye off the price (Mohs 1999).
A current snapshot of the frequent flier campaigns is presented below
8
:
• There are more than 120 million members worldwide, 74 million members of whom
are from the U.S.
• 27-28% of the members are active.
• AAdvantage of American Airlines remains as the largest frequent flier program with
more than 45 million members. On average, more than 11 thousand new members
enrolled in AAdvantage per day in 2001.
• The programs grew by 11% on average with the fastest growing segment being
“mileage consumers.”
• Approximately 40% of the miles earned are not from flying and credit cards are the
most popular form of earning such miles.
• An award is estimated to cost airlines $13.93 on average.
• 14 million free tickets were awarded in 2001.
• 82-87% of members have web access.
5.1.5.3 Distribution: The Rise (and Fall) of Travel Agents
There is no doubt that the travel agencies grew in importance after deregulation.
As the choice of flights available to the public increased in number, the influence that the
agents had in the decision making also increased. The airlines tried to get the leading
agents on their side by use of commissions. Two main conflicting forces are at work
when the motives of travel agents are considered. On one side is the issue of the override
commission and the percentage baseline commission, which motivates the agent to
influence the customer to buy the most expensive ticket from the airline that offers the
8
Source:http://www.webflyer.com/company/press_room/facts_and_stats/frequent_flyer_facts.php
171
best override commission, on the other hand is the need to keep their customers satisfied
and happy to succeed in the long run. Travel Agent Commission Overrides (TACOs) are
special bonus commissions paid to travel agents by a specific airline for meeting a
targeted proportion or number of passengers booked (Oster and Strong 2001). It appears
that this scenario was more beneficial to major airlines who offered better commission
rates than low-cost carriers. Another public policy concern is that travel commission
overrides were typically designed to be in favor of the carrier with the largest market
share (1996a).
The inability of the new entrants to cope with commission overrides (TACOs)
was a main reason why they exited certain markets. Southwest airlines’ decision to pull
out of Indianapolis – Detroit (one of the rare exit events for Southwest) and Midwest
Express’ exit from Milwaukee-Detroit and other markets have been linked to commission
overrides (Oster and Strong 2001). The largest travel agencies also admitted that
overrides had an important effect on the booking patterns (1996a). The sophistication in
the computer reservation systems also fueled the need for agents.
However, the growth of discounting and the explosion of the internet as a viable
distribution medium drastically changed the equilibrium as the major airlines were being
forced to cut their commissions due to increasing cost pressures. Namely, Delta,
American and Continental decided to drop the commissions altogether and others
followed suit (Brannigan and Stringer 2002). Some major airlines have announced that
they will start for charging extra for paper tickets. Major airlines also realize that selling a
ticket through their own web site costs them 25% of what it cost through a travel agent,
172
therefore the travel agencies who typically sold 80% of the tickets face a gloomy future
(Fonti 1999). This trend can be readily observed in Figure 5.3.
Figure 5.3: US Travel Agency Overall Airline Commission Rates and the Number of
Agency Offices, 1976-2000
Source: Harris / Travel Weekly, 2000 Travel Agent Survey
5.1.5.3.1 Computer Reservation Systems (CRS)
The CRS systems had tremendous impact on the development of complex pricing
strategies and revenue management systems. These systems enabled the travel agents to
track fare and capacity changes instantaneously. Although several competing systems
were initially introduced (American’s Sabre vs. United’s Apollo), many of these systems
(as well as the largest travel agencies) later merged. Some airlines chose to join existing
systems and pay fees for listings rather than develop their own systems. Table 5.7
presents the travel agents’ market share of competing systems:
173
Table 5.7: 1999 Market Share for CRS
CRS Market
Share
Sabre 34%
WorldSpan 24%
Apollo 24%
System One 20%
Note: Some travel agents use more than one system
Source: Harris / Travel Weekly 2000 Travel Agent Survey
The DOJ actually sued the airlines and owners of the systems because of
signaling and price fixing (1994b). That the systems were designed to give their owners
certain (anti-)competitive advantages was a concern:
An airline whose CRS is used by travel agents has access to a very
accurate picture of both its own and its rivals’ business patterns. Through
the CRS an airline can track the effect of price changes, see roughly how
much of a rival’s seat inventory is assigned to a given discount fare
classification, measure how much full-fare business it attracts compared
to rivals, and track changes in city-pair flows… It can even see how loyal
its own frequent flyers are. A CRS owner can then use this information to
distort market signals to its rivals, leading them to make incorrect
decisions. When a CRS owner sees travel agents making bookings on a
rival airline’s flights, it can intervene through targeted incentive programs
in an attempt to switch business. By responding selectively, it can
temporarily distort signals the market sends to competitors, in order to
persuade the rivals to abandon fares, schedules, or even routes where,
absent these secret interventions, its offerings would be preferred by
customers (Kuttner 1996, p.261).
Computer Reservation Systems (CRS) provide much of the information needed
for a competitive response. Competitors, prices, capacity and availability can be observed
through these systems. The incumbents can even gather the scope of new entry because
the schedule and fares are filed before they are put in effect (Oster and Strong 2001).
Therefore, the incumbent has much of the knowledge it needs for a competitive response
174
decision. It can target its direct (e.g., price, capacity on route) and indirect predatory
weapons (e.g., frequent flier miles, commission overrides) discussed earlier to coordinate
a successful campaign against the entrant (Oster and Strong 2001). Since these efforts can
be highly targeted, they do no not signal hostility to provoke retaliatory response from
other major carriers.
5.1.5.4 Discounting and Competition
“Today, one of every seven domestic passengers is flying because of the
increased competitiveness resulting from low fare service.” (1996c)
In April 1996, the DOT released a report full of hope for the airline industry. Its
title read “The Low Cost Airline Service Revolution.”(1996c). One conclusion of the
report was that the consumers heavily benefited from the efficiency and competition that
the low cost carriers (LCCs) were bringing in the airline industry. It was also suggested
that there was evidence that network (major) carriers and low cost carriers could co-exist.
The fundamental cost advantages that low cost carriers have over network carriers were
emphasized and the global implications were discussed. Other positive implications
included those on industry labor force (e.g., union relations) and economic growth and
benefits to consumers, communities, travel related industries and the aerospace industry
in general. The report also included a warning that the premiums at network hubs where
there was no low-cost competition were high and increasing.
It can be argued that the discounter LCCs entered the market in two waves. The
first wave emerged right after deregulation in 1978 and continued roughly until 1982.
Even though these discounters helped reshape the industry and brought many benefits to
175
consumers, the pre-deregulation incumbents did not let the LCCs steal away their
customers. They responded with price discrimination utilizing sophisticated revenue
management techniques. Some of their deep-discount fares were coded “FU” and the
signal to the new entrants could not get clearer than that! (Kuttner 1996). As a result
many LCCs went out of business before they got a fair chance to compete.
The classic example of a discounter during the first wave was undoubtedly People
Express. People Express was founded in April 1981. Don Burr’s strategy was simple:
very low operating costs and offering flight experience to the masses (Peterson and Glab
1994). The service might not be extensive but it would be warm (Petzinger 1996). Every
employee was required to have a second job and many employees (e.g., the CFO) served
as a flight attendant in their second jobs. Don Burr was a charismatic leader and the
employees embraced his style of his leadership. He announced the precepts –the code of
behavior for every employee of People Express:
One: Service –Commitment to the growth and development of our people.
Two: To be the best provider of air transportation.
Three: The provide the highest quality of leadership.
Four: To serve as a role model for others.
Five: Simplicity.
Six: Maximization of profits. (Petzinger 1996, p.134)
As the most successful airline launch ever, the company grew from 3 to 17 planes
and was flying to 17 destinations out of Newark by the end of the same year. In five
years, it had 217 planes operating in domestic and international routes (Pearson 1996).
Their target market was the middle-class, working class, and students. “At one point it
was cheaper to fly to Florida than to take a bus or drive” (1994a, p.35; Pearson 1996).
The prices were fixed and you bought tickets as you boarded the plane, on a first come,
176
first serve basis. However, unable to respond to fierce price competition and revenue
management by American, it was already bankrupt in 1986. In fact, the same fate was
typical of the LCC of the era, among the 58 start-ups launched between 1978 and 1990,
only America West and Midwest Express have survived (McBride 1999).
The second wave of LCC entries happened after the industry once again became
profitable in 1993 and continued until the ValueJet crash in 1996. Ironically, the typical
example of a second wave entry would be ValuJet which started its operations in Atlanta
in 1993 and had grown from 5 to 30 cities until the crash. Second wave entries were also
effectively stopped by the major players because they were already fortified in their hubs
and were not about to give up the premiums. The typical competition pattern, also called
the “homicidal cycle” (Dempsey 2000a; Dempsey 2000b) was observed in the airline
industry in these five steps:
1. Incumbent (major carrier) monopolizes a hub and raises prices to supra-
competitive levels.
2. Low Cost Carrier (discounter) enters the market with low fares.
3. Incumbent responds by matching fares regardless of its cost structure, and often
couples the response with increased capacity and/or frequency. Incumbent also
uses TACOs (commission overrides) and other tactics to drive the entrant out
before it can establish a foothold in the market.
