Description
The ongoing integration of the world economy has not stopped at the real sectors, but has also taken place in the financial sector, where both the crossborder provision of financial services as well as the globalization of banks are on the rise.
Topics in
Multinational Banking
and
International
Industrial Organization
Inaugural-Dissertation
zur Erlangung des Grades
Doctor oeconomiae publicae (Dr. oec. publ.)
an der Ludwig-Maximilians-Universit¨at
M¨ unchen
2007
vorgelegt von
Peter Beermann
Referent: Professor Dr. Dalia Marin
Korreferent: Professor Dr. Monika Schnitzer
Promotionsabschlussberatung: 06. Februar 2008
Contents
1 Introduction 1
2 Multinational Banking - A Literature survey and the case of
Bank Austria in Central and Eastern Europe 6
2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Multinational banking - what do we know? . . . . . . . . . . . 7
2.2.1 Entry motives of multinational banks . . . . . . . . . . 7
2.2.2 Which banks become multinational? Firm and home
country characteristics . . . . . . . . . . . . . . . . . . 10
2.2.3 Which markets attract multinational banks? . . . . . . 12
2.2.4 How do the entry and market strategies of multina-
tional banks look like? . . . . . . . . . . . . . . . . . . 15
2.2.5 How successful are multinational banks operating in
foreign markets? . . . . . . . . . . . . . . . . . . . . . 19
2.2.6 What are the e?ects of multinational bank entry on
the host country? . . . . . . . . . . . . . . . . . . . . . 22
2.3 A case study of successful multinationalization: Bank Austria
Creditanstalt in Central and Eastern Europe . . . . . . . . . . 29
2.3.1 Overview of objects and sources of the case study . . . 29
2.3.2 Bank Austria Creditanstalt: An Overview . . . . . . . 30
2.3.3 Bank Austria Creditanstalt in Eastern Europe . . . . . 31
2.3.4 Entry motives, entry modes and market strategy of
Bank Austria Creditanstalt in CEE . . . . . . . . . . . 32
2.3.5 The future strategy of Bank Austria Creditanstalt in
CEE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.3.6 Same region, di?erent strategy: The case of Erste Bank
AG in Central and Eastern Europe . . . . . . . . . . . 41
2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3 M&A versus Green?eld - Optimal Entry Modes into Markets
with Sequential Entry 47
I
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 The basic model . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.2.1 Analyzing entry (mode) decisions of the sequential en-
trant in period T=2 . . . . . . . . . . . . . . . . . . . 56
3.2.2 The optimal entry mode of the early mover . . . . . . . 61
3.2.3 Benchmark Case (No sequential entry) . . . . . . . . . 64
3.2.4 How does the threat of sequential entry change the
entry mode decision of early movers? . . . . . . . . . . 65
3.2.5 A note on completely endogenous market structure . . 68
3.3 Welfare analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.4 Markets with restricted takeover possibilities . . . . . . . . . . 71
3.5 Country-speci?c learning-by-doing e?ects . . . . . . . . . . . . 73
3.5.1 Sales volumes in period T=1 . . . . . . . . . . . . . . . 74
3.5.2 Sequential entry probabilities in period T=2 . . . . . . 74
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4 When do Banks Follow their Customers Abroad? 84
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.2 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.2.1 Bank pro?t maximization under given liquidation value 91
4.2.2 Loan provision mode choice and endogenous liquida-
tion value . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2.3 When do banks follow their customer abroad? . . . . . 104
4.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5 The E?ect of Bank Sector Consolidation through M&A on
Credit Supply to Small and Medium-Sized Enterprises 115
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.2 Bank Sector Consolidation: An Overview . . . . . . . . . . . . 119
5.2.1 Revenue enhancement as a motive for bank sector M&A120
5.2.2 Cost saving as a motive for bank sector M&A . . . . . 121
5.2.3 E?cient Risk diversi?cation as a motive for bank sector
M&A . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.2.4 Managerial motives for bank sector M&A . . . . . . . . 124
5.3 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.3.1 Inside the small bank . . . . . . . . . . . . . . . . . . . 128
5.3.2 Inside the large bank . . . . . . . . . . . . . . . . . . . 130
5.3.3 Small business ?nancing in small and large banks . . . 138
5.4 The Role of Consolidation . . . . . . . . . . . . . . . . . . . . 139
5.4.1 Changing credit supply within the merged bank . . . . 140
5.4.2 SME credit supply at the market level . . . . . . . . . 141
II
5.4.3 Small ?rms with multiple bank relationships – The odd
?rm out . . . . . . . . . . . . . . . . . . . . . . . . . . 142
5.5 National versus Multinational Consolidation through M&A
and heterogeneous countries . . . . . . . . . . . . . . . . . . . 144
5.6 The Empirics of Small Business Lending and Consequences of
Bank M&A . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
5.6.1 Characteristics of Small Business Lending . . . . . . . 146
5.6.2 Consequences of (Cross-border) Consolidation on Small
Business Lending: The Existing Literature . . . . . . . 146
5.6.3 Bank sector consolidation and relationship manage-
ment strategies of small businesses . . . . . . . . . . . 149
5.7 Findings and Shortcomings . . . . . . . . . . . . . . . . . . . . 149
6 Conclusion and Outlook 157
Bibliography 162
III
List of Figures and Tables
1.1 Foreign Bank Control (% of assets)…………………………………………... …….2
2.1 BA-CA Operations in Central and Eastern Europe…………………………… …...32
2.2 Erste Bank AG Operations in Central and Eastern Europe…………………… …...42
3.1 Time Structure of the Model………………………………………………….. …...52
3.2 The General Structure of Entry Mode Decisions……………………………... …...55
3.3 The Effect of Early Entry Mode on Sequential Entry………………………… …...61
3.4 Comparison of Benchmark and Sequential Entry Case………………………. …..67
3.5 Contingent Sequential Entry Probabilities and Transport Costs……………… …..76
4.1 Time Structure of the Model………………………………………………….. …...91
4.2 Equilibrium Bank Strategies - The Case of Severely Cash-Strapped or
Moderately Cash-Strapped/Small Effort Problem firms………………………
….108
4.3 Equilibrium Bank Strategies – The Case of Moderately Cash-Strapped Firms
with a Large Effort Problem…………………………………………………...
….108
5.1 Volume of M&A Activity in the Banking Sector…………………………….. ….116
5.2 Correlation Analysis of Bank ROE among Nations…………………………... ….124
5.3 Relationship Length and Specialisation Decision…………………………….. ….130
5.4 Bank-Firm Relationships Affected by M&A…………………………………. ….140
IV
Acknowledgements
During the last years working on this thesis I have experienced immense
support, both academically and personally, by a large number of people.
First of all, I am deeply indebted to my supervisor Prof. Dalia Marin for
her wonderful support, giving me the opportunity to pursue my research as
a member of an exceptional faculty. She continuously encouraged me and
o?ered plenty of ideas for this thesis. Discussions with her were extremely
rewarding and instructive. Moreover, her continued rigorous demand for clar-
ity in internal seminars really pushed my understanding and argumentation
to another level.
I am also thankful to my colleagues Alexander Raubold, Thorsten Hansen,
Maria Lehner, Christa Hainz and Georg Gebhardt for countless fruitful scien-
ti?c discussions. Numerous other members of the faculty made this research
experience rewarding.
Special thanks go to Mr Gerhard Smoley, Ms Jana Schwarze and Mr
Christoph Schropp, whose exhaustive practitioner’s insights helped my un-
derstanding of the complex banking industry considerably.
I am deeply grateful to Ms Sabine Schmidberger for encouraging me to
keep working through numerous setbacks one experiences when developing
such a body of research. Last but not least I would like to thank my whole
family for supporting me throughout my academic career.
V
Chapter 1
Introduction
The ongoing integration of the world economy has not stopped at the real
sectors, but has also taken place in the ?nancial sector, where both the cross-
border provision of ?nancial services as well as the globalization of banks are
on the rise.
Though no way a new phenomenon
1
, the internationalisation and multi-
nationalisation of the banking industry has picked up steam (again) in recent
years, with the real new feature being the unprecedented scale of foreign di-
rect investment (FDI) in the banking sector.
This trend was aided by world-wide ?nancial liberalization and deregula-
tion for the banking sector, an increased globalisation of the real sector and
technological advances reducing information costs in international ?nancial
services activities (e.g. Soussa (2004)[121].
In some countries, primarily in Latin America and the former transition
economies in Central and Eastern Europe, foreign direct investment in the
banking sector has gone so far as foreign bank subsidiaries now being the
dominant market players in these countries, cumulatively holding market
shares of close to or more than 50% in many countries in these regions (see
the following table).
1
Jones (1990)[87] traces multinational banking back to the Middle Ages, when Italian
bankers established branches in foreign countries to assist in international banking service.
1
Table 1.1: Foreign Bank Control (% of assets)
0 10 20 30 40 50 60 70 80 90 100
Czech Republic
Hungary
Poland
Turkey
Mexico
Argentina
Peru
Chile
Venezuela
Brazil
Colombia
Malaysia
Korea
Thailand
1994
2001
Source: Inter-American Development Bank
In general ?nancial sector FDI to emerging market economies increased
sharply from 6 $billion in the period 1990-1996 to nearly 50 $billion from
1997 to 2000 (numbers as in BIS report[63]). At the same time an increasing
number of large cross-border merger and acquisition deals has taken place in
the banking industry in OECD countries. In Europe, for example, the top 10
Cross-border M&A deals in banking alone amounted to an investment volume
of 101.4 $billion from 1995 to 2006 (source: The Economist (2006)[127]), with
the largest intra-European bank merger to date having been the acquisition
of HypoVereinsbank AG group (GER) by Uni Credit Group (IT) at a value
of 22.3 $billion.
Amid this surge both new political concerns and scienti?c questions have
arisen. From a normative perspective, governments are in need of answers
concerning the e?ects of foreign bank entry on the economy, while banks are
in need of best practices to shape their international expansion strategies.
Taking a positive perspective, economic science has to come up with expla-
nations concerning politics towards multinational banking (e.g. legal entry
barriers) and the observed patterns of international expansion by multina-
tional banks.
The banking sector is widely perceived to be one of the main important
”‘strategic”’ sectors in an economy, due to its role in providing capital to
2
all other sectors in the economy. Unsurprisingly then, the increasing glob-
alisation of banking has invoked both fears and hopes among governments
concerning the e?ects of foreign bank entry into their local markets. Whereas
there is general hope in emerging market economies, that such entry might
help in accessing international capital sources and aid in the e?ciency de-
velopment of the domestic banking sector, fears have arisen about potential
negative e?ects of foreign bank entry on the survival probability of domestic
banking institutions, potentially harmful ”‘overconsolidation”’ of the sector
and entry e?ects on loan provisioning for informationally opaque market seg-
ments, especially to small and medium-sized enterprises.
Besides scienti?c interest in the latter direct economic policy-relevant
issues, questions on why, when and how banks go abroad are of major interest,
especially because entry motives and strategies might shape foreign bank
entry e?ects on the host economy. For example, as will be discussed in the
literature overview in this thesis, it is observed, that foreign bank entry via
the set-up of a new physical structure (Green?eld Investment) has di?erent
e?ects than entry via the acquisition of a local incumbent bank (M&A). Also,
e?ects of such entry might depend on which banking segment the entrant
foreign bank focuses on in the host country market.
In this thesis, besides trying to give both an overview as well as a prac-
titioners view on the general issues of multinational banking, I focus on
theoretically exploring three questions, concerning banks’ reasoning for in-
ternational expansion due to follow your customer-considerations, the opti-
mal entry mode into foreign markets and how consolidation via M&A in the
banking sector might a?ect credit supply for the most likely victims of such
a development, small and medium-sized enterprises.
Chapter 2 of this thesis starts out with ?rst giving a short literature
overview about six basic questions that arise in the context of multinational
banking, namely
1. What are the motives of banks for expanding internationally?
2. Which banks become multinational players?
3. Which markets attract multinational banks?
4. How do entry and market strategies of multinational banks look like?
5. How successful are multinational banks operating in foreign markets?
6. What are the e?ects of multinational bank entry on the host country?
3
Additionally a case study of a bank highly successful in its international
expansion, Bank Austria Creditanstalt, studying its penetration into the
Central and Eastern European region, is conducted. Based on ?rst-hand in-
formation from a Bank Austria practitioner, I derive a number of insights for
a general discussion about the international expansion strategy of multina-
tional banks. Generally, the study shows how strongly various dimensions of
international expansion strategy are intertwined and how such strategies are
shaped by ex ante characteristics of the bank and its background. Also, the
?ndings shed some further light on the importance of the follow your cus-
tomer motive in banks’ decisions to go abroad, as well as e?cient strategies
in identifying host market characteristics underlying the location decision of
multinational banks.
Chapter 3 tries to answer the question of whether ?rms in general should
enter a host market via the establishment of a completely new structure
(Green?eld Investment) or via the acquisition of a domestic incumbent. I
theoretically discuss this question in a setting of sequential entry and ?nd,
that one reason for the dominance of Green?eld Investment concerning the
number of occurrences might be its strategic advantageous e?ect of deterring
potential further entry into the market. While the theory adds to the general
literature on entry modes in foreign direct investment, the results are valid for
the banking industry, too. Extensions of the model are proposed to discuss
entry modes in industries with limited local takeover possibilities and strong
country-speci?c learning-by-doing e?ects, of which the latter might further
help in understanding entry mode choice in speci?c banking segments such
as retail banking.
Chapter 4 discusses the decision of bank entry into foreign markets under
the follow your customer-motive. This motive is among the most discussed
reasons for banks to go abroad, however no formal theory had been brought
forward to my knowledge. In a setup of a double moral hazard problem
within a real sector ?rm I discuss its home bank’s incentive to follow the ?rm
abroad to provide ?nancing for a host country subsidiary to be set up. I ?nd
that the decision of a bank whether to provided its client with ?nancing from
its home base, via actually establishing a physical presence abroad or not at
all, depends on client, bank and host country characteristics, namely the
relative magnitude of the two moral hazard problems the parent ?rm faces
in interaction with its foreign subsidiary manager, the general e?ciency of the
bank in the liquidation of project assets and the factor endowment or stage
of development of the ?rm’s FDI host country. These single characteristics
are found to not in?uence the banks behaviour uniquely, but rather interact
with each other to shape the respective banks service provisioning strategy.
In chapter 5 I propose a general theory on how active consolidation (via
4
M&A) in the banking sector might a?ect the credit supply for small and
medium sized enterprises (SMEs). Though set up as a general theory for the
banking sector, the model also applies to changes in the industrial organi-
zation in banking markets induced by the entry of foreign banks. I study
this topic by modelling capital allocation decisions within banks, formally
introducing relationship lending considerations. Restricting the analysis to
banks’ incentives to lend to SMEs, the model yields the result that consol-
idation in the form of at least one existing bank becoming part of a larger
organisational structure leads to a potential reduction of credit supply to a
fraction of SMEs in the market. Additionally, I give a smaller literature re-
view on why banks might want to grow larger and how consolidation a?ects
credit supply to SMEs empirically.