4. LCC, not having the deep pockets as the incumbent is forced to withdraw.
5. Incumbent increases price to previous or higher levels and decreases capacity.
As the same pattern was observed over and over again, the reputation effects
become more than a theory and new entry to markets were stifled. New market
entries diminished in the nineties (Dresner et al. 2001) and there were no new entry
applications to DOT during 1997 and 1998 (see Figures 5.4 and 5.5 below). However,
the tremendous effect discounting has had on fares levels is probably most apparent
from reviewing actual figures.
177
0
1
2
3
4
5
6
1992 1993 1994 1995 1996 1997 1998
New Entry Applications
Figure 5.4: New Entry Applications
Data Source: (1999b)
Market Entry Patterns in the Airline Industry
0
1
2
3
4
5
6
7
8
9
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
Figure 5.5: Entry/Exit Patterns in the Airline Industry
Note: The bottom (blue) stack represents new entering and remaining airlines for
the year, the top (red) stack represents the entry and exits in the same year.
178
Table 5.8: Average Fares in Markets With and Without Low-Fare Competition
Market Type 1994 1995 1996 1997
Without Low-Fare Competition $168 $183 $177 $180
With Low-Fare Competition $86 $95 $91 $100
% Difference 97% 92% 93% 80%
Source: Domestic Airline Fares Consumer Report 1997 (Q4)
Low-fare carriers usually do not fly longer distances so it may be argued that a
comparison between long haul and short haul market fares may not be justified.
However, a deeper look in any given year indicates that the benefits of low-fare
competition exist over the range of flights:
Table 5.9: Average Fares Sorted by Distance
Mileage Block Market Type
0-250 251-
500
501-
750
751-
1000
1001-
1500
1501-
2000
Over
2000
Without Low-Fare
Competition
$142 $165 $171 $167 $179 $198 $223
With Low-Fare
Competition
$67 $75 $101 $121 $133 $163 $176
% Difference 111% 121% 70% 38% 35% 22% 27%
Source: Domestic Airline Fares Consumer Report 1997 (Q4) (1997c)
The above Tables 5.8 and 5.9 demonstrate that discounting has a robust negative
impact on the prices. Focus on specific markets is also telling in that it reveals the
competitive impact on prices and passenger levels.
179
Case in point: Air-Tran Entry and Exits
Table 5.10: Atlanta Markets with AirTran Entry
Atlanta to:
Passengers per
day
Average Fare
97 Q3 98Q3
Change in
Passengers
97 Q3 98 Q3
Change
in
Average
Fare
Buffalo, NY 269 443 65% $169 $119 -30%
Dayton, OH 261 521 100% $178 $116 -35%
Greensboro,
NC
310 575 85% $229 $113 -51%
Hartford, CT 514 885 72% $242 $124 -49%
Houston, TX 974 1409 45% $211 $130 -38%
Richmond, VA 426 700 64% $223 $118 -47%
Group
Average
459 756 65% $213 $122 -43%
Table 5.11: Atlanta Markets with AirTran Exit
Passengers per
day
Average Fare
Atlanta to:
97 Q3 98Q3
Change in
Passengers
97 Q3 98 Q3
Change
in
Average
Fare
Charlotte, NC 875 623 -29% $110 $181 65%
Columbus, OH 728 546 -25% $114 $182 60%
Louisville, KY 589 391 -34% $101 $188 86%
Group
Average
730 520 -29% $108 $184 70%
Source: Domestic Airline Fares Consumer Report 1998 (Fourth Quarter); (2001e)
LCCs grew rapidly over time in spite of anti-competitive practices in the industry.
Their market share grew to almost 15% in 1999 from 8.8% in 1986 (2001a). This growth
is also evident in Table 5.12 below:
180
Table 5.12: Airports with Most U.S. Destinations Served by Low Fare Airlines
Low Fare Services 1992 2002
Phoenix 48 84
Las Vegas 33 59
Chicago Midway 9 54
Orlando 5 40
Atlanta 2 39
Baltimore 3 38
Kansas City, Mo. 6 35
Denver 2 33
Tampa 2 32
Nashville 4 28
Houston 14 25
Los Angeles 6 25
Fort Lauderdale 1 23
Albuquerque 11 22
Oakland 7 22
St. Louis 13 20
New Orleans 3 19
New York JFK 3 19
Austin 7 17
Columbus, Ohio 11 17
Seattle/Tacoma 3 17
Indianapolis 6 16
Salt Lake City 2 15
San Diego 6 15
Detroit 2 14
Portland, Oregon 3 14
Dallas Love Field 13 13
El Paso 12 13
San Antonio 6 13
Birmingham 2 12
Dallas/Fort Worth 2 12
San Jose 1 12
Source: (Adams 2002)
181
5.1.5.5 Price Wars
“I knew the industry would be ruggedly competitive, but I did not expect
its leaders to engage in extended kamikaze behavior.”
Warren Buffet regretting his investment in USAir in 1989 to his
shareholders (Labich 1993).
The competition between the major airlines seems to have stagnated during the
last ten years or so. However, this was not always the case, and irrational price wars have
also taken place in the colorful history of the airlines. Dempsey (2000b) describes one
such war between American and Northwest. In April 1992, American Airlines launched
its “value pricing” campaign. They had concluded that the price structures had become
too complex for the consumer to comprehend. Their new value pricing had only four
levels (first class, regular coach, 14-day advance purchase, 21-day advance purchase).
The unrestricted coach fares went down by almost 40% on average. The Economist
argued that American was pricing below operating costs and could cause more chapter 11
files in the industry (1992a). American anticipated that the newly stimulated demand
would more than offset the reduction in prices. However in May 1992, Northwest
announced its “Grown-Ups Fly Free” campaign (i.e., an adult flying with a fare paying
child would fly for free) to steal the thunder from American. American responded by
cutting its lowest fares by 50%. As a result, the prices went down and the industry
experienced the worst times in its history. Northwest and Continental sued American
airlines with claims of predatory pricing, but as it has been the case in almost any
predatory pricing case since Areeda-Turner, they could not get a verdict against the
182
defendant. However, American had to spend in excess of $20 million to cover legal costs
(Clouatre 1995). Eventually, complex fare structures resurfaced. Accusing American
Airlines with predatory pricing, Northwest CEO John Dasburg then stated:
In the long run, predatory pricing will reduce the number of airlines,
ultimately cutting the number of flights and choices available, particularly
in smaller markets. This will leave the few surviving airlines free to price
just as high as they want for just as long as they want.
(Dempsey 2000b, p.3).
A Relatively Recent Incident:
American seems to take the lead in most price changes but sometimes with
differing results. In March 2002, American eliminated a three-day advance fare
(Woodyard and DeLollis 2002). The impact of this would be to force the customers who
used to buy this class of tickets (i.e., mostly business travelers) into buying a higher class
fare (i.e., 10% fare increase). However, other major carriers, which usually readily went
along with price increases, did not think that the increase was justified this time. They
already had low load factors, and with the exception of Continental, they did not comply
with American’s price increase. Surprisingly, rather than taking the increase back to its
previous level, American “retaliated” to the refusal to follow suit by dropping its three-
day-advance fares by as much as $99 on routes where it competed with the rebel carriers
and required no Saturday-night stay (Woodyard and DeLollis 2002). It offered these
discounted fares in ten markets that it competed with United, Delta, and US Airways each
(McCartney 2002), but not with Continental. Normally, this type of big brother
punishment would be expected if a member breaks the rules of a cartel such as OPEC.
This was a case of blatant signaling and retaliation in the face of refusal to comply. This
183
incident demonstrates the unhealthy nature of competition in the airline industry.
However, this time the end result was beneficial to consumers. Northwest would not take
any of American’s swashbuckling and reacted by offering round trips for $198 in 10
American markets, and when American did not give in, it offered the same discounts in
20 markets. American insisted on its discounts so Northwest finally offered $189 round
trips in some 160 American markets (Woodyard and DeLollis 2002) while American’s
comparable prices were averaging $1650 (McCartney 2002). Southwest, always in search
of an opportunity, joined the battle by re-offering its “friends fly for free” campaign after
five years (Woodyard and DeLollis 2002).
Southwest Effect: Southwest undoubtedly proved to be the most successful LCC after
deregulation. It actually had begun its operation as an intra-state carrier before
deregulation but quickly extended its operations to more states. Southwest does not use
major hubs and relies on secondary airport for its operations. It also typically does not
serve long distance markets and focuses on short and medium haul instead. Most entries
by Southwest are to markets with connecting service by incumbents. Southwest enters
these markets with non-stop service and lower prices. The result is an overall decrease in
the fare levels in the market, also called the “Southwest effect.” (1998a).When the price
wars got tough in mid 90’s because of competing discount services such as Calite
(Continental), Shuttle (United), and Reno Air, Southwest’s Pilots Union accepted no
increases for five years in return for options, which in effect enabled the airline to cut its
costs by hundreds of millions of dollars (Banks 1995). At the height of the battle, United
Shuttle even purchased 1-800-Southwest and used it as its reservations number.
184
The major carriers were not successful in their campaigns against Southwest
because Southwest had both lower costs and deep pockets. “In the 1998 calendar year,
the total domestic operating cost in cents per available seat-mile for the network airlines,
adjusted for distance, ranged from 7.737 cents for America West and 9.123 cents for
Delta to 11.582 cents for US Airways. The comparable costs for the low-fare airlines
ranged from 6.083 cents for Southwest to 8.626 cents for Frontier. Thus every low-fare
airline had adjusted costs per available seat-mile that were significantly below the costs
of any network airline except America West.” (2001g). Therefore, the major carriers
admitted failure against Southwest’s strategy and attempted to contain it rather than
defeat it. “When major network airlines were subject to entry by either Southwest or by
another major network airline, the response was typically either a very slight fare
reduction with no significant increase in capacity or a fare increase. We did not find cases
where the response was as aggressive as when a new-entrant low fare carrier entered a
market.” (Oster and Strong 2001, p.15).
A Word on a Market Niche: It should be noted that not every new entrant in the airline
industry has attempted to be a discounter. Despite the tough competition with the
incumbents, several start-ups have attempted to carve themselves a niche in the first class
business travelers’ market. For example, Midwest Express has secured itself a niche in
the tough industry. Its formula is offering superior service at slightly higher fares. The
seats are leather, and there is no center row. As a result, there are 30% fewer seats on its
DC-9s than expected (Oliver 1995). Meals come in china with silverware and “crystal
salt and pepper shakers” (though how this was impacted after September 11 remains a
185
question) (Oliver 1995). Midwest spends twice as much on its meals than other airlines
(Oliver 1995). Midwest had been owned by Kimberly-Clark Corporation but went public
in September 1995 (Oliver 1995). Recently, JetBlue, National, and Legend were among
the other names who attempted to capture the same niche in different markets. So far,
JetBlue seems to be successful, and Legend is already out of business due to aggressive
responses by American.
5.2 Competitive Outlook
“Capacity is how we compete in this business”
Bob Crandall CEO, American (Labich 1993)
The revenue of the U.S. airline industry was around $97 billion in 2001. The
profit margins have not been impressive. Revenue management is used to subsidize
discounted tickets with the gains from first and business class fares. When the demand
falls for cyclical or other reasons, the industry experiences red ink. The following section
exposes the industry with some factual information.
5.2.1 Industry Snapshot
On the day the Wright Brothers flew at Kitty Hawk if there had been a capitalist
down there, the guy should have shot down Wilbur.
Warren Buffet. (Smith 1995)
Due to increased competition from LCCs and price wars, the profitability of the
airline industry has displayed a highly cyclical pattern over the years. Table 5.13
illustrates this point. It should be noted that the price wars in 1982 and 1992 happened
while the industry was already in a recession as the airlines tried to increase their load
186
factors.
Table 5.13: The Cyclical Nature of the U.S. Airline Industry
Cycle Years Duration Airlines
Boom 1960-68 9 years Expansion
Bust 1969-74 6 years Recession
Boom 1975-79 5 years Expansion
Bust 1980-82 3 years Recession
Boom 1983-89 7 years Expansion
Bust 1990-94 5 years Recession
Boom 1995-2000 6 years Expansion
Bust 2001-2005 5 years Recession
Note: 1995-2000 expansion was a forecast which eventually proved to be accurate. 2000-
2005 recession was predicted by AirTran CEO Joe Leonard at his speech at Georgia
Pacific Headquarters on Aug 19, 2002.
Source: Aviation Week and Space Technology, 13 March 1995
One recent trend has been international growth through global alliances. For
example, American Airlines and British Airways; United Airlines and Lufthansa; Delta
Air Lines, Air France and Aero Mexico all joined forces for international expansion. The
industry appears to be leading to a new wave of consolidation. However, the merger
between United and U.S. Airways was blocked by the DOJ. Alternative expansion
strategy for the majors has been controlling some or all of regional start-ups. For
example, Delta owns Delta Express, Atlantic Southeast, Comair and has launched Song.
Another interesting response by major airlines to low cost carriers was in the form
of founding their own low-cost versions. United Airlines was the first airline to try this
strategy in October 1994, and the service was called “The Shuttle” by United (Oster and
Strong 2001).
9
Delta started Delta Express and US Airways started Metrojet in October
9
Continental had had an earlier effort with Continental Lite but the initiative lacked a consistent strategy
187
1996 and June 1998 respectively (2001g; Oster and Strong 2001). These efforts were
generally defensive in nature but also allowed the majors to use the low-cost model of the
entrants through their flexible labor agreements. Pilot unions ultimately restrained the
expansion of most of these initiatives by imposing limits on total hours that could be
flown as a proportion of total flights.
Table 5.14 illustrates the aggregate income statement of the industry over time.
The expansion and recession periods of the industry can be tracked within this Table as
well.
and eventually failed (Oster and Strong 2001).
Table 5.14: Aggregate Industry Snapshot Over time
Source: Air Transport Association Annual Report (2001c)
1
8
8
The table illustrates that the profit margins of the industry have not been
impressive even in good years. Airlines have earned a net profit between one and two
percent on average, compared to an average of five percent or more for other industries.
Only 15 have survived among the 43 pre-deregulation airlines, and two thirds of 226
post-deregulation airlines have not been able to succeed (Dempsey 2000a). On the
positive side, the airlines were highly profitable in the late 90’s and their net profits
totaled $23 billion from 1995 through 2000 Q3. The Average passenger load was around
72% percent for the same year while the break-even load for most carriers was around
65% (2001g).
More than 130 airlines have gone bankrupt since deregulation, and currently there
is little incentive for new entrants (Wysocki Jr 2001). Airlines already schedule more
flights than the eight most crowded airports in the nation can handle, which only adds to
the congestion problems at the hubs (Wald 2001). The recent activities (e.g., reports,
DOT proposed guidelines, complaints and finally the DOJ lawsuit) in the airline industry
strongly suggest that anticompetitive practices (e.g., predatory pricing) may indeed be an
issue.
The unique nature of the airline industry enables the network carriers to engage in
predation in a variety of ways. Observed tactics have included: “dropping prices sharply;
eliminating advance purchase and Saturday night stay-over restrictions; expanding the
inventory of low-fare seats offered; increasing the number of flights and/or the size of
aircraft; scheduling departures in close proximity to the new entrant’s flights, sometimes
boxing them in; offering passengers bonus frequent flyer miles; paying travel agent
commission overrides to steer traffic toward the incumbent in the new entrant’s markets;
189
190
paying higher upfront commission rates on routes where it competes with a new entrant;
biasing its computer reservations systems against non-affiliated interline connections;
refusing to enter into ticketing-and-baggage, joint-fare, and code-sharing relationships
with the new entrant; refusing to lease gates, provide services, or sell parts to the new
entrant; restricting airport operators with majority-in-interest clauses to prohibit the
construction of gates and other infrastructure for new entrants; and prohibiting affiliated
regional feeder airlines from entering into marketing agreements with the new entrant”
(Cooper 1999; Dempsey 2000b, p. 24). Half of the informal complaints received by DOT
between 1993-1999 were regarding unfair pricing and capacity increases, where as more
than 30% were regarding restrained access to gates and other facilities or services
(1999d).
Increasing complaints and actual observations of unfair conduct motivated
the DOT to propose a set of guidelines to protect competition in the airline
industry in 1998. The guidelines essentially proposed that predation would be
inferred if:
(1) the major carrier adds capacity and sells such a large number of seats
at very low fares that the ensuing self-diversion of revenue results in lower
local revenue than would a reasonable alternative response,
(2) the number of local passengers that the major carrier carries at the
new entrant’s low fares (or at similar fares that are substantially below
the major carrier’s previous fares) exceeds the new entrant’s total seat
capacity, resulting, through self-diversion, in lower local revenue than
would a reasonable alternative response, or
(3) the number of local passengers that the major carrier carries at the
new entrant’s low fares (or at similar fares that are substantially below
the major carrier’s previous fares) exceeds the number of low-fare
passengers carried by the new entrant, resulting, through self-diversion, in
lower local revenue than would a reasonable alternative response
(1998c).
191
The responses for the above guidelines were mixed, and generally the
parties who would be distracted by the application of it opposed to it while those
who suffered from/disturbed by anticompetitive practices supported it. Therefore,
the major airlines, their trade association, certain academics, and unions opposed
the guidelines (Gattuso and Boudreaux 1999), whereas the low-fare airlines,
GAO, and Alfred Kahn (1998) supported them.
An interesting counter-proposal to the DOT guidelines came from Foer (1999).
He argued that the DOT guidelines would not be effective since the courts take too long
to resolve the cases (if ever) and that the low-cost carriers go out of business long before
then. He emphasized the need for greater predictability and proposed a “Safe Harbor”
option. A major carrier (defined as 50% or more marker share from a hub) could take the
Safe Harbor option when challenged by a new entrant. A major carrier that chose to
comply with the Safe Harbor option would not be subject to any predation lawsuits. That
option would require them to make a public statement that they will commit to their
response level (low) fares and increased level of capacity for at least two years. They
would have to file a public notice to the DOT. And with certain exceptions, they would
not be able to increase their prices or decrease their capacity for a period of two years. He
suggested that the majors would not alter their price and capacity unless they were
willing to live with it and suffer the losses for two years. The pattern where the major
goes back to the high (monopoly) levels of pricing and capacity would be effectively
constrained with the Safe Harbor (Foer 1999). Naturally, ongoing antitrust law would
have to be modified before Safe Harbor could become successful. Otherwise, no major
carrier would opt for it when they have the option to predate and not be found guilty
192
under existing laws.