I ?nally conclude this thesis by summarizing my ?ndings and discussing
potential future trends that will shape the further evolution of multinational
banking.
5
Chapter 2
Multinational Banking - A
Literature survey and the case
of Bank Austria in Central and
Eastern Europe
2.1 Overview
The purpose of the following chapter is to discuss the main ?ndings of em-
pirical literature on multinational banks as well as to discuss entry motives
and entry strategies by the example of the most important foreign players in
Central and Eastern Europe, namely Austrian Bank Austria Creditanstalt.
For the literature review I will focus on the following questions surround-
ing multinational banking.
1. What are the motives of banks for becoming multinational or generally
acting in foreign markets?
2. Which banks become international/multinational players?
3. Which markets attract multinational banks?
4. How do the entry and market strategies of multinational banks look
like?
6
5. How successful are multinational banks operating in foreign markets?
6. What are the e?ects of multinational bank entry on the host country?
Of course, at least part of these questions are intertwined with each other.
For example, the motives of banks should critically hinge on what kinds of
banks they are, which again also determines whether they become interna-
tional players. Of course, especially taking into account the literature on
vertical versus horizontal FDI, the geographic pattern should also di?er in
the entry motives. Entry motives also will play a role in how the respective
multinational bank will setup their entry and market strategy.
One striking point, is that the e?ect of multinational bank entry should
very much depend on the entry motives. For example, if a foreign bank
enters a host country market to provide services that are neither available
from foreign banks nor close substitutes to the service provided by the latter,
the e?ect of foreign bank entry on the domestic banking sector might be
negligible.
I proceed with a case study of Austrian banks’ entry in Central and East-
ern Europe to shed some further light on these issues. Among the ?ndings
are that home country push factors play a strong role in banks going abroad,
that the availability of follow your customer-strategies signi?cantly supports
international expansion and that the choice of entry mode is strongly con-
nected with the respective market strategy of banks. The latter point also
is a reminder that banks are ?nancial conglomerates, potentially o?ering a
wide variety of services, therefore treating banks as a homogenous group in
a general discussion of multinationalization strategies might be misleading.
2.2 Multinational banking - what do we know?
In the following I want to give a small, by no means complete, review on the
literature about multinational banking.
2.2.1 Entry motives of multinational banks
Broadly speaking, there are two main strands of motives on why banks enter
foreign markets. The one I will not discuss in detail here are managerial
motives. Obviously, entering a foreign market is usually connected with ?rm
growth. Empire-building tendencies of bank managers might therefore be
a simple reasoning why banks want to enter foreign markets, especially if
growth possibilities are restricted in the home country e.g. due to anti-trust
7
considerations. A large literature deals with the empire-building tendencies
of managers and the underlying motives such as status, power, compensa-
tion and prestige of managers of large ?rms (see e.g. Baumol (1959)[10],
Williamson (1974)[131] and Jensen (1986)[85]). An additional managerial
motive for expanding into foreign markets is (potentially ine?cient) risk di-
versi?cation (e.g. Berger et al.(2000)[12]) to prevent cases of bank liquidation
leading to job loss.
The other strand of motives, which I want to discuss in more detail, can
be subsumed under ”‘pro?t-maximization motives”’. One can further di?er
between the motive to expand business in general and the motive to expand
in a speci?c business segment/geographic market.
Concerning motives for general expansion, economies of scale and scope in
various dimensions might play a role. Besides the generally discussed revenue
and cost economies of scale and scope, another focus in the analysis of the
banking industry is on risk diversi?cation economies of scale and scope.
The empirical evidence on the former is mixed as is discussed in more
detail later in of this thesis. However, practitioners strongly support the
view that such economies of scale and scope exist in international expansion.
For example, Spanish bank managers active in the expansion into the Latin
American market perceive a wide variety of such economies e.g. due to the
possibility to develop relatively homogeneous ?nancial products or centralize
back o?ce and transaction processes (Guillen and Tschoegl (1999)[74]. The
majority of studies on this topic also might su?er from old data. Due to tech-
nological advances and changing market possibilities, such economies might
be available to a higher degree today. The evolution of internet banking, the
arrival of Automated Teller Machines (ATMs) and the beginning specializa-
tion of banks along the value chain have most probably increased available
economies of scale (The Economist (2006)[127]). At the same time bank size
and diversi?cation might be an important requirement to be able to place
”‘strategic bets on future markets such as China without putting the whole
bank at risk”’(The Economist (2006), page 4[127]).
The other two well-discussed potential pro?t-enhancement motives for
entering a foreign market are to win new customers in this country (market-
seeking foreign direct investment) or to keep existing domestic customers and
enhance business volume with them (follow your customer-strategy).
The former is the most obvious and generally acknowledged motive for
entry into a foreign market. However, special to the multinational banking
literature, there had been an ongoing debate about whether a foreign bank is
actually capable of successfully entering local retail and commercial banking
markets (e.g. Nolle and Seth (1996)[108]). More recent literature, as well
8
as the overwhelming experience, suggest that at least some foreign banks in
some host markets are able to penetrate local markets on a su?cient scale
(e.g. Berger et al.(2000)[12]).
Compared to the discussion of FDI in other industries, the follow your
customer motive plays a much larger role in the multinational banking lit-
erature. From a ?rm perspective this follow your customer behaviour seems
to have been of large importance for a long time now, as, e.g. in the survey
by Pastre (1981)[111] 52% of U.S. multinational ?rms reported to use one of
their domestic banks for operations in foreign jurisdictions
1
.
The follow your customer motive includes both o?ensive and defensive
strategic traits. For one, a domestic bank might engage in stand-alone non-
pro?table follow your customer-FDI, to secure the respective client’s home
market business with the bank and keep them from switching to another bank
providing global service network capabilities. This so-called defensive expan-
sion approach has been mentioned by e.g. Grubel (1977)[73] or Williams
(1997)[130]. However, following its customer also might be pro?table for
the respective bank per se, as the latter is able to broaden the volume of
business conducted with the respective ?rm, taking over additional trade
?nancing and local cash management services for this ?rm. Also, the geo-
graphical expansion of the bank’s network might attract additional customer
from its home country looking for such ”‘global capabilities”’
2
.
Another pro?t motive for international expansion of banks seems to have
become less of an issue, but has been a big reason for the big wave of inter-
national expansion of U.S. banks in the 1960s and 70s, namely the search for
cheap sources of re?nancing for home market ?nancing activities. This need
arose through ”‘Regulation Q”’ enacted by the Glass-Steagall act in 1933
in the United States, which did put a limit on the interest rates that banks
could pay on deposits in the United States, therefore leading to a shortage
in capital supply for the banks’ loan business. The response was to enter
especially European markets on a large scale to use European deposits to
re?nance U.S. bank ?nancing activities (e.g. Huertas (1990)[80]).
1
Further empirical evidence is discussed at the beginning of chapter 3 of this thesis.
2
Huertas (1990)[80] e.g. cites Frank Vanderlip, former CEO of Citibank on the banks
expansion into Latin America: ”‘ .. I hope to get a very considerable return by o?ering
facilities that other banks cannot o?er to exporters, and thus attract their accounts to
Citibank”’.
9
2.2.2 Which banks become multinational? Firm and
home country characteristics
Just like in the general literature on Foreign Direct Investment (FDI), the
issue of which institution-speci?c characteristics in?uence a banks multina-
tionalization decision has started to garner interest just recently.
This newly arising question is strongly linked with the recent advances
in trade and FDI theory incorporating the fact that ?rms are heterogeneous,
e.g. in the papers by Melitz (2003)[99] and Helpman, Melitz and Yeaple
(2004)[56].
The main proposition of these papers is, that only su?ciently productive
(e?cient) ?rms supply foreign markets. This group can the be further split
into relatively less productive ?rms which will serve foreign markets via ex-
porting, whereas the most e?cient/productive ?rms will establish a physical
presence in foreign markets conducting Foreign Direct Investment.
A similar, and maybe even stronger, view is existent in the multinational
banking literature. As local banks should have inherent advantages over for-
eign banks in the market, due to intimate knowledge about e.g. borrowers’
risks or retail customers’ speci?c preferences, a successful entry by a foreign
bank should only be feasible, if the latter has su?cient advantages in other
bank characteristics to o?set the incumbents’ informational advantages (e.g.
Grubel (1977)[73] or Berger et al.(2000)[12]). Examples for such multina-
tional bank advantages that could be leveraged on a foreign market could
be managerial skill, enhanced risk management and IT systems (e.g. Berger
et al.(2000)[12]) or the ability to re?nance in the capital market or home
deposit market at lower costs.
The empirical literature on e?ciency characteristics is surprisingly scarce
to date. Focarelli and Pozzolo in two studies (Focarelli and Pozzolo (2001)[61]
and Focarelli and Pozzolo (2003)[62]) indeed ?nd, that the more e?cient
3
a
bank, the more likely this institution will run branches and/or subsidiaries
in foreign countries. Also, Buch and Lipponer (2004)[29], studying a sample
of German banks, discover that more pro?table banks
4
are more active
3
Measured as return on assets in these studies.
4
Indeed pro?tability in their study setup is a good proxy for general bank e?ciency, as
the authors control for di?erent business portfolios of the banks as well as for bank size.
Whereas the former therefore controls for pro?t di?erences arising from concentrating on
di?erent segments of the banking industry, the latter ensures that pro?tability is not only
a measure of economies of scale, but rather for underlying X-e?ciency of the respective
bank.
10
internationally, both in the sense that they undertake more foreign direct
investment as well as generate more revenue from international business in
general. However, the authors do not analyze how di?ering pro?tability
a?ects the choice of serving foreign markets predominantly via FDI or cross-
border provision of services.
Bank size per se might be an advantage for banks in their pursuit of inter-
national expansion as well, as large size and scope might enable institutions
to bear larger risks as well as get cheaper re?nancing rates at the capital
market, due to potential economies of risk diversi?cation
5
. A number of
studies indeed ?nd size to have a positive in?uence on the degree of multina-
tional activity of a bank (e.g. Focarelli and Pozzolo (2001)[61],Focarelli and
Pozzolo (2003)[62], Buch and Lipponer (2004)[29] and Tschoegl (2003)[126]).
The latter study even ?nds that, for the US market, it is predominantly the
largest bank from the respective home country, that is most active in the
host market. One reasoning might be, that these respective banks are most
limited in their further domestic growth due to anti-trust regulations.
However, Curry, Fung and Harper (2003)[42] mention that the causality
between the degree of multinationalization and the size of a bank remains a
bit unclear, as bank size might be in?uenced by the fact that the respective
institution is involved internationally and not vice versa.
Furthermore the two studies by Focarrelli and Pozzolo[61],[62] show evi-
dence, that the product focus of banks is a further determinant of the global
scope of its operations. To be precise investment banks, or banks that gen-
erate a high share of their total revenue from non-interest income in general,
are found to be more globalized than traditional loan-processing banks.
Finally, Buch and Lipponer (2004)[29] also show, that previous inter-
national experience increases the probability of a bank to enter a foreign
market
6
.
Incentives and capabilities of banks to become multinational are most
probably endogenous, depending on the characteristics of markets the bank
has been operating in, speci?cally its home market.
The ?nancial development of the home country, the degree of bank sector
competition in the bank’s main market, as well as regulatory conditions in
this home market are deemed by the literature to have an in?uence in both
developing the capabilities of banks to become multinational as well as their
incentive to do so (e.g. Berger et al.(2000)[12], Aliber (1984)[3] and Curry,
5
From this perspective, size is an e?ciency factor, determining available economies of
scale and maybe scope.
6
This ?nding is also backed by information found in conducting the below case study.
11
Fung and Harper (2003)[42]). Theoretically a large market should bread
tendentially larger banks or enable access to larger deposit volumes. Indeed
Brealey and Kaplanis (1996)[26] and Fisher and Molyneux (1996)[59] ?nd a
positive in?uence of home country size on the multinational activity of banks.
Also strong bank sector competition in the host country should lead to
pressure to become more e?cient, while also restricting the possibility to
make abnormal pro?ts, giving incentives and enhancing capabilities to ex-
pand into other (less competitive) markets.
Home market saturation is seen among the main reasons banks start to
look abroad. Guillen and Tschoegl (1999)[74], conducting interviews with
Spanish bank managers, discovered, that the Latin American expansion by
Spanish banks was mainly pursued due to a very saturated Spanish home
banking market featuring strong margin pressure and restricted growth op-
portunities.
Additionally a well-developed capital market might give home banks an
opportunity for cheaper or more ?tting re?nancing options. The role of
home market ?nancial market conditions for general foreign direct investment
has been discussed by Klein, Peek and Rosengren (2002)[88], who ?nd that
the reduction in Japanese Foreign Direct Investment in the 1990s could be
explained by the Japanese banking crisis, which constrained Japanese ?rms’
ability to ?nd su?cient ?nancing for FDI activity.
The role of regulation on incentives to go abroad has already been partly
discussed in the motives for US banks to go abroad (”‘Regulation Q”’). Spe-
ci?c to the U.S., another regulatory restriction for domestic growth might
have played a large role in these banks’ decision to grow abroad. Precisely,
until the Riegle-Neal act in 1994, US banks faced severe restrictions on entry
into multiple US states, making large-scale interstate banking infeasible[12].
So geographic expansion for U.S. banks was mostly only feasible outside of
the U.S..
As mentioned before, another reason for international expansion might be
anti-trust considerations for large domestic banks, which might make further
inorganic (via acquisition) growth in the home market legally infeasible (e.g.
Tschoegl (2003)[126]).
2.2.3 Which markets attract multinational banks?
One area of research on multinational banking, that has already been ex-
plored to a relatively large degree by economic literature, is the location
choice of multinational banks.
The characteristics of an attractive host country can be sorted into two
12
main categories, stand-alone and bilateral host-home country characteristics.
Multiple dimensions of stand-alone characteristics of host country markets
can be distinguished.
Macroeconomic conditions
The macroeconomic conditions attracting banking foreign direct invest-
ment are very much alike these attracting FDI in general.
Numerous studies ?nd a positive in?uence on banking sector FDI of Gross
Domestic Product (GDP) and GDP per capita, e.g. Focarelli and Pozzolo
(2003)[62], Brealey and Kaplanis (1996)[26], Sabi (1987)[116], Buch and Lip-
poner (2004)[29] and Buch and de Long (2004)[28], which is to be expected
from the theoretical literature on horizontal foreign direct investment, e.g.
Markusen and Venables (2000)[98]. Generally, all else equal, a higher GDP
should equal a larger demand for any kind of product. Besides similar rea-
soning, higher GDP per capita might be associated with a higher demand
for sophisticated, high-margin banking services, such as asset management.