Canada, facing similar predatory problems in their airline industry passed
Regulations Respecting Anti-competitive Acts of Persons Operating a Domestic Service,
S.O.R./00-324 in year 2000. Predation by incumbent would be detected based on two
rules with respect to AAC:
(a) operating capacity on a route or routes at fares that do not cover the
avoidable cost of providing the service; and
(b) increasing capacity on a route or routes at fares that do not cover the
avoidable cost of providing the service (cf. West 2000).
Even though an average avoidable cost test has its own complexities such as
determining which costs were avoidable and the appropriate time frame, it has been
successfully employed in at least one case in Canada (West 2000).
An impartial body, the Transportation Research Board of the National Research
Council encouraged DOT to continue their investigations. They observed that “a cursory
review revealed some actions that were difficult to reconcile with fair and efficient
competition. Particularly difficult to reconcile were cases in which incumbent carriers
added nonstop service in low- to moderate-density markets they had not previously
served directly, coincident with a new entry. In some of these cases, the incumbent
bypassed its own hub to initiate the service, a strategy seldom employed outside of high-
density markets. The logical inference is that such responses are probably temporary—
possibly calculated to protect the incumbent’s hub traffic and to dissuade similar
challenges elsewhere—and would seem to warrant additional scrutiny” (1999d, p.33).
193
Inspired by DOT’s guideline proposal, DOJ decided to file a lawsuit against
American Airlines in 1999. This was the first DOJ predatory pricing lawsuit in the airline
industry since deregulation (Nannes 1999). DOJ alleged that American had used
predatory tactics in order to drive Vanguard, Western Pacific, and Sunjet out of its
markets (1999e, para 23-28):
American has monopoly power in most of its DFW city pairs and faces
little current competition and little prospect of entry on those routes. Its
monopoly power allows it to charge supracompetitive fares. American’s
fares on DFW city pairs are substantially higher then its fares on
otherwise comparable routes where it faces competition….[W]hen a [low
cost carrier] entered a DFW route and it appeared the [low cost carrier]
would be economically viable if American simply followed a profit-
maximizing business strategy, American would instead saturate the route
with enough additional capacity at low fares to keep the entrant from
operating profitably. American would also take further steps, such as
matching the [low cost carrier’s] connecting fares with its own nonstop
fares, to keep traffic away from the [low cost carrier]. To evaluate the
success of its strategy and determine whether to intensify its response,
American would investigate the financial resources of [low cost carriers],
determine their break-even factors, and conduct head counts at the
departure gate to monitor their passenger loads.
As a result of these practices all three entrants had to leave the markets
they entered, and all were bankrupt before September 11, 2001.
Table 5.15: American Airlines’ Actions Relevant for the DOJ Lawsuit
American Before During After
From Dallas
To
Flights Price Flights Price Flights Price
Kansas City 8 $108 14 $80 11 $147
Long Beach 0 --- 3 $86 0 ---
Colorado
Springs
5 $150 7 $81 6 $137
Source: (Carney and Zellner 2000).
194
Vanguard markets: American operating at 65% load factor charged $108 before
Vanguard’s entry on DFW to Kansas route. They reduced their rates to $80 following the
entry matching the fares. American also added 6 more round trips to the route, driving
Vanguard out of the market in the process. After Vanguard pulled out, American
decreased the daily round-trips to 11 from 14 and increased the prices by 80%. It was
also documented that the route went from the worst performing to the best performing
American route after the exit of Vanguard.
SunJet markets: American responded to the entry announcement of SunJet to the DFW –
Oakland route. They matched SunJet’s fares and also entered the DFW-Long Beach route
in which it also matched Sun Jet. SunJet was forced to move out of all DFW routes.
American increased prices by over 30% in the two months following Sun Jet’s exit.
Western Pacific markets: American increased its capacity significantly to put pressure on
Western Pacific. The frequency was increased from 5 to 7 flights per day. As a result of
this response, Western pulled out of the routes and declared bankruptcy. American’s fares
were in the range of $81-105 when Western was in the market. After they pulled out
American went back to its original frequency of 6 and increased the prices to $137 on
average.
As it was discussed previously, cases of predatory pricing a very hard to prove in
the airline industry, and the judge of DOJ’s American lawsuit dismissed it by summary
judgment in May, 2001. DOJ appealed the decision but cold not overturn the ruling.
195
The DOT issued a series of reports in January 2001 re-stating their concern about
the anti-competitive practices in the airline industry (2001e; 2001g; Oster and Strong
2001). They concluded that “incumbent airlines at times have responded to new
competition with fare cuts, capacity increases, and other practices that are apparently
designed to eliminate or reduce competition” (2001g). The DOT reminded that they were
authorized by the Congress to prevent unfair practices in the airline industry and warned
that they had the right and the obligation to prevent such practices even if they do not
violate antitrust laws. After studying thousands of responses their proposed guidelines
generated, the DOT concluded that they would examine each case independently due to
the complexities and differentials in the marketplace, and one governing guideline would
be too simple of an approach. Instead, the DOT expressed a wish to develop a “body of
caselaw based on a more thorough examination of cases of apparent predatory-type
behavior” (2001g, p.11).
The second report was entitled Predatory Practices in the U.S. Airline Industry by
Professors Oster and Strong (2001). It described the patterns of entry and competition,
and included an analysis of predatory conduct and predatory pricing with examples from
the airline industry. This report examines the potential for predatory practices, or unfair
methods of competition, in the U.S. domestic airline industry. One of their two main
conclusions was that “predatory practices may have occurred in the past and are a
recurring possibility in the U.S. domestic airline industry.” They detected high use of
market power in the airline industry, hub premiums and recoupment of losses suffered
during predation against new entrants. Their second conclusion was that the antitrust
laws, as they have been applied in other industries, may not be sufficient to identify some
196
types of predatory practices in the airline industry. Their work is highly relevant to
understand he competitive context of the airline industry, and several of their figures have
been cited and reproduced in this chapter.
The third report issued by the DOT in January 2001 was entitled Dominated Hub
Fares (2001e). This report summarized, updated, and provided new support for the
existence and demonstration of hub fares and the effects of LCC service on fare levels. It
argued that quality service and reasonable fares can co-exist just like major carriers and
LCCs can. Atlanta and Salt Lake City were provided as examples for co-existence. The
report concluded that “the key to eliminating market power and fare premiums is to
encourage entry into as many uncontested markets as possible.”
The DOT received 32 informal complaints from 1992 until 1999. Half of these
were allegations of unfair pricing and capacity responses –the dumping of low-fare
capacity in the city-pair market and, in some cases, added flights. Others were about
unfair marketing, airport handling relationships (higher travel agent commissions) (Oster
and Strong 2001, p.10). The administration concluded that “apparent unfair practices
have occurred in the airlines industry”(2001g, p.114). “Free markets do not exist in a
state of nature. Free markets are things that have to be defined by custom and law”
(Dempsey 2000a, p.487).
5.2.2 The Role of Barriers to Entry
It has been shown that the theory of market contestability did not apply to the
airline industry because of high barriers to entry (Baker and Pratt 1989; Hurdle et al.
1989). The threat of potential entry may not force the airlines to cut their prices since
197
they can readily and effectively react to any entry with price cuts and capacity increases
(Levine 1987). Instead, the incumbents may attempt to raise the barriers to entry and
develop a reputation for predation by competitive and aggressive reactions to entry to
their hub markets. Therefore, these barriers gains special importance when conducting
research in this industry. Among the six generic barriers identified by Porter (1980),
several either do not apply or take on a proxy form in the airline context. For example,
cost advantage of incumbents’ would not necessarily apply since the LCC entrants are
presumably more efficient within this framework. However, barriers such as product
advantages of incumbents are captured by hub formation and code-share agreements.
Both hubs and code-share agreements have been associated with higher fares (2001e;
Hassin and Shy 2000). Figure 5.6 presents proposed proxies for Porter’s barriers to entry
for the airline context.
Porter’s six generic barriers to
entry:
Barriers as reflected in the Airline Industry:
Cost advantages of incumbents The LCC entrants have cost advantages
Product advantages of incumbents Hub Formation
Capital requirements Airport Congestion, Reputation Effects
Customer switching costs Code-Share agreements, Frequent Flier Miles
Access to distribution channels Gate, Slot and Noise Controls and Availability
Government policy Gate, Slot, and Noise Controls
Figure 5.6: A Comparison of Generic and Industry Specific Barriers to Entry
As discussed in a prior section, Karakaya and Stahl (1989) have proposed an
extended list of generic barriers to entry. Several of the items among their extended list
are of interest and captured either in different components of the proposed framework
(e.g., market concentration, price) or under the current proxies of barriers to entry (e.g.,
gate and slot controls and availability is a proxy for limited access to essential facilities).
198
Another barrier to entry suggested by Karakaya and Stahl (1989) is incumbent’s expected
reaction to entry and this has been included in the framework in the form of reputation
effects of incumbent’s predation. It might also be interesting to consider advertising,
product differentiation of incumbents and brand name or trademark as potential barriers
to entry from a marketing perspective. However, these involve further complications in
terms of measurement.
Slot and Gate controls: Certain airports are considered high-density airports. This
indicates that in these airports, an airline must have slots to schedule a flight (take-off or
landing) between 6 a.m. and midnight. For example, FAA’s high density rule permits 48
slots at LaGuardia from 6 am to midnight every day, resulting in (48x18) 864 slots per
day (1995b). Slot controls have been imposed in four major airports (Chicago O’Hare,
New York La Guardia, New York Kennedy, and Washington National) since late 1960s.
It has been estimated that the fares in slot-controlled airports have been 11-15% higher
than non-controlled comparable airports with the exception of New York Kennedy
(Morrison and Winston 1997). Congestion pricing (charging aircraft for their take-offs
and landings according to the delays it imposes) has been offered as an alternative to slot-
controls (Morrison and Winston 1997).
Joe Leonard, CEO of AirTran, has noted that they are able to serve Atlanta with
cheap fares because Eastern Airlines went bankrupt and freed up 22 gates for AirTran
(2001h). Similarly, JetBlue is serving New Yorkers because they were awarded 75 slots
at JFK airport by the FAA (2001h). Landing slots in most major airports are not
available. The incumbent is rarely, if ever, willing to lease extra slots to a new entrant.
199
Even in the rare case of lease, monopoly rents may be collected. The FAA “Buy-Sell”
slot rule lets deeper pockets ensure market share by getting the slots and enjoy
monopolistic pricing (Dempsey 1989). Slot control regulation needs to be significantly
revised so that LCC access to markets is possible and not left to coincidence or awards by
officials. Before the DOT permitted the landing slots to be resold, the largest eight
airlines controlled about 70% of the slots at Chicago O’Hare, New York La Guardia,
Kennedy, Washington National. In a couple of years the figure increased to 96%
(Dempsey 2000a, p.450).
Table 5.16: Percentage of Domestic Air Carrier Slots Held by Selected Groups
Airport Holding Entity 1986 1991 1996
O’Hare American and United 66 83 87
Other Established airlines 28 13 9
Financial Institutions 0 3 2
Post-deregulation airlines 6 1 1
Kennedy Shawmut Bank, American, and Delta 43 60 75
Other Established airlines 49 18 13
Other Financial Institutions 0 19 6
Post-deregulation airlines 9 3 7
LaGuardia American, Delta and US Airways 27 43 59
Other Established airlines 58 39 14
Financial Institutions 0 7 20
Post-deregulation airlines 15 12 2
National American, Delta and US Airways 25 43 59
Other Established airlines 58 42 20
Financial Institutions 0 7 19
Post-deregulation airlines 17 8 3
Note: Some Financial institutions have taken control of slots when incumbent carriers
went bankrupt.
Source: Anderson (1997)
Similarly gate controls are an issue, and the dominance of major carriers in
controlling the gates is illustrated in Appendix D (1999a). Low Fare carriers who cannot
200
secure access to gates are left to the mercy of the major carriers and have to pay hundred
per cent to two hundred per cent premiums to use the gates (Woellert 1998). Spirit
Airlines was reported to face such a fate since it had been outbid by the majors in Detroit
continuously for almost a decade (Ingersoll 1999). Thus, slot and gate controls can also
effectively serve as barriers to entry and lead to increased fares (Hurdle et al. 1989;
Morrison and Winston 1990).
5.2.3 After September 11, 2001
The general belief about the airline industry is that it is not really a profitable
industry and that it should be left alone. Although examples of apparently unfair practices
seem to be common, this general belief has provided the industry some relief with respect
to antitrust investigations. Such empathy regarding the industry is higher than ever after
September 11. The industry has received around $15 billion worth in direct aid and
guaranteed loans. The total direct aid payments by October 2002 have exceeded
$4,299,852,435.3 to some 382 airlines. Table 5.17 illustrates the direct aid that the larger
carriers have been provided by the federal government as of July 25, 2002.
Table 5.17: Direct Aids to Major Airlines by the Federal Government
UNITED AIR LINES, INC. $724,485,485.8
AMERICAN AIRLINES, INC. $656,032,849.3
DELTA AIR LINES, INC. $594,894,536.4
NORTHWEST AIRLINES, INC. $405,525,526.3
CONTINENTAL AIRLINES, INC. $343,398,895.9
US AIRWAYS, INC. $287,219,674.0
SOUTHWEST AIRLINES CO. $264,642,089.9
AIRTRAN AIRWAYS, INC. $27,690,969.2
Source:http://www.dot.gov/affairs/carrierpayments.htm
201
The airlines also laid-off employees and cut back on services in response to
falling demand and increasing security costs. However, the future of the major airlines
still looks gloomy. The consumers save $6.3 billion annually due to low-fare airline
competition (1996). On June 18 2002, The Wall Street Journal reported that ‘Wal-Mart’
airlines were crunching the biggest carriers. The number of passengers carried by the five
major airlines was 10% lower in May 2002 than it was in 2001. One would consider this
a result of September 11, and the turmoil it created, but the traffic of the five biggest low-
cost carriers increased by 11% in the meantime.
The future of the surviving LCCs looks brighter than it did before September 11.
Obviously, the same conclusion cannot be drawn for the major carriers who have yet to
re-structure their costs. Union relations play a big role in restructuring and labor expenses
represent up to 40% of the revenue of a major airline but only 25% of the revenues of
low-cost carriers (Trottman and McCartney 2002). Another advantage of low-cost
carriers is their higher aircraft utilization due to their point-to-point and secondary airport
operations. The transparency of (low) prices and distribution of tickets (e-tickets) have
also helped the low cost carriers. More than 40% of Southwest tickets are booked online
whereas the same figure is around 5% for majors (Trottman and McCartney 2002).
The fare difference between flying business class on a major carrier and
discounter has become so wide that some employers allow their employees to purchase
two discounter seats so that they can get some extra space to stretch out (and still spend
less money than it would cost on a major carrier) (Trottman and McCartney 2002). While
the passenger traffic in April 2002 was down by 10.5% from two years earlier, the major
202
carriers had increased the average cost of a 1000 mile coach ticket had by 14.7%
(Trottman and McCartney 2002).
5.2.3.1 Anticipating Trends
Our major competition over the last fifteen years was other people doing
the same things we do. That’s not the case anymore.
Don Carty, CEO, American Airlines (2002)
In July 2002, Delta announced that the competitive threat is not from the big
players such as American but from the discounters (Fonti 2002). It devised low-fare
competition strategy with McKinsey and Co. (Fonti 2002). In the meantime, Southwest
announced an aggressive expansion plan “connecting the dots” between city pairs
common for business travel (e.g., Baltimore to LA, the highest Southwest round-trip fare
was $598 as opposed to $1127 of United) (Trottman 2002).
In August 2002, Delta, Northwest and Continental announced a marketing
agreement which links flight schedules, frequent flier programs, and access to airport
clubs. A similar deal was earlier announced by United and US Airways (Woodyard
2002). In the meantime, the share of revenues from premium passengers (i.e., business
and first class), average industry yield and revenues have been decreasing (Hazel 2003).
Deriving from the above sections, I anticipate that the following trends could be
observed in the industry in the near future.
1. Industry Consolidation Attempts: As the competition and pressure for
profitability increases, the majors will attempt to consolidate. US Airways is
not in a position to compete with either majors or LCCs. It is likely that it will
203
either shrink to its roots as a regional carrier or be taken over by a major
carrier (Sheth and Sisodia 2001). If the second option is realized, other
mergers would follow. The CEO of America West Airlines recently predicted
that United, Delta and Northwest would remain as major network carriers, and
AirTran, Southwest, and Jet Blue would remain as discount carriers following
anticipated mergers (Sunnucks 2005).
2. Antitrust Action: In relation to the above scenario, the DOJ will have to revise
its criteria for evaluating mergers. DOT reports have shown the effect of the
presence of LCCs on fare levels, and their current measure (HHI) does not
take this into consideration. A new set of criteria could enable consolidation
action as long as there is sufficient LCC presence in the markets.
3. Revision of the Hub-and-Spoke System: The efficiency and cost of managing a
large scale hub-and-spoke system is being questioned. Major carriers may be
better off concentrating on medium and long haul markets and leaving the
short-haul spoke (feeding) operations to regional operators that they own or
control. Regional carriers have better cost structures that can effectively
compete with other LCCs.
4. Restructuring of Corporate Travel Policies: As audio and video conferencing
becomes widespread, corporations will revise what constitutes a need for
travel for whom and at what cost. More and more companies are already
booking flights on LCCs (Woodyard 1999). LCCs will want to get a larger
share of the corporate pie, and they will arrange umbrella deals with them. For
example, ProAir offered unlimited business trips for a flat rate, and Vanguard
204
offered 10-15% discounts for businesses that buy 10 or more trips (Woodyard
1999).