In?ation is found to reduce the attractiveness of the market by Focarelli
and Pozzolo (2003)[62], however Buch and Lipponer 2004)[29] do not ?nd a
signi?cant impact of in?ation
7
.
Country risk, as measured by an Index from Euromoney, is found to have
a negative in?uence on international bank activity in a respective country,
be it cross-border lending or bank foreign direct investment, in the study by
Buch and Lipponer[29].
Bank sector and regulation characteristics
One unanimous result across the empirical literature is that countries har-
bouring ?nancial centers (e.g. New York, London, Tokyo) attract a larger
volume of bank sector FDI (see e.g. Focarelli and Pozzolo (2003)[62] and
Buch and de Long (2004)[28]. This easily can be rationalized by a type
forward and backward linkages in the banking industry, as banks buy in-
vestment banking products for their portfolios from other banks as well as
sell such products to other banks. Proximity of respective banks’ investment
banking divisions supports these transactions with heavy information and
trust requirements.
Concerning bank sector regulation, studies ?nd, that the harsher activity
restrictions for multinational banks and the stricter banking regulation in
general, the less bank FDI the respective country will attract (e.g. Focarelli
7
However, Buch and Lipponer argue that positive (higher nominal returns) and negative
(higher instability) e?ects might cancel each other out.
13
and Pozzolo (2003)[62], Berger et al.(2000)[12], Buch and de Long (2004)[28]
and Curry, Fung and Harper (2003)[42]). However, this result does not seem
to hold for less developed countries (e.g. Sabi (1987)[116]).
Other pro?t-in?uencing factors determining whether banks enter the mar-
ket are the size of the banking market (+) (Sabi (1987)[116] the level of con-
centration in the host banking market (-) (Focarelli and Pozzolo (2003)[62]),
and the cost e?ciency of incumbent banks (-) (Berger et al.(2000)[12]). These
in?uence factors obviously shape the expected pro?tability of market entry
through determining the market volume and the degree of competition in the
respective market.
Bilateral home-host country characteristics
Screening the empirical literature, bilateral country characteristics seem
to play a large role in multinational banks’ location decisions, just as the
general theory on Foreign Direct Investment predicts (e.g. Markusen and
Venables (2000)[98]).
Looking at the broad picture one clearly sees strong bilateral patterns
in multinational banking, as found by Soussa (2004)[121]. For example, as
discussed in the case study below, Austrian banks abroad are almost exclu-
sively active in the former transition economies, with 93% of Austrian banks’
FDI stock concentrated in ?ve countries (Czech Republic, Slovak Republic,
Croatia, Poland and Bulgaria). Also, Spanish banks’ multinational activity
is predominantly restricted to Latin America, with a share of 94% of total
outward bank FDI stock in Brazil, Mexico, Argentina, Chile and Colombia.
Similar strict bilateral geographic patterns are found for banks from Belgium
and Italy (strong focus on CEE countries).
From a host country perspective, for example, Spanish banks account for
65% and 58% of total bank foreign assets in Argentina and Brazil, respec-
tively, while e.g. Austrian banks’ foreign asset share in the Czech Republic
is 39%.
A closer, regression analysis-based look on bilateral home-host country
characteristics shows, that the amount of bilateral trade and real sector for-
eign direct investment positively in?uences the volume of bank sector FDI
into the host country (see e.g. Focarelli and Pozzolo (2003)[62], Buch and Lip-
poner (2004)[29], Brealey and Kaplanis (1996)[26] and Fisher and Molyneux
(1996)[59] and Sabi (1987)[116]). This result is consistent with the idea, that
follow your customer-motives play an important role in multinational banks’
location choice.
Additionally common language, low distance and a common legal system
are found to positively in?uence bilateral bank sector FDI in some studies(e.g.
14
Buch and de Long (2004)[28]), however the impact is partly statistically
insigni?cant in other studies (e.g. Focarelli and Pozzolo (2003)[62] and Buch
(1999)[27])
8
.
Looking at how the Latin American and Eastern European markets shape
up concerning the source of inward bank FDI, especially in the Latin Amer-
ican case, language and cultural factors do seem to play a role, as the dom-
inance of Spanish banks among multinational banks in these countries is
blatant. Such a cultural factor is also often mentioned as the reason for the
market leadership of Austrian banks in Eastern Europe, as countries such as
the Czech Republic, Hungary and Slovenia used to be part of the Austrian
Habsburg empire for a long period of time in history.
2.2.4 How do the entry and market strategies of multi-
national banks look like?
Having discussed empirical results on why, which bank, stemming from which
country, enters where, the focus of discussion shifts to the existing literature
on how a bank enters a foreign market.
When deciding upon how to become actively involved in a foreign market,
a bank has two intertwined organizational dimensions to decide on. For one,
the bank has to choose the degree of market involvement (Curry, Fung and
Harper (2003)[42]) it wants to reach. A low level of involvement is achieved
by establishing correspondent banking (cross-border services with the help
of a correspondent incumbent bank in the host country) or by opening a rep-
resentative o?ce or agency. Common to these forms are restricted activities,
with the bank not enabled to engage in deposit taking or direct lending
9
in
the host country. In contrast, via establishing branches or subsidiaries, for-
eign banks are able to ”‘conduct the full range of banking activities”’(Curry,
Fung and Harper (2003)[42]). The di?erence between branch and subsidiary
organizations is, that the former type is not a legally independent organiza-
tional structure, whereas the latter is.
Ball and Tschoegl (1982)[6] ?nd that the main determinant of organiza-
8
In Focarelli and Pozzolo (2003) distance, as expected, in?uences foreign bank activ-
ity negatively and signi?cantly. However, a common language enters insigni?cantly and
ambiguously in the probability of a foreign bank operating either a branch or even a
subsidiary in the respective host country.
9
Agencies might be allowed to engage in commercial lending, but not in other loans or
deposit taking.
15
tional choice in the above dimension is a bank’s experience in the respective
host market as well as its general experience in foreign banking markets.
The former suggests, that markets are entered in a step-by-step approach
starting with low degrees of involvement (e.g. representative o?ce) with the
bank in the following growing its local structure up to possibly establishing
a subsidiary in the market.
Concerning branches and lower-level physical engagement forms, these
will usually be set up via Green?eld Investment.
The second dimension of organizational mode of entry concerns the setup
of a subsidiary in a host market. Entering via the establishment of a sub-
sidiary can be achieved either by setting up a completely new organizational
structure in the market (Green?eld Investment) or via the acquisition of a
local incumbent bank.
Very few empirical studies to date deal with the determinants of entry
modes of multinational banks.
According to de Haas and van Lelyveld (2006)[44], establishing Green-
?eld subsidiaries might be the entry mode of choice, if the parent bank wants
to exercise a high degree of control over the foreign structure. The authors
show evidence for this claim, ?nding that Green?eld subsidiaries are more
closely integrated within the parent organization operationally, having access
to the parent banks’ internal markets for capital and management resources.
In contrast, subsidiaries stemming from the acquisition of local incumbent
institutions enter the multinational bank group with an existing personnel
and capital (deposits and loans) portfolio, and are found to be less integrated
into the parent organizations’ internal markets, and also might need restruc-
turing such as to ?t into the group’s organization and product portfolio. On
a positive side, the promptly available deposits of an acquired bank may en-
able the foreign bank to grow its local loan business faster due to available
local re?nancing
10
. The need for access to the multinational bank group’s
internal capital market might therefore also be less pronounced.
A disadvantage of entry via M&A proposed in the literature especially ap-
plies to entry into non-OECD countries. Whereas acquired banks come with
a portfolio of potentially non-performing loans, a Green?eld subsidiary can
start o? local business without such baggage[35]. This might be an impor-
tant factor, especially when entering countries, that have experienced a loan
crisis lately or harbour a majority of ine?cient banks, also lacking the trans-
parency for multinational entrants to determine their level of engagement in
bad loans
11
.
10
see Curry, Fung and Harper (2003)[42]
11
Indeed, in the case study about Bank Austria Creditanstalt in this chapter, the prac-
16
An important determinant of the entry mode for multinational banks
should also be the respective bank’s segmental strategic focus in the respec-
tive host market. Due to the problems of acquiring soft information about
loan risks, ”‘green?eld banks have an incentive to focus on the most trans-
parent clientele”’(Havrylchyk and Jurzyk (2006)[76]). The two authors, for
banks entering Central and Eastern European banking markets, show indi-
cation, that Green?eld banks on average charge lower interest rates on their
loans than acquired subsidiary banks, which indicates that the former con-
centrate on low risk, informationally non-opaque clients, as this customer seg-
ment should experience more bank competition, and therefore lower interest
margins available, according to theory (Dell’Ariccia and Marquez (2004)[48]).
As the informational advantage of entry via acquisition versus via Green-
?eld investment hinges on the fact of potential customers being informa-
tionally opaque, one might be able to argue, that we should ceteris paribus
observe acquisition to be the dominant mode of entry in countries with in-
su?cient information-providing institutions and a relatively large volume of
informationally opaque potential customers. However these markets might
also be those featuring incumbent banks with a high share of non-performing
loans, which (partially) o?sets the advantage of entering via M&A.
Havrylchyk and Jurzyk (2006)[76] also decompose foreign subsidiaries’
and domestic banks’ pro?ts. A look at their descriptive statistics yields, that
tendentially foreign banks, that enter the foreign markets via acquisition,
show higher net interest margins and higher return on assets than foreign
banks entering via Green?eld Investment . This suggests that banks enter via
acquisition, if they are large, pro?t e?cient and generate a high proportion
of their pro?ts from the traditional banking activity of lending and deposit-
taking.
Some special political developments in countries also play a large role in
the entry mode decision of foreign banks. The main example are the tran-
sition countries in Central and Eastern Europe, where entry via acquisition
of incumbent domestic banks was supported by the large bank privatization
wave in the 1990s, providing available targets for sometimes relatively low,
political prices (ECB (2004)[8]. The above study also claims, that one of
the reasons for entering via M&A in these countries was, that the foreign
entrants main line of business in this region was retail and commercial bank-
ing, operations that require local market knowledge. The authors however
especially caption the importance of the privatization programme ‘All in all,
titioner mentioned uncertainty about the degree of exposure to bad loans of potential local
target banks as a signi?cant barrier to entry via acquisition in Central and Eastern Europe
in the 1990s.
17
the most relevant consideration in the investment strategy of foreign banks
in the accession countries has been to take advantage of the opportunities
provided by privatisation programmes in order to develop a wide and visible
presence in the host markets within a short period of time
12
.
The latter part stresses another crucial advantage of acquisition over
Green?eld Investment, namely entry by acquisition enabling foreign banks
to serve the market on a large scale fast, reducing the time needed to ramp
up business volume to signi?cant levels. For example, a lack of brand name
recognition for an obscure foreign bank trying to enter the retail banking
market via a completely new structure, should lead to slow business growth,
as reputation has to be slowly built up. The argument might however not
hold true for multinational bank icons such as Citibank entering markets har-
bouring incumbent banks with a collective negative reputation, for example
stemming from a recent national banking crisis.
A general empirical literature on entry modes into foreign markets yields
some additional insights, that might also apply to the banking sector.
Hennart and Park (1993)[77], looking at the entry mode decision of
Japanese multinational ?rms into the United States, match ?rm and industry
characteristics with entry mode choice. Results that may also apply to bank-
ing are, that the level of concentration in the respective market positively
in?uences the propensity to enter via Green?eld Investment. Concerning
product strategy, the more similar products of parent and subsidiary, the
more attractive the entry mode of Green?eld Investment compared to ac-
quisition seems to be. However, this might be due to di?ering motives for
acquisition and Green?eld Investment. In di?erence to the latter, the former
might be motivated by trying to add technological knowledge or diversify a
?rm’s business portfolio.
More to the point of a general organizational discussion, the above authors
also ?nd that Green?eld Investment is chosen if operations in the host market
are of small volume. This points back to the discussion about agencies and
branches versus subsidiaries, where the former operations are feasible for low
levels of activity in the host market and can most easily and cheaply be
achieved by setting up this small structure from scratch.
Andersson and Svensson (1994)[4], for a sample of Swedish multinational
?rms, come up with evidence, that older ?rms with strong organizational
skill
13
tend to enter foreign markets via acquisition of local incumbents,
whereas the role of ?rm size in the decision whether to enter via Green?eld
investment or acquisition is unclear. Concerning host country characteristics,
12
(ECB (2004), page 2)[8]
13
The authors use the number of existing a?liates as a proxy for organizational skill.
18
the authors ?nd that acquisition is the preferred mode of entry for developed
countries, as the probability of entering that way positively and signi?cantly
depends on GDP per capita.
Another dimension of entry strategy is concerned with which segment a
bank wants to serve in the respective host market. Tschoegl (2003)[126] dis-
cusses an interesting example how such a decision is, besides obvious market
characteristics, shaped by institutional/regulatory market conditions. Swiss
banks and Deutsche Bank were reluctant to get into retail business in the
United States because of fears they might run into problems with U.S. bank-
ing regulation agencies as an universal bank, as they already were heavily
involved in the securities market in the United States
14
.
2.2.5 How successful are multinational banks operat-
ing in foreign markets?
Having discussed what shapes the structure of multinational banking, I now
take a look at what literature has to say about the success of international
expansion of multinational banks. The following literature gives insights into
how the relative e?ciency of banks in foreign markets look like.
Results by Claessens et al.(2001)[34] suggest, that the relative perfor-
mance of multinational banks compared to local banks in foreign markets is
to some degree host country speci?c. The authors, using a large sample of
7,900 bank observations from 80 countries, ?nd evidence that foreign banks
are more pro?table
15
than domestic banks in developing countries
16
. How-
ever, this ?nding is turned upside-down for developed countries. The latter
result has been seen as puzzling, as from a Bertrand competition-point of
view there should then be no scope for pro?table multinational bank entry
into these developed countries
17
.
14
The Glass-Steagall Act prohibited universal banking structures and was only repealed
in 1993.
15
The authors discuss interest margins, tax payments and general pro?tability as vari-
ables for pro?t-e?ciency of banks.
16
This result is replicated by a number of other empirical studies e.g. Bonin, Hasan and
Wachtel (2005)[25] and Majnoni, Shankar, Varheggyi (2003)[96].
17
However, this need not be a real puzzle. As the banking industry can be divided into
multiple heterogeneous service segments, one could argue that multinational bank entry
into developed countries is mostly restricted to segments of the industry, where pro?tability
19
Berger et al.(2000)[12] use a more advanced estimation strategy deriv-
ing banks’ X-e?ciency from a standard banking cost function
18
and further
break down the analysis to country levels within developed countries. They
focus on ?ve countries, France, Germany, Spain, UK and the United States.
The authors generally con?rm the result by Claessens et al.(2001)[34], that
domestic banks are more e?cient than foreign banks in developed countries,
with the notable exception being Spain.
Decomposition into pro?t and cost e?ciency gives a hint on where this
pro?t di?erences between domestic and foreign banks stem from. In all of
the analyzed countries the average domestic bank shows signi?cantly higher
pro?t e?ciency than the average foreign bank. However, in Spain, this is
more than o?set by a lower cost e?ciency of domestic banks. Cost e?ciency
in the U.S. is also lower for domestic banks (2,9% higher costs), however
this is vastly lower than the pro?t e?ciency advantage for these domestic
institutions (25,1% higher pro?ts). Therefore the authors argue, that the
cost disadvantage might not be due to ine?ciency but rather ”‘these high
expenses more likely re?ect e?orts to produce a quality or variety of ?nancial
services that generate substantially greater revenues”’
19
.
One interesting result concerns home country e?ects in the discussion of
relative e?ciency. Though, as laid out, foreign institutions are on average
less e?cient than domestic banks in most developed countries, this result
does not hold for multinational banks from the U.S.. Indeed, Berger et
al.(2000)[12] observe, that U.S. banks are more e?cient than domestic banks
in France, Germany and Spain, the only exception to this pattern being the
UK. It therefore seems that U.S. banks are so overwhelmingly more e?cient
in general, that they are able to even outperform domestic banks in their
own backyards in most countries.
is lower. Domestic banks might voluntarily leave these segments to foreign players, instead
growing their business in other ?elds of banking. Also, the empirical literature partly
su?ers from the problem, that the majority of multinational banks entered foreign markets
rather recently. Therefore, these banks might predominantly be in a phase of pursuing
aggressive growth strategies in respective foreign markets, asking for low interest spreads
and providing services at low prices to attract away customers from incumbents.
18
The speci?ed cost function uses four variable outputs (consumer loans, business loans,
real estate loans and securities), one ?xed output (o?-balance-sheet activity), two ?xed
inputs (physical capital, ?nancial equity capital) and three variable inputs (purchased
funds, core deposits and labor). This setup therefore at least partially controls for di?erent
business portfolios/strategies of the respective banks.
19
Berger et al.(2000), page 57
20
Concerning developing countries, Green, Murinde and Nikolov (2004)[71],
exploring economies of scale and scope achieved by domestic and foreign
banks in selected Central and Eastern European countries from 1995-1999,
?nd selective counter-evidence on whether foreign banks are generally more
e?cient than domestic banks in lesser developed countries. The authors do
not ?nd signi?cant di?erences in scale and scope e?ciency between domestic
and foreign banks in the analyzed region. However, their results should be
handled with care, as due to the nature of the transition markets, foreign
banks had just recently entered the market prior to the observation period,
probably still in the process of reaching their e?cient scale and scope in
respective markets.
Havrylchyk and Jurzyk (2006)[76] study the relative e?ciency of for-
eign banks in Central and Eastern Europe in the period 1993-2004. Using
BankScope data they measure pro?tability via return on assets (ROA). They
?nd that the mode of entry plays an important role in the relative e?ciency
of foreign banks. Whereas foreign bank subsidiaries resulting from the acqui-
sition of domestic incumbents do not signi?cantly di?er from other domestic
banks concerning pro?tability, foreign subsidiaries established via Green?eld
Investment are signi?cantly more pro?table than domestic banks
20
. How-
ever, decomposition of the pro?tability variable, shows that this di?erence
in pro?ts might not be due to ine?ciency but rather due to di?erent strate-
gies and business segments. Green?eld banks seem to focus on low-cost, low
risk business segments, which do yield lower risk and service premia, but
these premia seem to be too low in transition economies for the risk born via
non-performing loans and high overhead costs
21
.
The two authors also discover, that Green?eld foreign banks show sig-
ni?cantly lower deposit-to-asset rates than either ”‘takeover foreign banks”’
or domestic banks (60% versus 79/76%). Whether this is by choice, as the
former banks want to re?nance via other sources, or by their ineptitude to
raise su?cient volumes of deposits, e.g. due to a missing large scale branch
network, is at ?rst sight debatable. A lack of access to local deposits in gen-
eral might be one of the reasons for the relatively poor performance of foreign
banks in developed markets. DeYoung and Nolle (1996)[51] ?nd that foreign
banks in the US work under an ine?cient input mix, predominantly having
to rely on re?nancing via purchased funds, whereas local banks use domes-
tically raised deposits to a much higher degree. For the US, this even seems
20
”‘Green?eld foreign banks”’ show a mean return on asset ratio of 1.45, whereas do-
mestic and ”‘Takeover foreign banks”’ show mean ROA of 0.86 and 0.87 respectively.
21
High overhead costs might stem from focusing on loan projects that require a personal
gathering of information about the respective client.
21
to hold true for ”‘takeover foreign banks”’. Peek et al.(1999)[86] ?nd that
foreign banks in the US tend to acquire targets that show an above-average
reliance on purchased funds, resulting in a low deposit-to-asset ratio.
2.2.6 What are the e?ects of multinational bank entry
on the host country?
From a political point of view, the main question concerning multinational
banking is, which e?ects the entry of foreign banks into respective markets
has on economic conditions. With banks still being one of the main providers
of ?nance to the real sector, questions about the entry e?ects on domestic
banks’ performance and loan supply are heavily discussed.
The e?ects of multinational bank entry depend on the respective entrants’
(relative) e?ciency, the product/segment strategy they implement and what
kind of entry mode they choose.
One commonly accepted e?ect of entry into any oligopolistic market is,
that market power of incumbents diminishes and market volumes increase, if
the entrant is su?ciently e?cient to put competitive pressure on the incum-
bents. Due to competitive pressure, incumbents additionally might or might
not be incentivised to operate more e?ciently, either by bank managers being
pushed to operate the structure more e?ciently, or by increasing/decreasing
the need/incentive for incumbents to invest in e?ciency-enhancing new tech-
nologies and practices.
These e?ects should theoretically be costly for the domestic banking sec-
tor up front, but should at ?rst sight improve loan market conditions for
borrowers.
Systematically, the potential e?ects of bank entry can be decomposed
into the direct market e?ect via the introduction of an additional (Green?eld
Investment) or at least the e?cient restructuring of an existing local bank (ac-
quisition), an indirect e?ect induced by domestic incumbents adapting their
strategies in this case, and some direct spillovers of banks technology and
institutional requirements on domestic banks’ and domestic regulatory/legal
institutions.
An increasing theoretical literature has provided both a basis for further
(and ex post explanation for existing) empirical analysis.
Concerning the e?ect of entry on domestic banking sector e?ciency,
Lehner and Schnitzer (2006)[93] in a setup of competition in horizontally
di?erentiated products, discuss the direct e?ects of increased competition
and direct spillovers in screening technology on the e?ciency of the domestic
22
banking sector, where the two direct e?ects have the negative side-e?ect of
reducing incumbent banks’ incentives to invest in screening themselves. Dif-
fering between e?ects from entry via Green?eld Investment and acquisition,
the authors ?nd, that for the case of weakly competitive market (high product
di?erentiation) tendentially entry via acquisition is relatively harmful in less
developed countries, whereas it is less harmful in developed countries. For
the case of low product di?erentiation Green?eld Investment is the welfare
maximizing mode of entry in all kinds of countries.
The e?ect of foreign bank entry on incumbent banks’ lending practices
is theoretically analyzed by Dell’Ariccia and Marquez (2004)[48], who de-
velop a model capturing information asymmetries in loan markets. One of
their main results is that entry by an ”‘uninformed”’ outsider (multinational
bank) should lead to incumbent banks shifting their loan portfolio towards
more informationally-opaque borrowers, due to increased competition in the
segment of non-opaque potential clients. From a general perspective this
could be seen as a bene?cial indirect e?ect of entry, if the opaque segment
had previously been ine?ciently loan-constrained by incumbent banks.
Claeys and Hainz (2006)[35], building on the above setup, develop a model
to discuss di?ering e?ects of foreign bank entry via Green?eld and acquisition
on bank lending rates in the respective host market. They ?nd that foreign
bank entry reduces required lending rates by incumbent domestic banks,
more so if the majority of entry is via Green?eld Investment.
Proceeding to test their hypothesis empirically, Claeys and Hainz (2006)[35]
indeed ?nd, that a higher foreign bank share in loans reduces average lending
rates in the respective market. Also they show evidence, that ”‘Green?eld
foreign banks”’ charge higher lending rates than ”‘takeover foreign banks”’.
Concerning di?ering competition e?ects of entry via Green?eld or acquisi-
tion, the former is observed to reduce average domestic bank lending rates
signi?cantly more than the latter.
Havrylchyk and Jurzyk (2006)[76] also discuss the di?ering e?ect of dif-
ferent entry modes on domestic banks’ performance. They discover that, no
matter the entry mode, a higher market share of foreign banking increases
costs for domestic banks. The authors note that this might seem to be coun-
terintuitive at ?rst sight, as one would expect higher competition to lead to
higher cost e?ciency, however they argue that this cost increase might be
short-term, due to domestic banks’ arising need for investing into competitive
risk, IT and management systems. In contrast to a high share of ”‘takeover
foreign banks”’, a large market share of ”‘Green?eld foreign banks”’ addi-
tionally decreases domestic banks’ non-interest income but also their loan
loss provision volume. In total, however, the e?ects seem to balance out each
23
other, as domestic banks’ pro?ts seem to be una?ected by the foreign banks’
market share in the respective market.
Other empirical studies ?nd that entry of foreign banks reduces the prof-
itability of domestic banks. Additionally to the result from Claeys and Hainz
(2006)[35] e.g. Claessens et al.(2001) [34] ?nd that this reduction in prof-
itability is mainly due to reduced net interest margins, a hint that the e?ect
on pro?tability indeed works through a reduction in market power of incum-
bent banks in the provision of loans and/or deposit taking.
The latter authors ?nding can also be perfectly related to the miss-
ing ”‘foreign bank market share-e?ect”’ found by Havrylchyk and Jurzyk
(2006)[76]. Indeed Claessens et al.(2001)[34] ?nd that the market share of
foreign banks is not an important determinant of domestic banks pro?tabil-
ity, but rather only the number of foreign banks has the expected negative
signi?cant e?ect, suggesting that the threat of foreign banks taking market
shares away already induces domestic banks to reduce net interest spreads,
allowing them to keep domestic bank market shares at a high level.
Another accommodating result is found by Levine (2003)[94], who uses
a unique data set on cases of regulatory institutions denying foreign bank
entry in 47 developed and less-developed countries to analyze the relation-
ship between political entry restrictions for foreign banks and bank interest
margins in a respective country. The author shows evidence, that the more
restrictive entry regulation for foreign banks in a country, the higher the net
interest margins for banks in the market. Foreign bank entry also is found to
be special, as restricting domestic bank entry does not alter operating banks
net interest margins. The results further con?rm the hint from Claessens
et al.(2001)[34] that the contestability of a respective host market primarily
determines the competitive behaviour of operating banks, not the actually
incurring amount of entry into the market. It seems that the existence of
multinational banks threatening to enter a market already disciplines incum-
bent banks.
One important point of note concerning the e?ects of foreign bank entry
is, that such entry from a global view went hand in hand with a consolidation
of the international banking industry. The chapter on bank sector consolida-
tion in this thesis discusses the e?ects of general consolidation in the banking
industry in detail. In that chapter the potential e?ect of multinational bank
entry on the availability of loans to a speci?c segment of the market, infor-
mationally opaque small and medium-sized enterprises (SMEs), is discussed
in detail.
Generally, foreign bank presence is found to increase access to loans, at
least for larger and transparent ?rms in emerging market economies (see e.g.
Mian (2006)[100], Giannetti and Ongena (2005)[66], Clarke et al.(2001)[37]).
24
Concerning informationally opaque smaller ?rms, e.g. Mian (2006)[100] how-
ever ?nds that
‘greater cultural and geographical distance between a foreign bank’s head-
quarters and local branches leads it to further avoid lending to ”information-
ally di?cult” yet fundamentally sound ?rms requiring relational contract-
ing”’
suggesting that the bene?cial e?ect of foreign bank entry might not extend
to this segment of the market. This results is independent of bank size, bank
risk preferences or legal institutions in the home country. This biased lending
strategy e?ect of distance also is identi?ed to be large enough to completely
exclude some types of borrowers in the economy from foreign bank ?nance.
Concerning the increased availability of bank ?nancing for at least a
fraction of borrowers in the host market through MNB entry, it at ?rst
sight seems as this is simply due to an underlying capital transfer into the
host country via the foreign banks’ internal capital market. However, as
Havrylchyk and Jurzyk (2006)[76] discovered foreign banks to predominantly
re?nance their local lending activity with local deposits, the international
capital transfer through entering banks seems less pronounced than origi-
nally thought. Especially for banks entering via acquisition the deposit-to-
asset ratio of 79% suggests minimal cross-border capital transfers within the
banks’ activity. However, if entering foreign banks are more e?cient screen-
ing potential borrowers than existing incumbents the availability of deposits
for commercial loan ?nancing might rise in total, as owners of liquid assets
might be more willing to extend deposits to the banking system (instead of
e.g. transferring assets abroad or invest in government bonds).
Finally, one topic that has also received a substantial amount of interest
in the economic literature is how the presence of foreign banks a?ect the
general stability of the host country banking sector and ?nancial market.
The empirical studies to date focus on the two main recipient less-developed
regions, Latin America and Central and Eastern Europe (CEE).
Theoretically, two speci?c advantages of multinational banks compared to
local domestic banks might determine whether credit supply is stabilized by
foreign banks or not. On the one hand, a multinational bank might have ad-
ditional/cheaper sources of re?nancing, such as better access to international
capital markets or to deposits in the home country or other countries of op-
erations. Therefore these banks might be able to support their subsidiaries
in cases of market-speci?c liquidity/bank crisis, therefore dampening local
shocks. On the other hand, due to operating in multiple markets, the bank
might transfer liquidity/assets from badly-performing/crisis-plagued coun-
25
tries to other regions, therefore strengthening local shocks.
Supporting a subsidiary should generally be a question of a parent bank’s
capability to do so, whereas substituting e?ects should be seen as an incentive-
based decision within in the bank structure, depending on banks’ opportunity
costs of keeping capital in a respective host country subsidiary.
Due to multinational banks operating in multiple regions, they might
therefore transfer shocks from one region to the other via reallocation of
assets within their international internal capital market.
Recent theoretical work by Morgan et al.(2004)[103], who develop a model
of a multinational bank (along the general banking model of Holmstr¨om and
Tirole (1997)[78]), that rebalances its international bank capital reacting to
shocks in bank and real sector capital in respective regions, show what kind of
country-speci?c shocks might be dampened or strengthened by the presence
of multinational banks. The model predicts that multinational banks dampen
local bank-capital shocks by supporting their local subsidiaries but increase
the volatility of the business sector by reallocating bank capital away from
regions experiencing real sector capital shocks.