5. Increased Market Access and Growth for LCCs: Permission of the mergers
with special access conditions for LCCs could serve to increase their presence
in key markets. The Federal government has already asked the airports to
reveal their plans for increased competition. More than a dozen hub airports
with high concentration were requested to file “competition plans” in order to
secure federal expansion funding or increase facility charges (Pinkston 2000).
These plans would include new gate plans, ticket counters, and strategy to
attract new (i.e., LCC) airlines. Slot controls are also likely to be revised and
relaxed for the benefit of LCCs. However, the presence of LCCs in bringing
lower fares and stimulating economic growth in communities is well
understood. If the Federal government acts slowly in opening up the markets
to LCCs, the communities are likely to take the matter in their own hands. The
following section demonstrates such a case. This is a very important
phenomenon as it demonstrates the insufficiency of public policy at the
federal level.
Case in point: Communities want LCC Service. Sam Williams, the President of Metro
Atlanta Chamber of Commerce, commissioned a study to examine the fare levels with
benchmark cities and stated “We want more airlines brought in here” (Saporta 1998).
More and more communities are becoming aware of the substantial effect that the
discounters have on prices. Since the public policy makers have been slow to open up
markets to LCCs, some communities have decided to be proactive and have taken the
205
matter in their own hands. For example, the business community in Wichita, Kansas have
raised $4.7 million so that they could attract AirTran (Fonti 2002). The money would be
solely committed for ticket purchases on AirTran for the next two years (Fonti 2002).
Air-Tran would also be able to draw funds if ticket sales fall below expectations. Finally,
the community would spend $600K for marketing AirTran flights in Wichita (Fonti
2002). Wichita was previously served by three majors: Delta, American, and United, and
the community’s multimillion dollar investment would bring much lower fares in return.
Similar deals with AirTran have already been successfully employed in six other markets:
Gulfport/Biloxi-Mississippi, Pensacola-Florida, Tallahassee-Florida, Grand Bahama,
Newport News, Va., and Rochester-New York (Fonti 2002). Admitting the failure of
current policy, the DOT itself stated that communities that have little or no service by
LCCs should actively seek their presence. The communities would have to market
themselves to LCCs, and it was advised that they seek advice from aviation consultants
(1996c).
5.2.4 European Landscape
European policy makers have observed the deregulation process in the US
closely. Unrestricted cross-border flights have been allowed since 1993 (McCormick and
Field 1997). The new start-ups in Europe have adopted Southwest’s business model and
tried to emulate its success. Successful European low cost carriers include Ryanair,
EasyJet, Virgin Express, and Debon Air:
10
10
This section draws heavily from (McCormick 1997) March 31 3B.
206
Ryanair: RyanAir is the oldest of the LCCs in Europe. It took advantage of early
deregulation and started its operation in 1985 flying between Britain and Ireland. It is the
leading carrier on the Dublin-London route. Continental Airlines chairman David
Bonderman owns 20% of Ryanair. Like Southwest, Ryanair intentionally ignores the
major airports due to their congestion and aircraft turnaround times. No frequent flyer
miles, no free snacks. However, this structure enables Ryanair to be able to break even at
50% utilization (and its load factors are around 75%) (Michaels 2000). Ryanair uses only
one type of aircraft (the Boeing 737) and is based in Dublin.
EasyJet: It appears that Easyjet is trying to emulate a combination of People Express and
Southwest. It offers a no-frills service: seats are not assigned (first come first serve), no
free meals, and casually dressed flight attendants. EasyJet also does not use travel agents
or a CRS system for ticket sales. The result is 25% cost savings per ticket and also fewer
employees per aircraft. Its base is London-Luton airport.
Virgin Express: Virgin Express was created when Richard Bronson (Founder and
Chairman of Virgin Group including Virgin Atlantic Airways) purchased European
Business Airlines. The initial problem Virgin Express faced was the high tax and cost
structure in its base, Brussels, Belgium. They also faced the problem of not getting the
slots they needed for efficient operations to their destinations. However, Richard Bronson
is reputed as a tough competitor, and Virgin Express is a strong brand throughout the
European Union. Its current base is Brussels-Zaventem airport.
207
Debonair: Debonair attempts to use a hybrid strategy: quality service at an affordable
price. They do offer quality snacks on board but no meals. Debonair has in-flight
entertainment at every seat and plans to introduce video-on-demand and in-flight gaming.
Like EasyJet, Debonair’s base is London-Luton airport.
As a result of increasing pressure from the emerging LCCs, some large players
who could not adopt to the new competitive field are facing trouble (e.g., Iberia, Alitalia,
Olympic). On the other hand, British Airways has aggressively made acquisitions in
Germany and France to increase its presence. Concentration and prices are expected to
decrease in the coming years in Europe as the benefits of deregulation are greeted by
more consumers.
208
CHAPTER 6
DATA AND METHODOLOGY
6.1 Data Set
The Airline Origin and Destination Survey (DB1B) is a 10% sample of airline
tickets from reporting carriers collected by the Office of Airline Information of the
Bureau of Transportation Statistics. This database is used to determine air traffic patterns,
air carrier market shares and passenger flows. The data includes origin and destination,
distances, passengers per day, average one way and round trip fares for all carriers, fared
pasangers, market shares per city-pair and airport, average yields, market ranks, and
revenues. The data can be aggregated for airport, for markets within an airport or at city-
pair detail. The Database Products Inc. makes the data commercially available in the
form of several CD-ROMs. Access was gained to these databases through industry
contacts. This available data of 28 quarters (from 1
st
quarter of 1993 to fourth quarter of
1999) allows for analysis of aggregate and individual markets. From this analysis, one
could understand and identify market defense behavior in light of the principles discussed
in this dissertation. The content of the databases has been summarized as follows in Table
6.1 (Dixit 2000):
209
Table 6.1: Contents of O&D databases
Variables/Database O&D Plus
Market
O&D Plus
Carrier
Onboard
Passengers yes yes yes
Market Share yes yes yes
Distance yes yes yes
Revenues yes yes no
Yield (cents per
passenger mile)
yes yes no
Onboard Passengers no no yes
Seats available no no yes
Market Rank Yes yes yes
Fared passengers yes yes no
Fare yes yes no
Enplaned Passengers yes yes yes
Other data for code-share agreements among carriers, gate and slot controls and
usage agreements, carrier financial resources, employees etc. come from various DOT
reports (see section on measures) and have been manually and built into the database
extracted through a cumbersome process.
6.2 Methodologies
Most empirical work to date seems to have utilized standard approaches such as
regression analysis as their methodology (e.g., Abunassar 1994). One exception is Dixit
(2000) who used neural networks and binary logit analysis. I have considered a wide
array of methodologies to ensure robust and unbiased results and to strengthen the
validity of the findings a la triangulation. Sophisticated methods such as binary and
ordered logit analysis, multi-level modeling with mixed coefficients (HLM), and event-
history analysis were determined to be feasible. Logit models enabled the analysis of
limited dependent variables; HLM enabled to account for the hierarchical structure in the
210
data; and even-history analysis overcomes the issues with right-censoring (i.e., exits
unobserved in the database). Split sample and double-cross validation have been
employed for validation purposes where applicable.
6.2.1 Logit models
The applicability of logit models to the airline industry context has been shown by
Dixit (2000). The model is quite robust, and its use is familiar in marketing.
The ordinal categorical/binary nature of the dependent variables in the model
(e.g., supra-competitive, competitive, entry/no entry; exit/no exit) facilitates the use of
ordered/binary logit models. Ordered logit implies that the dependent variable can be
ranked but the distance between the categories is not known (Aldrich and Nelson 1984;
Menard 2002). Binary logit implies a dependent variable with two known states.
The assumptions of the limited dependent variable models, similar to multiple
regression models, are as follows (Long 1997):
a. The x’s and ? are not correlated ( \ ) 0 E x ? =
b. Error terms have constant variance
c. Error terms are normally (or logistically) distributed with a mean of zero.
An ordered regression model such as logit (or probit) can be derived when an
infinite range for y is mapped to an observable variable and cut-off points (thresholds) are
established (Long 1997). For example, an ordered logit model with three levels for the
dependent variable would be specified as follows (Borooah 2002):
Yi = 1, if Fi ? ?
1
Yi = 2, if ?
1
?Fi ? ?
2
211
Yi = 3, if Fi ? ?
2
where ?
1,
?
2
? 0; ?
1
< ?
2
, and represent unknown parameters. Te classification for pricing
strategy depends on whether the observed fares surpass the respective thresholds. The
probabilities for Yi for each pricing strategy are:
Pr(Y
i
=1) = Pr(Z
i
+?
i
? ?
1
) = Pr(?
i
? ?
1
– Z
i
)
Pr(Y
i
=2) = Pr(?
1
? Z
i
+?
i
? ?
2
) = Pr(?
1
– Z
i
< ?
i
? ?
2
– Z
i
)
Pr(Y
i
=3) = Pr(Z
i
+?
i
? ?
2
) = Pr(?
i
? ?