Empirically, for Latin American countries, Dages et al.(2000)[43], Peek
and Rosengren (2000)[112], Goldberg (2001)[69], Crystal et al.(2002)[41] and
other studies, ?nd strong evidence that foreign bank presence increases the
stability of domestic banking sectors. Foreign banks seem to show stronger
and less volatile credit growth, and positive growth even in periods of ?-
nancial market crisis
22
. However, this line of reasoning does not seem to be
speci?c to the ownership structure of banks, but rather to the underlying rel-
ative ?nancial health of foreign banks in this region. Dages et al.(2000)[43]
observe that domestic banks show the same involatile behaviour as foreign
banks, if they are characterized by similar health e.g. similar levels of shares
of non-performing loans in their portfolio.
For the Central and Eastern European region, de Haas and van Lelyveld
(2006)[44] reinforce the notion of foreign banks as a stabilizing force as they
are found to keep up lending volume during times of ?nancial distress whereas
domestic banks strongly contract lending. However, this seems to hold true
for ”‘Green?eld foreign banks”’
23
only.
22
Indeed, it seems that foreign banks see these occasions as opportunities to expand
their market shares at the expense of ?nancially stricken local banks.
23
Again this is in line with the results of Havrylchyk and Jurzyk (2006)[76], which show
that Green?eld banks are much more embedded in a multinational bank group’s internal
capital market, whereas acquired banks within the group seem to be organized as rather
independent capital centers.
26
Home market and parent bank e?ects, a?ecting the whole bank group via
internal capital markets, also signi?cantly in?uence the lending activity of
multinational banks abroad.
De Haas and van Lelyveld (2006)[44] ?nd evidence for both a substitution
and a support e?ect
24
, depending on parent bank and home market condi-
tions. Concerning the former, they ?nd that foreign banks reduce credit
supply in foreign markets if GDP growth in the home country accelerates,
leading to more potential value-adding business in the home country market.
Again, this substitution e?ect only applies to ”‘Green?eld foreign banks”’.
The capability to support a subsidiary should critically hinge on the parent
bank’s ?nancial status. Indeed, the authors ?nd that parent banks showing
strong ?nancial health
25
have subsidiaries in the CEE region growing credit
volume faster than subsidiaries of weak parent banks. The latter result does
hold for all kinds of foreign bank subsidiaries, however the e?ect is more
pronounced for Green?eld operations.
The majority of empirical studies on home market e?ects con?rm the
results concerning the substitution hypothesis, as worsening home country
conditions seem to have led banks to enlarge their lending activity in foreign
markets (e.g. Moshirian (2001)[104] and Calvo et al.(1993)[30]).
Summing up, concerning the e?ect of multinational bank presence on the
stability of local banking markets, such presence seems to increase stability in
the analyzed less developed regions during times of ?nancial distress. How-
ever, there is strong evidence that, while foreign bank presence might dampen
?nancial crisis in the respective host countries, it could also strengthen or
even import these when home or third country market developments lead the
bank to substitute business and assets from the foreign to its home market
or third countries. The advantages and disadvantages of multinational bank
presence strongly depend on the level of integration of the respective local
subsidiary in the bank group’s international internal capital market. Both
disadvantages and advantages seem to be less pronounced for foreign sub-
sidiaries established via the acquisition of local incumbent banks, as these
structures seem to be more or less ?nancially segregated from their parent
banks.
After discussing what existing literature has to say about multinational
24
The following results are also con?rmed in a newer study by the same authors (de
Haas and van Lelyveld (2006b)[45].
25
The authors use the ratio of loan loss provision over net interest revenue as a proxy
for ?nancial health.
27
banking, the chapter now turns to a practitioner-oriented view, discussing
the behaviour of multinational banks with the help of a case study of one of
the success stories of multinational banking, Bank Austria Creditanstalt and
its successful expansion into Central and Eastern European markets.
28
2.3 A case study of successful multination-
alization: Bank Austria Creditanstalt in
Central and Eastern Europe
2.3.1 Overview of objects and sources of the case study
One of the hot spots of the evolution of multinational banks in recent years
have been the former Communist countries in Central and Eastern Europe.
Nowadays, concerning the scope of internationalization of their banking sys-
tem, these countries are special in that their banking system is dominated by
foreign bank subsidiaries. For example Bol, de Haan, Scholtens and de Haas
(2002)[24],for the year 2000, ?nd foreign bank asset share in total banking
assets to be 54% in Central Eastern Europe and 87% and 77% in South East-
ern Europe and the Baltic States respectively. This dominance holds true in
large countries like Poland (69%) and the Czech Republic (66%).
Within this environment one other fact of note is that the leading home
country of multinational banks operating in these transition economies is
Austria, with the regional market leader in this region being Bank Austria
Creditanstalt (BA).
These two facts rationalize choosing the respective bank in the respec-
tive region for a case study on Bank Austria’s operations in these (former)
transition countries.
The following insights have been won by screening annual reports and
presentations of Bank Austria and its Austrian competitor Erste Bank, but
the major insights were won by interaction with Mr Gerhard Smoley, Head
of Investor Relations of the Bank Austria Group at this time
26
.
I proceed as follows. First a general overview about Bank Austria Cred-
itanstalt (BA) and its history is given. In the following the focus is on the
bank’s operations in the Central and Eastern European (CEE) region, giving
an overview of the development of the bank’s market position. A discussion
of BA’s entry motives, modes and market strategy follow, yielding insights
into how, from a practitioner’s view, home country, host country and bank-
speci?c characteristics determined the internationalization strategy of Bank
26
The following insights strongly base on telephone interviews[109] as well as additional
email communication.
29
Austria. After discussing the future strategic focus of Bank Austria Credi-
tanstalt in the region, an overview of operations of Erste Bank, BA’s main
competitor in the CEE region concerning size, is given, allowing some in-
teresting comparisons between two very di?erent, yet both highly successful
market entry strategies. The case study ?nally is concluded by deriving gen-
eral insights into topics in multinational banking, that can be won from the
proposed case study.
2.3.2 Bank Austria Creditanstalt: An Overview
Bank Austria Creditanstalt today is the leading bank in its original home
country Austria (1.8 million clients at a country population of 8 million) and
what it now deems as its ”‘second home market”’[40] Central and Eastern
Europe
27
. Since November 2005 Bank Austria Creditanstalt, formally a part
of the HypoVereinsbank (DE) group, has become a member of the UniCredit
(IT) banking group via the latter acquiring the former. 95% of Bank Austria
shares are now held by UniCredit with 5% in free ?oat.
Concerning historical roots, Bank Austria was founded in 1991 via the
merger of ”‘Oesterreichische Landesbank”’, ”‘Zentralsparkasse”’ and ”‘Kom-
merzialbank”’. In 1997 Bank Austria took over the Austrian government’s
shares in ”‘Creditanstalt”’ fully integrating into the Bank Austria Credi-
tanstalt group in 1999. As ”‘Creditanstalt”’ was privatized by the Austrian
government only in 1990, all formerly independent parts of the newly arising
bank had formally been state-owned.
In the year 2000 Bank Austria Creditanstalt merged with HypoVereins-
bank and became the competence center for CEE business of the group.
BA fully concentrated on the Austrian as well as 11 selected Eastern Eu-
ropean countries, taking over HVB business in these countries while trans-
ferring other international business to the HVB organization. Bank Austria
embraced this friendly merger, as management saw Bank Austria’s further
growth possibilities limited due to the fact, that Bank Austria stand-alone
had grown too large for its home market but was too small to establish
strong operations in an integrating European banking market
28
. After the
HVB-UniCredit merger in 2005, Bank Austria became part of UniCredit
27
In Bank Austria Group de?nition this also includes CIS countries as well as Turkey.
28
Additionally Bank Austria had become too small for the risk it carried in its portfolio.
Especially its exposure to risks in the North American market was too large for further
growth stand-alone.
30
group structure and now acts as the primary holding and operations center
for the group’s Central and Eastern European business[40].
2.3.3 Bank Austria Creditanstalt in Eastern Europe
Bank Austria Creditanstalt was the ?rst mover in the Central and Eastern
European states, especially with Creditanstalt already having had established
agencies in Hungary (1975), Prague and Moscow (both 1987) even before the
fall of the iron curtain. Creditanstalt also was the ?rst foreign bank to take
over a domestic target in the CEE states, acquiring a majority stake in the
Slovenian ”‘Nova banka”’ as early as 1992[40].
In general, between 1989 and 1991 both Bank Austria and Creditanstalt
started to enter markets in Central and Eastern Europe on a large scale.
Nowadays Bank Austria Creditanstalt is by far the market leader, not
only among foreign but all banks, in Central and Eastern Europe, with to-
tal assets of 41bn and 39.000 employees in 1,800 branches serving roughly
18 million clients in the region. As the CEE competence center of Uni-
Credit group, this includes former subsidiaries of UniCredit and HypoVere-
insbank.Concerning assets, BA-CA-group subsidiaries are the No.1 bank in
Croatia, Bulgaria and Bosnia-Herzegowina, and are among the ?ve largest
banks in eight countries in total[39].
Table 1 gives a detailed look at the group’s positioning in the respective
countries
29
.
29
data as reported 31.12.2006
31
Table 2.1: BA-CA Operations in Central and Eastern Europe
Country Subsidiaries Total
Assets
(€bn)
Branches Market
position
Bosnia • HVB Central Profit Banka
• Nova Banjalucka Banka
• UniCredit Zagrebacka Banka
1.7 180 No. 1
Bulgaria • UniCredit Bulbank 4.2 300+ No. 1
Croatia • Zagrebacka banka 10 127 No. 1
Czech
Republic
• HVB Bank Czech Republic
• Zivnostenska Banka
9.2 79 Top 5
Estonia • HVB Bank Talinn 0.07 NA Top 10
Hungary • UniCredit Bank 5.3 76 Top 10
Lithuania • HVB Bank Vilnius 0.38 2 Top 10
Macedonia Representative Office only
Montenegro Representative Office only
Poland • Bank BPH
• Bank Pekao
33 1,292 No. 1
Romania • UniCredit Tiriac Bank 3.7 130 Top 5
Russia • International Moscow Bank
• Yapi Kredi Moscow
6.6 NA Top 10
Serbia • UniCredit Bank 0.86 46 Top 5
Slovak
Republic
• UniCredit Bank 3.6 93 Top 5
Slovenia • Bank Austria Creditanstalt
Ljubljana
2.2 14 Top 5
(Turkey) • Yapi Credi 29.5 653 Top 5
Ukraine • HVB Bank Ukraine
• UniCredit Bank
0.25 6 13
Source: Bank Austria Creditanstalt Corporate websitehttp://www.bankaustria.at
2.3.4 Entry motives, entry modes and market strategy
of Bank Austria Creditanstalt in CEE
The following section draws most of its information from communication with
Bank Austria group head of investor relations Mr Gerhard Smoley.
Three headline motives for BA’s entry in CEE countries became apparent
in discussions. I describe these hand in hand with what seems like underlying
or complementing characteristics of Bank Austria.
Home country push factors seem to have played a signi?cant role in
Bank Austria’s decision to expand internationally. The Austrian market
32
was severely overbanked in the early 1990s and no further signi?cant growth
in the home market was deemed feasible. In this surrounding, Bank Austria,
a recently privatized group of formerly state-owned bank, had to grow its
pro?ts fast after privatization to reach su?cient levels of shareholder value.
Therefore BA was one of the banks most committed to ?nding pro?table
business opportunities. The political development in Eastern Europe pre-
sented an unexpected window of opportunity for this.
‘By chance the possibility for feasible international expansion arose by
the fall of the iron curtain.”’(Smoley, 2007)(Smoley (2007)[109])
Host country factors in the beginning of the geographical expansion were
more restricted to indirect follow your customer considerations. The ?rst
clients in these markets were Austrian commercial clients of BA entering the
respective markets. According to Mr Smoley this was the main motive or
building stone in the beginning. ”‘If not for so many of BA’s customers from
Austria entering the CEE markets, Bank Austria would most probably have
entered these markets signi?cantly later in the process”’(Smoley (2007)[109].
Additionally the CEE markets turned out to be in large need for stable banks
(Bank Austria back then o?ered an AAA rating).
Also, Bank Austria, almost from the beginning, did not only service Aus-
trian clients in Eastern Europe, but also a variety of multinational ?rms from
other home countries, such as McDonalds and VW-Skoda. The bank might
have had an additional home country advantage in the latter business, as
Vienna functions as a management platform for a number of multinational
?rms operating in the CEE region. For example, McDonalds Eastern Euro-
pean operations center is located in Vienna, therefore close personal contact
between Bank Austria top management and McDonalds’ CEE management
was available quite easy.
A helpful ”‘snowball e?ect”’, stemming from business in the CEE re-
gion with such multinational customers, partly materialized, as these multi-
national customers often used local suppliers in the value chain. Getting
into business with these local ?rms was made signi?cantly easier by this
link[109]
30
.
The general identi?cation strategy of Bank Austria concerning attractive
host country markets, which was applied for location choice in the latter
process of transition, was based on basic macroeconomic indicators, such
as GDP, GDP growth and population. Additionally BA analyzed potential
cross-market synergies, trying to identify markets, where entry would also
30
This might be another avenue for research as one could discuss how the possibility of
follow your customer-motivated entry into foreign markets supports local market-seeking
operations later on.
33
leverage opportunities in BA’s global business network[109].
Compared to banks from other countries, Bank Austria might have had
an advantage concerning ”‘appetite for entry”’ by an indirect home market
e?ect. Bank Austria stems from a relatively small home market with a
population of 8 million. Therefore the bank was less reluctant to enter, what
banks from other home countries like Germany might have perceived as rather
too small, markets in CEE, as from the perspective of a small country bank
the market seemed to be su?ciently large[109]. Additionally the bank had
experience how to pro?tably operate in a small market. These two points
might also explain the fact, that Austrian banks in general have been heavily
involved in the CEE region already early in the transition process.
Another reason why banks from the small host country Austria were
?rst-movers in the CEE region, outpacing German banks who, according
to other theoretical considerations, such as the availability of large deposit
volumes and the e?ect of strong competition in large markets on e?ciency
of market participants, should have bene?ted from their larger home country
market while also being close geographically, might have been the fact, that
German banks, in the early period of transition, very much concentrated
on growing their business in the former GDR. This focus possibly restricted
organizational capabilities left to explore the CEE markets.
The early geographic pattern of Bank Austria’s expansion into CEE mar-
kets also was closely linked to the follow your customer-nature of its early
FDI projects. Bank Austria ?rst expanded into markets close to Vienna,
namely Prague in the Czech Republic, Bratislava in the Slovak Republic and
Budapest in Hungary. Whereas in the economic literature (e.g. Buch and
de Long (2004)[28]) the in?uence of distance on bank FDI is predominantly
perceived to base on an intuition of information costs, Mr Smoley stated,
that, in the case of Bank Austria, distance was simply negatively correlated
with the amount of activity of potential real sector customers to be followed
abroad[109].