2
– Z
i
)
The ordered logit model assumes that the error term (?
i
) is logistically distributed
(on the hand, an ordered probit model would assume the error term to be normally
distributed. Greene (2000) suggests that while it is hard to specify this distribution based
on theory, the results are equivalent in most applications). Therefore, the cumulative
distribution function of the random variable X is:
exp( ) 1
Pr( ) ( )
1 exp( ) 1 exp( ))
x
X x x
x x
? = ? = =
+ + ?
and the logit model becomes:
1
1
1
Pr( 1) ( )
1 exp( )
i i
i
Y Z
Z
?
?
= = ? ? =
+ ?
2 1
2 1
1 1
Pr( 2) ( ) ( )
1 exp( ) 1 exp( )
i i i
i i
Y Z Z
Z Z
? ?
? ?
= = ? ? ? ? ? = ?
+ ? + ?
2
2
1
Pr( 3) 1 ( ) 1
1 exp( )
i i
i
Y Z
Z
?
?
= = ? ? ? = ?
+ ?
212
The estimates for
1 2
, ,
k
? ? ? are estimated through the maximum likelihood
function. When the intercept is explicitly included, the following general notation
is revealed:
0 0
1
ik
K
i k i i i
k
F X Z ? ? ? ? ?
=
= + + = + +
?
The binary model is simpler in that only one threshold point has to be defined for
the dependent variable.
6.2.2 Multi-level mixed coefficient models (Hierarchical linear modeling)
Multilevel analysis focuses on the analysis of data with complex patterns of
variability, with a focus on its nested sources (Luke 2004; Snijders and Bosker 1999).
Much of the phenomena in social and behavioral sciences consist of nested layers (Kreft
and Leeuw 2002). Variables may relate to individuals, groups, teams, departments,
organizations, countries, trading blocks and so forth. Multilevel models have been
developed to deal with hierarchically structured data. These models are referred to as
mixed-effect and random effect models in biometric applications, random coefficient
regression models in econometrics, and covariance component models in statistics
literature (Raudenbush and Bryk 2002). It is also referred to as Hierarchical linear
modeling in the social sciences because of popular software package with the same name.
A random coefficients multi-level model is said to exist when there are at least two
levels, and the coefficients at the lowest (micro) level are treated as random variables at
the second level. In contrast, mixed effect models assume that some of the coefficients
are fixed and some are random (Snijders and Bosker 1999).
213
Adding new level(s) to the model renders it more general and often more useful
(Kreft and Leeuw 2002). Analysis from data aggregated at different levels may produce
different results when a hierarchy exists. This can be determined by intra-class
correlations. If intra-class correlation is practically zero, then group differences do not
exist and can be ignored in the analysis. However, by assuming and modeling this
correlation, the implicit nested structure of the data is accounted for. On the other hand, if
this correlation is substantial and is ignored, the reliability of the results become
questionable (Kreft and Leeuw 2002). Therefore, the choice for the level of aggregation
used to analyze the data becomes important. If we aggregate the data, we may potentially
discard meaningful data and also introduce an aggregation bias. On the other hand, if we
conduct our analysis at the individual (i.e., micro) level, we violate the independence
assumption.
Overall, multi-level models help the researcher in three main ways (Raudenbush
and Bryk 2002, p.7):
1. Improved estimation of effects within individual units (e.g., developing an
improved estimate of a regression model for an individual school by borrowing strength
from the fact that similar estimates exist for other schools)
2. The formulation and testing of hypotheses about cross-level effects (e.g., how
varying school size might effect the relationship between social class and academic
achievement within schools)
3. Partitioning of variance and covariance components among levels (e.g.,
decomposing the covariation among set of student-level variables into within- and
between-school components).
214
These features enable researchers to construct more realistic models and test
hypotheses across levels with greater accuracy and reliability. The use of multilevel
models has been recommended unless there is reason to believe that hierarchical structure
in the data is not an issue. The method is emerging and more and more researchers in
marketing have been utilizing it (e.g., Malhouse et al. 2004; Van den Bulte 2000).
Luke (2004, p.13-14) categorized multilevel models in three main classes:
1. Unconstrained
1:
ij oj ij
Level Y r ? = +
00 0
2:
oj j
Level u ? ? = +
Mixed-Effects model becomes
00 0 ij j ij
Y u r ? = + +
This model is used as a null model to estimate between group effects.
2. Random intercepts:
1:
ij oj ij
Level Y r ? = +
00 01 0
2:
oj j j
Level W u ? ? ? = + +
or with explanatory variable
1
1:
ij oj j ij ij
Level Y X r ? ? = + +
00 0
2:
oj j
Level u ? ? = + ;
1 10 j
? ? =
Mixed-Effects Models are
00 01 ij j oj ij
Y W u r ? ? = + + + and
00 01 ij ij oj ij
Y X u r ? ? = + + +
respectively.
3. Random intercepts and slopes
0 1
1:
ij j j j ij
Level Y X r ? ? = + +
0 00
2:
j oj
Level u ? ? = + ;
1 10 1 j j
u ? ? = +
215
with cross-level interaction terms (W
j
represents the interaction term):
0 1
1:
ij j j j ij
Level Y X r ? ? = + +
0 00 01
2:
j j oj
Level W u ? ? ? = + + ;
1 10 11 1 j j j
W u ? ? ? = + +
The mixed effects models are
00 10 0 ij ij j ij ij ij
Y X u u X r ? ? = + + + + and
00 01 10 11 0 1 ij j ij j ij j j ij ij
Y W X W X u u X r ? ? ? ? = + + + + + + respectively.
6.2.3 Event-History analysis
11
An event is described as a qualitative change that occurs at a specific point in
time. Therefore, while a price change no matter how drastic, does not necessarily
constitute an event, market entry and exit are considered events. It has been suggested
that the best way to examine the causes and consequences of such events is through
event-history analysis. If standard methods are applied to event-history data, the results
can be severely biased due to censoring and time-varying explanatory variables. The
event-history analysis has found many applications including but not limited to
unemployment studies, consumer behavior studies (i.e., brand choice over time), medical
studies on the course of illness, learning experiments in psychology and instruction
research, insurance and accident studies, studies of migration, analysis of family
formation and fertility, criminology studies ad legal research, organization and
management research (cf. Blossfeld et al. 1989). There are several dimensions that need
to be considered before one can apply event-history analysis. I consider these in view of
my model and data set.
11
This section is based on Allison (1984) except where cited otherwise.
216
Distributional versus regression methods. Most of the early research that used
event-history analysis has focused on the distributional issues such as the time until an
event or the time between events. More recently, the focus has shifted to regression
models in which the event is the dependent variable where its occurrence depends on
explanatory variables. Similarly, the factors that lead to market entry/exit rather than the
time that passes until a market entry is of interest for the current research.
Repeated versus nonrepeated events. Events can be repeated (i.e., marriage) or
non-repeated (i.e., death). Not surprisingly, repeatable events are more complex to model
but can be handled with event-history analysis. There can be more than one entry/exit in a
given market in a longitudinal study or one can simply assume that it is a one-time event
for a market.
Single versus multiple kinds of events. Events can be treated as identical or
distinguished from one another (i.e., voluntary versus involuntary job termination). As
such, market entry and exit in my models can be treated as identical or categorized as
inter-entry and intra-entry.
Parametric versus nonparametric methods. Non-parametric methods make few or
no assumptions about the distribution of the data. Alternatively, the researcher can
assume a specific distribution (e.g., Weibull). Hybrid approaches have also been
developed in which the regression model has a specific functional form but the
distribution of the event times is not specified (resembles the linear models where no
distribution is assumed for the error term).
Discrete versus continuous time. If the researcher is able to measure the “exact”
event time or assume a continuous scale, then continuous models can be used. Otherwise,
217
discrete units (e.g., quarterly data) can also be handled by the event-history method.
Discrete time models are also easier to implement.
In simple form event-history analysis is conducted in the form of life-tables. This
essentially implies that the survival time data is grouped.
If we define time intervals as I
j
where j=1,….J : I
j :
[t
j
, t
j+1
), where
D
j
: the number of failures observed in interval I
j
M
j
: the number of censored spell endings observed in interval I
j
N
j
: the number at riskof failure at start of interval
S
j
, F
j
: survival and failure functions for interval j respectively
If we assume that the transitions are evenly spread, then the notation is (Jenkins 2004):
2
j
j j
d
n N = ? and
1
ˆ
( ) (1 )
j
k
k k
d
S j
n
=
= ?
?
since S(t)=1-F(t):
1 1
ˆ ˆ ˆ ˆ
( 1) ( ) ( ) ( 1)
ˆ
( )
j j j j
F j F j S j S j
f j
t t t t
+ +
+ ? ? +
= =
? ?
hazard rate is estimate for mid-point of the interval becomes
ˆ
[ ( )]
ˆ
( )
( )
f j
j
S j
? =
where
ˆ ˆ
( ) ( 1)
( )
2
S k S k
S k
+ +
=
While the notation becomes more complex when we introduce explanatory
variables to the model, a Cox proportional hazards model with two time-constant
variables can be simply generalized as:
Log h(t) = a(t) + b
1
X
1
+b
2
X
2
;
where a(t) can be any function of time (Allison 1984).
218
The main methods that are utilized are contrasted with each other and other
alternatives in Table 6.2 in order to demonstrate their relative advantages.