Additionally though, the transition economies closest to the EU-15 also
pro?ted most from the fall of the iron curtain, experiencing faster economic
growth than more distant countries. According to the practitioner, and prob-
ably due to considerations of eastern enlargement of the EU, a more rapid
increase in institutional quality also was evident in these countries. There-
fore they were also the most interesting markets from a pure foreign market-
seeking perspective in the region.
Digging deeper into the expansion strategy of Bank Austria, to under-
stand the entry modes preferred by the bank one has to take the general
34
strategic setup of Bank Austria in the 1990s as well as the changing land-
scape in the ?nancial sector in CEE into consideration.
Both Bank Austria and Creditanstalt started to enter the CEE markets
in an extensive way, ?rst setting up small agencies in the countries’ capi-
tals. Bank Austria’s general strategy was to enter the market via Green?eld
investment. Usually its ?rst step was to enter via its investment banking
division, which acted as a path?nder to screen respective market character-
istics and the legal environment. As laid out before, primary early clients
for BA in the region were Austrian corporate customers and other multina-
tional ?rms. In general Bank Austria, in Austria as well as in CEE, heavily
focused on the commercial wholesale segment, at least in the 1990s. Due to
this strategic positioning, a large branch network to attract retail and SME
customers and soft information about these informationally opaque poten-
tial clients were not mandatory for successful operations in these markets.
Indeed, for conducting wholesale business, it is often su?cient to establish a
handful of regional headquarters in the big cities[109]. This latter argument
might make clear why Bank Austria, compared to other multinational banks
in CEE, was less active in the takeover market for local banks.
However, as margin pressure in commercial banking increased in Eastern
Europe, due to the development of capital markets and an increasing number
of foreign, e?cient banks operating in the region in this segment, Bank Aus-
tria added retail banking as an additional focus for growth in the region[109].
This rededication to retail banking in the region was further motivated by
the rapidly growing demand for services such as asset management and tra-
ditional retail banking arising through the economic development of these
countries.
One of the most attractive, due to being the largest, markets for retail
banking in Central and Eastern Europe is Poland. Bank Austria took a share
in the Polish bank PBK in 1997, becoming majority shareholder by 2000
31
.
However, Bank Austria was not able to grow its business via acquisi-
tions at this time in many CEE countries strongly. The reasoning shows how
strategic choice in one market is in?uenced by an institutions’ business devel-
opment in other markets. Bank Austria was one of the very few international
banks heavily involved in Russia when the Ruble crisis materialized in 1998.
The incurred losses of Bank Austria in its Russian business restricted the
possibility for growing its business in other parts of Eastern Europe for the
following years, due to a lack of ?nancial power. BA was not able to partic-
31
Polish regulation at that time was such that only 10% stakes in a domestic bank could
be bought per round. Therefore reaching a majority share in a Polish bank could not have
been attained immediately.
35
ipate full scale in the bank privatization rounds in Eastern Europe in 1999
and 2000[109]. Its main Austrian competitor in Eastern Europe size-wise,
Erste Bank AG, took advantage of this strategic weakness very actively tak-
ing part in takeover/privatization markets in Eastern Europe. Erste Bank
AG also was less reluctant to pay high strategic prices for banks especially
o?ering su?cient retail and SME client contacts, as these business segments
also happen to be the strategic focus of Erste Bank in its Austrian home
market.
So in general, Bank Austria was relatively weakly involved in signi?cant
acquisitions in Eastern Europe. BA tried to grow its retail business via
acquisitions in the Czech Republic, Slovak Republic and Hungary, but failed
to do so, due to a regional takeover market ever more characterized by bidding
wars on attractive targets. From an ex post point of view the bank’s forced
low acquisition activity in the early years however seems to have been a
blessing in disguise, as Bank Austria was one of the foreign banks in the
region not negatively a?ected in their growth possibilities by having to take
on a large share of non-performing loans from an acquired local bank[109]
32
.
An additional explanation for the bank’s relatively low M&A activity,
according to Mr Smoley[109], was that Bank Austria had no punctual geo-
graphic expansion strategy, therefore being less dependent on speci?c acqui-
sition, rather focusing on entering the Eastern European market on a broad
regional basis. With this implemented strategy Bank Austria has also been
able to gain signi?cant economies of scale on the regional basis via synergies
both on the pro?t as on the cost side. On the cost side, Bank Austria Cred-
itanstalt was able to build up centralized transaction centers for the whole
region and also introduced a, though partially adapted to local needs, com-
mon IT infrastructure. Additional economies of scale and scope according to
the practitioner are available on the revenue side[109]. As the CEE region
itself strongly integrates trade-wise, with intra-region trade becoming ever
more important, as well as an ongoing legal integration due to the adoption
of common EU law, two potentially pro?t-enhancing opportunities arise for
a bank with a complete network across the whole region. For one, the trade
32
Indeed with this ”‘Green?eld”’ strategy Bank Austria itself was surprised with the
low ex post risks in their loan portfolio. Ex post risks for some time now have been lower
in BA Eastern European operations than in Austrian operations, which was a completely
unexpected development for the bank. One advantage according to the practitioner is
social culture in the CEE states, in the respect that e.g. in Poland not paying back a
loan is still seen as a personal shame, leading to a perceived lack of moral hazard in loan
provision.
36
integration increases the need of local ?rms for banking services in multiple
countries in the region, favouring a bank with a complete network. Concern-
ing the legal integration, the possibility for selling homogeneous products in
the whole region increases, yielding cost-saving and reputation-spillover pos-
sibilities for Bank Austria. These available economies suggest, that banks,
analyzing single country markets stand-alone (and deeming them too small),
might miss an important point. As Mr Smoley stated
‘Bank Austria is the No.5 bank in Slovenia, a country with a population
of 2 million. Therefore looking at the market stand-alone one could ques-
tion whether signi?cant value-added can be achieved by entering this market.
However being present there creates network e?ects and value-added for com-
mercial customers in Poland, Czech Republic, etc.. One has to take this into
account when thinking about entering a market.”’(Smoley, 2007[109])
Bank Austria also indirectly grew by acquisition in Eastern Europe, tak-
ing over Creditanstalt in 1997, who itself already had a signi?cant physical
presence in numerous market in this region, such as for example the Slovenian
”‘Nova banka”’ Creditanstalt had acquired back in 1992. One of the success
factors in this integration was that brands stayed independent for ?ve years
to keep goodwill immanent in both brands in these countries, while at the
same time exploring economies of scale centralizing back o?ce operations for
both brands.
Additionally Bank Austria’s CEE operations grew via two other M&A
deals Bank Austria was part of. As discussed before Bank Austria today
bundles all former CEE operations of HypoVereinsbank and UniCredit
33
.
For Bank Austria Creditanstalt the merger with HypoVereinsbank brought
a number of signi?cant advantages for CEE operations with it. BA was able
to signi?cantly strengthen its network in Eastern Europe. The HVB opera-
tions taken over were further successfully leveraged, as Bank Austria already
was a more established brand name in Eastern Europe. BA was also able to
manage existing HVB operations more e?ciently due to its greater experi-
ence and vaster market knowledge in CEE markets
34
. One clear-cut synergy
was that HVB had already also taken over a Polish domestic bank. As the
respective Polish subsidiaries were merged the newly created subsidiary now
33
Except one polish subsidiary bank which became part of the UniCredit organization
due to political restrictions by the Polish government.
34
The former claim can be backed by the fact that for a while HVB changed the name
of subsidiaries of the group in the region to HVB (country). The brand name was not
as well known as Bank Austria and stunted business growth for some time in the CEE
region.
37
became the third-largest bank in the Polish market and even more impor-
tant, the lone universal bank with a strong retail business in Bank Austria’s
Eastern European portfolio[109]. As stated before, this was of signi?cant
interest especially in Poland. The increased size of the group’s Eastern Eu-
ropean operations also allowed the bank to take on ?xed costs risks of further
expansion into the retail sector.
As Bank Austria tried to grow its retail business, it grew its branch net-
work in Poland, Romania, Croatia and Hungary via Green?eld Investment.
As Hungarian target banks became too expensive for takeover, BA tried to
grow organically on a large scale, establishing 100 new Green?eld branches
in short time.
Additionally, in recent years, Bank Austria became more active in the
takeover market, acquiring banks ”‘Splitska banka”’ in Croatia and ”‘Biochim”’
in Bulgaria in 2002 as well as ”‘Central Pro?t Banka”’ in Bosnia in 2003.
In the retail sector ”‘Bank Austria started out with Green?eld Investment
and then tried to speed up the growth process via acquisitions, even though
acquisition prices grew higher, too.”’(Smoley, 2007[109]. Concerning target
strategy, Bank Austria, contrary to other banks in the region, focused on
acquiring healthy local banks and paying the high prices for these, whereas
other entrants sought a low acquisition price ?rst and foremost, banking on
being able to restructure badly-performing banks burdened with a high share
of non-performing loans[109].
Based on this discussion, one may be able to derive that in general full-
scale retail banking services can probably only be provided in a market en-
tered via M&A. In almost all countries BA was not involved in takeovers it
still focuses on the commercial clients segment[109].
The underlying reasons can be identi?ed by evaluating the banking mar-
ket in the Czech Republic. Bank Austria there has limited scope winning
retail clients away from other banks due to a lack of network size. Whereas,
for example, Erste Bank AG subsidiary Ceska sporitelna has 630 branches
in this country, Bank Austria only operates 40. So due to missing local
structures, there is no real possibility to attract retail customers as well as
SME customers in some regions of the Czech Republic. Still, Bank Austria
in terms of value of assets is the fourth largest bank in CZ concerning assets,
due to its leading position in the ?eld of industry and trade (commercial)
clients[109].
Strong existing bank-client relationships and therefore a general lack of
customer mobility in the retail segment are reasons why Bank Austria is re-
luctant to build up large Green?eld branch structures in most CEE countries,
even more so, as BA tries to position itself as a quality-, not price-leader,
where the former strength can hardly be marketed to potential customers
38
locked in an existing relationship with another bank.
Besides engaging in retail and commercial banking, Bank Austria has a
very well-positioned investment banking division in the CEE region. Due
to having been one of the ?rst-movers in the markets, BA has excellent
long-lasting relationships especially with government agencies in Central and
Eastern Europe, also evidenced by the bank being awarded a price for ”‘Best
Investment Bank in CEE”’ by ”‘Financing New Europe”’. In this ?eld BA
also still pro?ts from the fact, that most large investment banks early on
deemed this regional market to be too small to put a strategic focus on it.
According to Mr Smoley these relationships to local customers that are in
place now are more important for successful investment banking than strong
relationships to the capital market where products are placed
35
. Now being
part of UniCredit group also strengthens BA’s position in the latter respect,
such that BA expects to keep its market leadership in CEE in this area of
business[109].
One ?nal interesting point concerning Bank Austria Creditanstalt’s strat-
egy is the mode of re?nancing of operations in Central and Eastern Europe.
Start-up investments and acquisition prices were ?nanced by the Austrian
parent bank. However concerning the re?nancing of operating business such
as loan provision, the pecking order is local ?rst, global only if needed. Sub-
sidiaries should ?rst and foremost re?nance their operations via raising local
deposits. BA tries to re?nance all of its local loans in the respective local
market. However the degree of local re?nancing di?ers among countries (be-
tween 100% and 80%). The larger the subsidiary concerning the number of
branches, the higher tendentially the percentage of local re?nancing
36
. Early
on in the transition process. as loan demand exceeded wealth, the picture
percentage-wise looked di?erent, with the bulk of loans by CEE subsidiaries
re?nanced by the parent in Austria. One important part of local deposits for
BA, as the group’s main strength is still with commercial clients, are large
deposits from its commercial clients, who hold liquid assets in local currency
in a substantial amount[109].
If a subsidiary comes short of complete self-re?nancing, Bank Austria
operates an internal capital market structure trying to optimize capital across
the group. Usually the subsidiaries should not re?nance themselves via the
capital market themselves, as Bank Austria has better re?nancing conditions
35
However placement power has to be su?ciently large to keep the trust of customers
in the bank being actually able to secure ?nancing.
36
This again points at the need for a large branch network to attract a large volume of
retail customers.
39
there due to a better rating[109].
The following reasoning can be put forward for such an implemented
pecking order of re?nancing. First, local deposits should be the main and
?rst source of re?nancing, due to a complete lack of currency risks
37
when
re?nancing that way, as well as an increasing distaste of stock analysts for
intra-group cross-subsidiaziation[109].
If local deposits are however not available to a su?cient degree, the in-
terest rate advantage of Bank Austria Creditanstalt over its subsidiaries in
capital markets should be made use of.
2.3.5 The future strategy of Bank Austria Creditanstalt
in CEE
With Bank Austria Creditanstalt now a part of UniCredit group, the geo-
graphic focus of BA’s business is shifting further to the east. Bank Austria
has recently entered markets in Russia and Turkey and is in the process of
doing so with a large commitment in Kazahstan and the Ukraine. In Rus-
sia, Bank Austria group is present via the International Moscow Bank. In
Turkey, BA holds a 50% share in the Yapi Credi Bank. In Kazahstan and
the Ukraine Bank Austria is in the process of taking over ATF Bank and
Ukrsotsbank respectively[40].
This further eastern expansion is driven by market dynamics in the CEE
regions already serviced as well as by a perceived comparative advantage of
Bank Austria over other competitors in the new markets.
Large CEE markets have already become quite consolidated and further
signi?cant growth is hard to achieve there. One prime example is the Czech
market. The ?rst three market positions are ?rmly established there, espe-
cially in the retail banking sector characterized by a low customer switching
rate. The market is more or less divided in stable market shares, and BA is
restricted to keep operating within its niche strategy there. While organic
growth can hardly be achieved, pro?table growth via acquisitions is also not
feasible as attractively priced and available targets are missing[109].
In contrast, vast pro?table growth potential in Russia and Turkey exists,
especially due to the fact that a lot of foreign banks are still not present in
37
Note that additional to the direct costs of currency risks, Basel 2 guidelines force
banks to either completely costly hedge these risks or build up legal reserves for them,
which is also very costly as the respective amount of capital can then not be used in more
pro?table opportunities.
40
these markets due to perceived political risk. There is high pro?t potential
in the retail segment as well, as especially in Turkey demographic trends lead
to an arising interesting market for ?nancial products[109].
According to Bank Austria, two comparative advantages of the bank over
competitors exist concerning these markets. For one, Bank Austria is one of
the very few banks that have build up vast relevant experience from being
present in the very early years in transition economies. With target countries
for further expansion being similar to CEE countries in the early years of
transition (Smoley, 2007[109]), Bank Austria might have best practices to
deal with such surroundings including political risk, a lack of institutional
quality and a just developing modern real sector.
Additionally the integration into the large UniCredit group enables Bank
Austria to expose itself to such higher risks. Bank Austria Creditanstalt, with
strong ?nancial backing of the complete group, is able to pursue a long-run
oriented strategy. Turkey and Russia might even be candidates for ?nancial
crisis in the next year in the eyes of BA, but the is be able to sustain 2-3 loss-
making years in the market, being able to focus on the long-term potential,
especially for Turkey.
2.3.6 Same region, di?erent strategy: The case of Er-
ste Bank AG in Central and Eastern Europe
Erste Bank AG also started out as a purely Austrian bank, rooting from
the mutual savings bank structure in Austria. Like Bank Austria the pub-
licly stated geographic focus of further business development is Central and
Eastern Europe[1].
In contrast to Bank Austria Erste Bank AG started to enter the CEE
markets relatively late, starting with the acquisition of Mezbank in Hungary
in 1998. This ?rst step was followed by further acquisitions of Cakovecka
banka, Bjelovarska banka and Trgovacka banka in Croatia in 1999
38
, Ceska
sporitelna in the Czech Republic in 2000
39
, Rijecka banka in Croatia in 2002,
Postabank in Hungary in 2003 and Novosadska banka in Serbia as well as
Banca Comericala Romana S.A.
40
in Romania in 2005. In July 2007 Erste
Bank, like Bank Austria Creditanstalt, expanded further eastwards, acquir-
ing Bank Prestige in the Ukraine[1].
38
These three banks were later merged into Erste&Steiermaerkischen banka.
39
To be precise Erste Bank bought a 52% majority share in this bank in 2000 gradually
increasing its share to 100% by 2005.
40
BCR is the largest bank in Romania with a market share of 32%.
41
Erste Bank AG operations in the respective countries today are subsumed
in Table 2.2.
Table 2.2: Erste Bank AG Operations in Central and Eastern Europe
Country Subsidiary Clients
(million)
Branches Market
position
(number
of clients)
Croatia • Erste Bank Croatia 0.6 114 No. 3
Czech
Republic
• Ceska Sporitelna 5.3 637 No. 1
Hungary • Erste Bank Hungary 0.9 186 No. 2
Romania • Banca Comerciala Romana 3.5 485 No. 1
Serbia • Erste Bank Serbia 0.3 NA No. 9
Slovak
Republic
• Slovenska Sporitelna 2.5 279 No. 1
Ukraine • Erste Bank Ukraine Just founded in December 2005
Source: Erste Bank AG Corporate websitehttp://www.erstebank.com
The main di?erence concerning general strategy between Bank Austria
Creditanstalt and Erste Bank AG is, that the former predominantly focuses
on the wholesale banking segment, including loans to larger commercials
as well as investment banking services, whereas Erste Bank AG’s primary
focus is the retail segment and, to a lesser degree, the provision of ?nancial
services to small and medium-sized (SME) ?rms[1]. This general di?erence
also seems to at least partly explain both di?erent timing and mode of entry
between these banks. Whereas the strategy of Erste Bank AG requires a large
branch network as well as soft information about informationally-opaque
potential loan clients, Bank Austria’s need for these due to its focus on
the wholesale business, was less pronounced (at least until 2000 when BA
started to increase its interest in the retail segment in the CEE region).
Therefore Erste Bank was much more active acquiring these needed assets or
capabilities via the acquisition of local banks. Erste Bank did not enter any
market on a large scale via Green?eld Investment but predominantly choose
an acquisition strategy from the beginning[1].
Also, the fact that Erste Bank AG entered the region way later than Bank
Austria Creditanstalt could also be seen in light of di?erent core business
segments. Whereas historically BA had a large portfolio of large Austrian
?rms equipped to enter the Eastern European markets, and therefore scope
42
for applying follow your customer-strategy, Erste Bank AG commercial client
portfolio was more skewed towards SMEs, who were less ready to engage in
the CEE region from the beginning. Additionally the retail segment in these
markets itself might have been to risky and small to enter in the early phase
of transition.
Erste Bank AG also focuses on less markets in the region, being active
in only 7 countries[1]. It has a very strong position in the retail segment
especially in the Czech and Slovak Republic and Romania, as well as to
a lesser degree in Hungary. Interestingly Erste Bank AG is not active in
Poland, deemed to be the most interesting market for retail banking in CEE
due to its large population.
Concerning future growth strategies Erste Bank AG tries to make use of
the banking sector development cycle in the heterogeneous markets in the
region. Whereas the Ukraine, Serbia and Romania are now seen as ”‘emerg-
ing markets”’ characterized by a low market penetration rate of banking
services, where growth short-term is driven by demand for simple banking
services such as savings, payment transfers and debit cards, these countries
will become what Erste Bank AG calls ”‘developing markets”’ like the Czech
and Slovak Republic, Hungary and Croatia, where bank penetration rates
are somewhat higher therefore experiencing tougher competition, but at the
same time enabling growth in high-margin products such as mortgage loans,
consumer loans, credit cards and wealth management products[1].
2.4 Conclusion
Quite a lot can be learned for multinational banking in general from the
above case study.
Concerning home country e?ects in multinational banking, it is con-
cluded, that banks from saturated markets are more likely to expand abroad,
due to missing signi?cant growth opportunities in the home market. This
result has already been discussed in the empirical literature (e.g. de Haas,
van Lelyveld (2006)[44]). Interestingly home country characteristics in the
majority of studies are most often seen as shaping ?rms’/banks’ capabilities
to become multinational (e.g. Buch and de Long (2004)[28], Focarelli and
Pozzolo (2000)[60], not so much as a push factor in incentives to become
multinational.
New to this discussion are soft home market e?ects mentioned by Bank
Austria Creditanstalt, namely that the relative size of the home country
43
has an e?ect on whether a potential target country is perceived by a multi-
national bank as su?ciently pro?table (large). Whereas this result seems
somewhat irrational, an additional home market size e?ect concerning ge-
ographical characteristics of expansion might be, that banks from a small
market might simply be better equipped to be pro?table in foreign small
markets, due to experience in this type of surroundings.
This knowledge in similar markets might also explain why some banks are
able to enter risky accending markets at an earlier stage than others. Bank
Austria seems to be more equipped to enter markets such as the former CIS
states due to experience in similar markets, namely the Central and Eastern
European countries at an early stage of transition.
One very clear result of the case study is the importance of follow your
customer strategies in multinational banking. Especially having customers
to follow seems to allow early entry in accending markets, which stand-alone
might not be attractive enough at this stage of development.
Additionally the geographic pattern of bank internationalization, at least
in non-OECD countries might to a large degree be explained by FYC strate-
gies. This result adds an additional explanation on the in?uence of distance
on multinational bank location choice, as distance might play a minor role
in this decision directly through information costs considerations (e.g. Buch
and de Long (2004)[28]) but a larger role indirectly as distance is an im-
portant variable in the location decision of real sector ?rms to be followed
who actually face a decision concerning the trade of physical goods over this
distance
41
.
Another mentioned bene?t of follow your customer strategies is that local
?rms might be easier to attract for a bank if the former has business ties to
multinational customers of the respective bank.
The most interesting point of the case study might be the ?ndings on how
a bank’s product strategy shapes entry modes and entry timing into foreign
markets, as well as the geographic pattern of such entry.
One conclusion derived is that banks focusing on the wholesale business,
characterized by informationally non-opaque clients with low switching costs
41
Along the FDI theories of e.g. Markusen and Venables (2000)[98] we would expect
market-seeking horizontal foreign direct investment between distant countries and vertical
production-cost minimizing foreign direct investment between neighbouring countries. For
less-developed countries we would rather expect in?owing vertical FDI. So in the early
CEE case we would expect countries close to Austria to be main recipients of vertical
FDI, therefore close countries to be main recipients of Austrian bank FDI.
44
and low requirements concerning the scope of the branch network, are more
likely to enter via Green?eld investment, whereas banks focusing on the retail
and SME
42
business are more likely to enter a foreign market via the acquisi-
tion of a local bank due to the client group’s informational opacity (especially
SMEs), high switching costs from existing bank relationships (especially re-
tail customers) and large requirements concerning the branch network scope
(especially retail customers). These assets might most easily be acquired via
the acquisition of an existing bank, whereas wholesale-focused entrants might
probably not be willing to pay for these assets, as they are of second-order
importance for their business focus.
Additionally, timing of entry in a respective market
43
might be a function
of product strategy of a bank. As discussed above, a retail-oriented bank not
having large real sector clients in its portfolio ex ante, might want to wait
to enter an accending market until the retail market in this country has
grown su?ciently large. In contrast, a wholesale-oriented bank might be
able to enter a market earlier on, banking on early pro?t from follow your
customer-business, business with third-country multinational ?rms as well
as investment banking services for the host country government and large
corporations.
For the economic literature these results propose a research strategy both
treating banks as conglomerates as well as on a market level taking care of
the fact that banking is not a single homogeneous industry. Therefore multi-
national bank behaviour might require di?erentiated theories for respective
banking segments. Entry modes, further strategic decisions and potentially
associated di?erent impacts of bank entry into foreign host markets might
have to be discussed separately for banks with a retail and SME ?nancial
services focus and wholesale/investment banking-oriented institutions.
The empirical literature, when discussing the location choice of multina-
tional bank subsidiaries, primarily focuses on host country e?ects. Taking
a result of the above case study into consideration, this picture might be
partially misleading. Indeed location strategies might take complete regions
into account, not single countries stand-alone. An isolated view on a re-
spective country market might be insu?cient, as due to potential network
e?ects within a region the market’s attractiveness also depends on potential
cost or revenue synergies with other markets in the region. The notion of
network economies of scope across regions also seems rather new to economic
42
Small and Medium-sized enterprises
43
This could also be seen as locational choice, as a market A in period t
i
might mimic
a market B in period t
j
.
45
theory concerning banking, but has been a standard feature of business lit-
erature discussing international one-stop-shopping possibilities for clients as
well as follow your customer strategies, which constitute a special case of
such network economies of scope.
The ?nal point I want to stress is the result obtained on the re?nancing
of multinational bank activity in host countries. Indeed it seems that in
su?ciently developed markets foreign bank activity does not lead to a large
volume of capital in?ows into this market, as the subsidiaries’ re?nancing
might primarily take place in the host country (via deposits). Therefore
internal capital markets in multinational bank organizations might only play
a residual role in the ?nancing of these subsidiaries.
46
Chapter 3
M&A versus Green?eld -
Optimal Entry Modes into
Markets with Sequential Entry
3.1 Introduction
One of the topics concerning foreign direct investment, that has just recently
become a core focus of economic literature, is the choice of the exact entry
mode of ?rms into foreign markets.
Foreign direct investment can take place via the acquisition of a local ?rm
in the target market (M&A), via the set-up of a completely new structure
(Green?eld Investment) or some impure organizational designs in between
these two modes (e.g. Joint Ventures, Brown?eld Investment).
The most general advantage for a ?rm entering via M&A instead of Green-
?eld lies in the di?ering e?ect on the ex post host market structure and
therefore on the degree of competition in the market ex post. A ?rm enter-
ing a market with d incumbents via Green?eld Investment will ex post face
d competitors in the market, whereas, when entering via M&A, it will only
face d-1 competitors (abstracting from the possibility of additional entry).
In markets with imperfect competition, e.g. due to horizontal product dif-
ferentiation, all else equal, pro?ts for the entrant should therefore be higher
if the ?rm enters via M&A. This positive characteristic of entry via M&A is
for example modelled by M¨ uller (2001)[105] and G¨org (2000)[70].
47
However, this bene?cial ”competition” e?ect might be a short-sighted
motivation for entry via M&A. If one would consider foreign ?rms to enter
markets sequentially, such that some ?rms due to various reasons
1
move into a
speci?c host market faster than others, the entry mode of early movers might
a?ect the entry decisions of sequential entrants in the future. Obviously, if
the early mover entered via Green?eld Investment, a sequential entrant would
face the decision whether and how to enter a market with d+1 incumbents,
whereas with early mover entry via M&A he would face a decision on a market
with d incumbents. Therefore sequential entry might be accommodated by
the early moving ?rm entering via M&A instead of Green?eld Investment.
So the static positive ”competition” e?ect of entry via M&A might to some
degree be o?set up by an increase in the likelihood of further sequential entry.
Two types of questions arise in such a ”dynamic” setting, the ?rst one
being ”ex post”, how the entry mode of the early mover a?ects potential
sequential entrants’ decisions, and therefore ”ex ante” how the optimal entry
mode choice for the early mover looks like given its e?ect on sequential entry.
To analyze entry mode decisions in a less static setting a simple two-period
model, featuring a potential early and sequential entrant, is proposed.
To focus on pure market structure e?ects of entry modes I, at least in
the basic model, abstract from other incentives for entry via M&A such
as asset complementarity between acquirer and target (see e.g. Nocke and
Yeaple (2007)[107]), but from the possibility, that an entrant might be bet-
ter equipped to operate a company in general compared to its host country
target. In this simple setup ?rms are assumed to be symmetric in their mar-
ginal costs, only di?ering in what can be called their ?xed costs of operating
in a market. In such a model I show, that entry via Green?eld Investment
indeed has a valuable strategic entry deterring e?ect for early entrants, such
that sequential entry in general is less likely than if the early entrant enters
via M&A. It is also shown, that the chosen entry mode of later entrants dif-
fers in expectations for di?ering early mover entry modes, as do acquisition
prices for domestic targets. Which early movers should enter via Green?eld
Investment or M&A, and how modelling potential sequential entry reduces
the relative probability of early movers choosing M&A over Green?eld In-
vestment as entry mode, is then derived.
To sum up, the model will have the following intuition. The general trade-
o? faced by ?rms, when choosing their entry mode, is between increasing net
variable pro?ts (variable pro?t minus acquisition price) and not having to
1
For example practitioners mentioned that German banks did not enter the market
in Eastern Europe early on because they focused on building up their business in East
Germany at this time.
48
sustain the ?xed costs of Green?eld entry when entering via M&A, versus
working in the market at lower ?xed operating costs per period when entering
via Green?eld Investment. The heterogeneity in possible entrants’ general
?xed operating cost levels determines their respective ?xed cost disadvantage
when entering via M&A.
Modelling potential sequential entry reduces the expected variable pro?t
advantage of entry via M&A, as this kind of entry is shown to increase the
probability of pro?t-reducing sequential Green?eld entry. Additionally, the
pro?t reducing e?ect for the early entrant of sequential entry via Green?eld
Investment is more pronounced if the early entrant entered via M&A, than if
it had entered via Green?eld Investment. In sum these two e?ects are shown
to reduce the attractiveness of M&A entry for all types (concerning ?xed
operating costs) of early entrants, even though the threat of sequential entry
also works to reduce the acquisition price to be paid for takeover targets. As
the ?xed cost disadvantage of M&A is una?ected by such an entry threat,
a smaller relative fraction of potential early entrants (namely only the very
?xed-cost e?cient ?rms) chooses entry via M&A compared to a static setting
not taking the threat of sequential entry into account.
The main result of the basic model is, that taking into account the threat
of sequential entry reduces the incentive for early movers to enter via M&A
compared to a static world. This result holds true for all markets where both
kinds of entry modes are principally probable
2
. These market characteris-
tics should be present in the majority of real-world markets, as we generally
observe both kinds of foreign direct investment modes in the respective coun-
tries.