Table 6.2: Comparison of Alternative Methodologies
Methodologies Pros Cons
Multiple Linear
Regression
• Easy to interpret and
communicate
• Interaction and dummy
variable coding
• Not suitable for binary
dependent variables
Logit Models • Can effectively deal with
binary/ordered/ multinomial
dependent variables
• Arbitrary choice for
dependent variable
• Cannot deal with censored
observations
Hierarchical Linear
Modeling
• Considers the nested structure
of the data and cross-level
correlations
• Time-consuming to prepare
data
Event-History
Analysis
• Can effectively deal with
censored observations
• Utilizes full information for
the dependent variable
• No assumptions on the nature
and shape of the hazard
function (Cox)
• Can deal with time-varying
explanatory variables
• Data hard to get and time
consuming to prepare
• Feasible only for qualitative
changes with known times
Structural
Equation Modeling
• Tests the full model rather
than a stepwise approach
• Not feasible when the
variables have single
indicators or formed of
formative (not reflective)
scales
Time-Series Models • Considers autocorrelation in
the data and can establish
cause and effect relationships
• Data hard to get and very
time consuming to prepare
Game-Theoretic
experiments
• Enables precise modeling of
timing and information levels
• Difficulty with complex
models prohibits
construction of a
generalizable model
• Hard to find executive
subjects
6.3 Measures
The operationalization of the variables for data analysis are described in this
section.
219
Relevant markets: An inquiry of market defense immediately raises the question as to
what constitutes a relevant market in the airline industry. One possible approach would
be to assess the whole nation as the market. It may be argued that even though no carrier
serves every city, there are no such restrictions prohibiting from doing so. However, with
this definition of the relevant market --six major carriers, three of which are about equal
size, and dozens of smaller carriers-- one would be misled to conclude that the industry is
indeed not concentrated. For all practical purposes, considering regional markets as the
relevant market level does not clarify the customer’s dilemma either. Let us consider the
case of the Southeast region: For a passenger who wants to fly from Atlanta to Savannah,
the competitive service that Air-Tran offers on its Atlanta-Orlando flights provide little
relief. A product/service market should have its own elasticity of demand as derived from
consumer preferences. Therefore, from a marketing perspective, the relevant market for
the consumers is the substitutable airport/city pairs. It should be noted that several DOT
reports and recent literature have also taken this view regarding the definition of the
relevant market in the airline industry (e.g., 2001d; Dempsey 2000b; Dixit 2000; Nannes
1999; e.g., Oster and Strong 2001). “We have consistently found that relevant airline
markets are generally no larger than city-pair routes” (1996b, p.5).
12
Pricing Strategy: This was the average yield (Average price / distance) for the incumbent
for the market in question for a given quarter.
Distance: is the physical distance in miles between the two cities that make the market
12
The relevant airline markets may be narrower than city pairs (1996b, p.6). A relevant
market investigation at the segment level (e.g. business class city-pair market, coach class city-
pair market) would be fully justified. Unfortunately, such an inquiry was not possible with the
available data.
220
Market Size: Number of total passengers in the market for a given time frame (i.e.,
quarter).
Market Power: The market power of the incumbent was measured by the Herfindahl-
Hirschman (HHI) index. The index is essentially the sum of the squared market shares of
all firms in a market (i.e., revealing 10000 maximum for a pure monopoly). The primary
advantage of the HHI is that it is more sensitive to the size distribution of firms than a
four firm concentration ratio. HHI performed better than simple market share during
preliminary analysis. An alternative measure to the HHI index was developed but did not
perform better (see Appendix B).
Barriers to Entry: Barriers to entry were formed as a composite variable. The factors that
were considered in the formation of the variable were as follows:
Market-specific barriers: Whether any of the cities that form the market is a hub; is
subject to slot and noise controls; whether the gates are utilized at capacity; whether the
gates are predominantly leased for exclusive use by one carrier (1999a; 1999d; Dresner et
al. 2001).
Firm-specific barriers: Content analysis was undertaken in trade press to identify carriers
with a reputation for predation for a period of ten years (1989-1999, i.e., pre-dating
DOJ’s American Airlines case). Keyword searches were conducted for “predatory”,
“aggressive”, “anticompetitive”, and “antitrust.” The results consistently revealed the
same pattern of frequency associations. American Airlines had its own a cluster with the
highest frequency count (3), Continental, Northwest and United were grouped together
next (2), Delta, TWA, and U.S. Airways (1) were grouped third.
221
Code-share agreements that were in effect for the major carriers were identified
from the DOT reports. Code-share agreement barriers were coded to exist if the
incumbent and one of its code-share partners operated in the same market.
Strategic Assessment: Strategic assessment was the sum of the composite variables for
entrant resources, strategic fits, and market growth.
Resources: This was the sum of the full-time employees of the low cost carriers that
operated at either end of the market (i.e., either city pair). It was transformed into a
composite variable. Alternative financial measures such as resource slack (working
capital over total assets) were employed as well but did not produce satisfactory results.
For example, the working capital for Southwest airlines was consistently negative,
whereas some smaller airlines’ were positive. Therefore, I have opted for employees as a
measure of potential entrant resources.
Strategic fit: This was the sum of the aggregate market shares of the low cost carriers that
operated at either end of the market. This was transformed into a composite variable.
Market growth: For quarter n, this was measured as the percentage of growth in the
number of total quantity of passengers flying in the market:
1 5
5
n n
n
QP QP
QP
? ?
?
?
. This was
transformed into a composite variable.
Entry event: Entries by the following carriers, JetBlue Airways, Frontier Airlines, Tower
Air, AirTran Airways, ValueJet Airlines, Kiwi International, Carnival Airlines, Nations
Air Express, National Airlines, Vanguard Airlines, Spirit Airlines, Pro Air, Reno Air, Sun
Country Airlines, American Trans Air, Western Pacific Airlines, Air South, and Casino
222
Express (list acquired from the DOT officials, identical with that in the DOT special
feature study (2003a)) with more than twenty passengers per day (2000c). Pioneer entries
into markets that did not exist before were excluded. Entries by Southwest airlines were
excluded because of the previously discussed “Southwest effect.” Competitors do not
retaliate against Southwest and Southwest very rarely exits from markets it enters.
Southwest’s market capitalization is more than double the rest of the industry combined
(Hazel 2003, p.13).
Incumbent’s response: This was calculated as the percentage difference between the
average incumbent yield before entry (Q
n-1
), and the average of the incumbent yield at
entry and four quarters following entry (Average Q
n
to Q
n+4
). The signs were reversed so
that they measure retaliation rather than change, which makes it easier to interpret.
Incumbent’s retaliation response for exit hypothesis was the percentage difference
between the average yield of the incumbent in the quarter before exit (or the quarter of
last observation) and the average incumbent yield before entry. Use of average responses
(i.e., average of yields after entry) would not necessarily support hypothesized
relationships for the retaliation–exit linkage because an incumbent may keep prices high
for a few quarters and then decrease it sharply (in multiple markets) to drive the entrant
out (which was an observed pattern in the data). Such predatory actions would not be
captured by an average response measure.
Other variables:
Hubs: Major and regional airline Hubs for the purposes of this research are listed in
Appendix C.
HubPower: Weighed average HHI of all markets linked to the hub.
223
HubPremium: Weighed average yield for the hub
HubSpecificBarriers: These are market specific barriers excluding hub formation.
224
CHAPTER 7
DATA ANALYSIS AND RESULTS
H1: Market power of the incumbent and the market price premium will be
positively associated.
H2: The positive relationship between the market power of the incumbent and the
market price premiums will be positively moderated by barriers to entry.
H3: Barriers to entry and the market price premium will be positively associated.
H4: The positive moderating effect of firm specific barriers on the incumbent’s
pricing strategy will be higher than that of market specific barriers.
H5: The positive effect of firm specific barriers on the incumbent’s
pricing strategy will be higher than that of market specific barriers.
H6: The positive relationship between the market power of the incumbent
and the market price premiums will be negatively moderated by the
incumbent’s strategic assessment (i.e., resources, strategic fit,
market growth) of the potential entrants.
H7: Incumbent’s strategic assessment and the market price premium will be
negatively associated.
These hypothesis were tested using multiple linear regression, logit and HLM
models. Top 4000 markets in the U.S. by passenger volume for the 4
th
quarter of 1999
(the most recent quarterly data available) were chosen for the analyses.
13
This was also
the quarter by which extensive code-share agreements had come in effect. The general
MLR model for hypothesis testing with moderating variables (Sharma 1981) was:
6 6
0 3
1 4
i j ij k i ik i
j k
Y X X X ? ? ? ?
= =
= + + +
? ?
Where Y
i
is the average yield for the incumbent for market i
Xi
1
is the distance between city pairs that form market i (control variable)
13
This represented a comprehensive approach as the top 2500 markets approximately account for
90% of passenger traffic.
225
Xi
2
is market size for market i (control variable)
Xi
3
is market power of the incumbent for market i
Xi
4
is strategic assessment of the incumbent for market i
Xi
5
is firm specific barriers to entry for market i
Xi
6
is market specific barriers to entry for market i
After cases with missing data and outliers were excluded, the final sample
consisted of 3948 markets. A quick preliminary run showed that the main effects existed
in the expected directions. All entered variables were significant (p