The possibility of market incumbents to deter or accommodate further
entry is well-established in economic literature, starting with the work of
Stackelberg (1934)[122] and continuing with contributions by e.g.Bernheim
(1984)[20] and Gilbert and Vives (1986)[67]. However, according to my
knowledge, no one has linked entry deterrence in models with sequential
entry to the entry mode choice of multinational corporations. Also, the ma-
jority of theories have been restricted discussing entry deterrence in markets
with Cournot competition.
The proposed basic model can therefore contribute to the understand-
ing why Green?eld Investment still makes up a substantial share of foreign
2
This means, that structures are such that ?rms entry choice will be heterogeneous in
equilibrium, with some ?rms entering via Green?eld Investment, some via the takeover of
an incumbent ?rm and some will not enter the market at all.
49
direct investment. For example, Ra?, Ryan and Staehler (2006)[113] ?nd,
that for Japanese ?rms Green?eld Investment is the dominant mode of entry,
with cases of Green?eld entry outnumbering M&A entry by a factor of 2.5 to
1. Even concerning value of investments, Lorentowicz, Marin and Raubold
(2002)[95] show evidence that for German Direct Investment in Eastern Eu-
rope Green?eld Investment makes up 56% of total FDI ?ows. Concerning
number of projects, this dominance of Green?eld Investment in their study
is even more striking when one considers that the mean size of M&A invest-
ments is roughly 2.4 times the mean size of Green?eld Investments in the
authors’ dataset.
Recent theoretical contributions on optimal entry modes have however
stressed additional reasons of ?rms to enter via M&A, especially the impor-
tance of asset complementarity (e.g. Nocke and Yeaple (2007)[107]). Con-
cerning the found dominance of Green?eld Investment in the number of oc-
currences, we think a counterbalancing theoretical contribution stressing the
advantage of Green?eld Investment is needed.
Also, when controlling for ?rm size e?ects, Ra?, Ryan and Staehler
(2006)[113] ?nd, that the widespread perception that the most e?cient ?rms
enter via Green?eld Investment, does not hold true. Indeed, when controlling
for ?rm size, they ?nd, that tendentially the more e?cient ?rms enter via
M&A, though the e?ect is not signi?cant. Andersson and Svensson (1994)[4]
?nd, that ?rms with strong organizational skill tend to enter foreign markets
via the acquisition of local incumbents rather than via Green?eld Invest-
ment. Theoretically the latter point is theoretically rationalized in the paper
by Nocke and Yeaple (2007)[107], as an internationally mobile asset such as
organizational skill can be leveraged by combining it with non-mobile capa-
bilities of acquired local market incumbents. I incorporate these ?ndings,
that have else been neglected in the literature, to date, with the general per-
ception being that the most e?cient ?rms enter via Green?eld Investment
(e.g. Mueller (2001)[105]), in my model setup.
Additional to the above main result it is shown, that the advantageous
e?ect of early entry via Green?eld Investment is more pronounced when
takeover possibilities in a market are limited, such that early entry via M&A
deletes the option of M&A entry for sequential entrants. Interestingly de-
creasing potential competitors options then has a negative e?ect for early
movers. I therefore come up with a new, strategic indirect, ”‘perverse”’ ef-
fect of missing takeover targets on the choice of entry mode of ?rms into
foreign markets.
Welfare analysis within the proposed basic model yields, that in a world
50
with sequential entry, the negative e?ect welfare e?ect of entry via M&A is
even more pronounced than in a static setting. This is due to the fact, that
additional sequential entry, which is accommodated by early entry via M&A,
in the proposed setup reduces pro?ts of foreign ?rms more than it increases
consumer rent.
A ?nal proposed extension of the basic model is to incorporate what can
be called country- or market-speci?c learning by doing-e?ects, which changes
the results from the basic model for some types of markets. Indeed it is
shown, that when the degree of product di?erentiation in the market is low
and learning-by-doing e?ects are su?ciently strong, an ”e?ciency e?ect”
dominates the ”competition e?ect” of entry modes, such that in this case
early entry via M&A deters sequential entry. In this setting then taking into
account the threat of sequential entry increases the incentive for early movers
to enter via M&A.
One could argue that this latter extension might yield an especially valu-
able insight into the retail and commercial banking sector, where ex ante
product di?erentiation between banks is generally perceived to be low and
learning-by-doing e?ects perceived to be both strong as well as predominantly
country-speci?c, due to the heterogeneity of banking regulation across coun-
tries. Indeed the result, that in this industry M&A entry should be very
dominant due to its additional sequential entry deterring nature, is getting
support from the ?ndings in the case study in this thesis.
From a general theoretical point of view, this chapter contributes to the
existing literature by being, to my knowledge, the ?rst to leave a static world
of entry mode decision analysis to implement a more forward-looking behav-
iour of potential foreign direct investors. Additionally, within the model, I am
able to discover yet undiscussed potential e?ects of limited takeover possibil-
ities on entry mode choice, as well as the e?ect of country- or market-speci?c
learning by doing e?ects on the choice of entry modes, in at least a stylized
way.
The rest of the chapter proceeds as follows. Section 2 lays out the basic
model and assumptions, analyzes the contingent sequential entry structure
and the derives the optimal entry mode decision of early movers, ?nally
comparing the results to a benchmark world without sequential entry. Section
3 discusses welfare implications Two extensions/setup changes are introduced
in the following. In Section 4 the e?ect of limited takeover possibilities is
analyzed. Then the model is extended by including country-speci?c learning-
by-doing e?ects in Section 5. The ?nal section concludes and discusses open
questions.
51
3.2 The basic model
The building stone of the model is a host country market characterized by
Bertrand competition in horizontally di?erentiated goods. The market is set
up as a Salop circle of size Y = 1 and transport costs or degree of product
di?erentiation t > 0.
Before any foreign entry into the market, there are two incumbent do-
mestic ?rms A and B operating in the market at marginal costs c
A
and c
B
and ?xed costs of operation per period of O
A
and O
B
, respectively.
The simple timing structure of the model is illustrated by the following
graphic.
Figure 3.1: Time Structure of the Model
Foreign
firm 1 entry
(mode)
decision
T=1: First
period of
market
competition
Foreign
firm 2 entry
(mode)
decision
T=2: Second
period of market
competition
T
Foreign entry happens sequentially, such that an early-moving ?rm enters
the market one period before the next potential entrant.
At the beginning of period T = 1 there is one potential foreign entrant
F
1
with marginal costs c
1
and ?xed costs of operation per period of O
1
.
The potential entrant ?rm has three options concerning entry into the
host country market. It can either enter the market via the acquisition of
an incumbent ?rm (M&A), by establishing a completely new ?rm structure
in the host country (Green?eld Investment) or not enter the market at all.
After the entry decision the ?rst market game in the host country market
takes place.
At the beginning of period T = 2 a second potential entrant F
2
with
marginal costs c
2
and ?xed costs of operations per period of O
2
, with O
2
52
assumed to be uniformly distributed between 0 and 1
3
, decides on whether
and how to enter the market. After the sequential entry decision the second
market game takes place.
For simplicity symmetric marginal costs of ?rms are assumed, such that
c
1
= c
2
= c
A
= c
B
= c. Therefore it is straightforward that ?rms will locate
equidistantly to each other on the Salop Circle
4
.
Additionally, incumbent ?rms are assumed to be perfectly symmetric,
such that O
A
= O
B
= O
D
> 0, where O
D
is then the level of ?xed costs of
operation per period for all domestic ?rms.
Concerning entry modes the setup is as follows.
If a ?rm enters via Green?eld Investment it bears ?xed costs of entry of
F > 0. The newly setup structure will then work at marginal costs c and
?xed costs O
i
.
If it enters via M&A, the ?rm has to pay an endogenous acquisition price
A for the respective target. Bargaining power is assumed to reside with
the acquirer, such that the acquisition price will equal the outside option of
the target ?rm, which is the respective targets (expected) foregone pro?ts
when staying independent. The acquired structure will work at marginal
costs c and ?xed costs ?O
i
with ? = 2 for simplicity. The results would be
unchanged as long as we assume ? > 1.
Therefore ?xed costs of operation are assumed to be higher under M&A
then under Green?eld Investment. ? determines how relatively large this
di?erence is. The usual explanation for higher (operational) costs under
operating with an acquired organization is, that company cultures between
target and acquirer may clash (see e.g. Feely and Kompare (2003)[57])or that
there might be some costs of restructuring the target to ?t into the acquirer’s
organizational structure (see e.g. M¨ uller (2001)[105]). In the proposed setup
the ?xed costs of operation are twice as large under M&A compared to stand-
alone Green?eld operation of entrants.
Special to the proposed setup is then, that the absolute negative e?ect of
operating with an acquired organizational structure instead of a completely
new setup structure (Green?eld Investment) on operating ?xed costs depends
on the general ?xed costs e?ciency O
i
of the entering bank. The intuitive
idea is, that a generally well managed ?rm should also be better equipped to
handle post-merger integration problems than a badly managed ?rm.
3
The actual realisation of O
2
of the potential entrant is then drawn by nature.
4
Obviously this would be the pro?t maximizing location choice. As in this model
incumbents have a ”‘location history”’, it is implicitly assumed that location switching
costs are zero.
53
Note that for reasons of tractability the model abstracts from directly
arising di?erences in marginal costs between Green?eld and M&A operation.
Therefore I do not claim to fully discuss the e?ect of a ?rms general e?ciency
on its choice of entry mode. I’ll be content to discuss e?ects of non-marginal
cost e?ciency on entry mode choice.
The ?nal restriction made to reduce cases to be analyzed is what can
be called a ”no passive consolidation”-clause, which means that the focus of
the analysis will be on cases of transport costs t and incumbent ?rms’ ?xed
costs O
D
such that incumbent ?rms are not driven out of the market by
competition
5
. Such passive consolidation would occur if pro?ts of incumbents
would be negative in the case of four players in the market, so i? t < 16O
D
6
.
To sum up, general pro?t functions of ?rms then look as follows:
For ?rms entering via Green?eld Investment:
?
i,j
(GF) = (p
i,j
(GF) ?c
i
) ×x
i,j
(GF) ?O
i
?F (3.1)
For ?rms entering via Acquisition:
?
i,j
(MA) = (p
i,j
(MA) ?c
i
) ×x
i,j
(MA) ?2O
i
?A
j
(MA) (3.2)
with i ? {1, 2} denoting early and sequential entrant and
j ? {MA, GF, NE} the entry mode of the other entrant.
For incumbent ?rms, pro?ts of operating in the market are simply respec-
tive variable pro?ts depending on the entry mode(s) of entrant(s) minus the
?rms’ respective ?xed operating costs.
?
A,j
= ?B, j = (p
A,j
?c
A
) ×x
A,j
?O
D
(3.3)
In appendix 1 the respective variable pro?t for ?rms for di?erent cases
of entry are derived. In this basic model the only di?erence between entry
modes concerning variable pro?ts lies in the respective number of ?rms that
compete in the market.
5
Besides reducing cases to be analyzed let us abstract from passive consolidation among
incumbents to keep the analysis non-trivial. If e.g. any kind of entry of ?rm 1 would lead
incumbent ?rms to leave the market in period 1 then there would never be an incentive
for 1 to enter the market via M&A and no possibility for 2 to enter the market via M&A
in period , as there would be no targets to be acquired.
6
The respective pro?ts under a given number of market participants are derived in
appendix 1.
54
Before getting onto the core questions one intermediate result is derived
to help determine equilibrium acquisition prices.
Lemma 1
The structure of entry decisions is such, that ?rms will enter via M&A if
they have low ?xed costs of operation, via Green?eld Investment for medium-
level ?xed costs and will not enter the market at all with high ?xed costs of
operation.
The Lemma is proved in appendix 2 and the following graph illustrates
this result.
Figure 3.2: The General Structure of Entry Mode Decisions
O
i
?
i
?>1
1
Greenfield entry M&A entry No entry
?
GF
?
MA
?
i
Õ
i
Therefore one can conclude that ?rms making an acquisition o?er to
an incumbent ?rm will be those that would enter via Green?eld if the ac-
quisition fails to materialize. Therefore the outside option of the target is
its (expected) pro?ts under the case that the potential acquirer enters via
Green?eld Investment.
First now the ex post question will be analyzed to determine how entry
patterns di?er for the sequential entrant in period 2 depending on the early-
mover’s choice of entry mode.
55
3.2.1 Analyzing entry (mode) decisions of the sequen-
tial entrant in period T=2
In general the entry decision of ?rm 2 can be subsumed as follows
a) Firm F
2
enters via M&A i?
?
2,j
(MA) > ?
2,j
(GF) and ?
2,j
(MA) > 0
7
b) Firm F
2
enters via Green?eld i?
?
2,j
(GF) > ?
2,j
(MA) and ?
2,j
(GF) > 0
c) Firm F
2
does not enter the market i?
?
2,j
(GF) < 0 and ?
2,j
(MA) < 0
8
From this general decision structure sequential entrants’ decisions contin-
gent on early entrant entry mode decisions can be derived.
Early mover entry via Green?eld Investment
In this case ?rm 2 decides upon entry and entry mode into a market with three
incumbents, the two domestic ?rms and the early entrant ?rm 1. Therefore
if ?rm 2 enters via Green?eld four ?rms will divide market pro?ts between
them. Firm F
2
pro?t then is
?
2
(GF) = ?
V ar
(4) ?F ?O
2
=
t
16
?F ?O
2
Firm F
2
will not enter the market at all if Green?eld pro?ts are negative,
so i?
t
16
?F ?O
2
< 0 or ?xed costs of operation above
O
GF
2
=
t
16
?F (3.4)
Firm F
2
pro?t when entering via M&A in this case is
?
2
(MA) = ?
V ar
(3) ??
V ar
(4) +O
D
?2O
2
9
7
Where we do know that the former is binding due to the structure of the entry mode
decision.
8
Where the former is binding again due to the structure of the entry mode decision.
9
As the acquisition price to be paid in this case equals ?
V ar
(4) ?O
D
.
56
Firm F
2
will enter via M&A if M&A pro?ts are larger than Green?eld
pro?ts, so i?
?
V ar
(3) ??
V ar
(4) ??
V ar
(4) +F +O
D
?2O
i
=
t
9
?
t
16
?
t
16
+O
D
+F ?O
2
> 0
or ?xed costs of operation below
O
2
GF
= ?
t
72
+O
D
+F (3.5)
Firm F
2
will enter via Green?eld Investment if such entry yields both
positive pro?ts and higher pro?ts than entry via M&A, so i?
t
16
?F ?O
2
> 0
and ?
t
72
+O
D
+F ?O
2
< 0, or ?xed costs of operation O
2
in the range
O
2
GF
< O
2
< O
GF
2
(3.6)
Early mover entry via M&A
In this case ?rm F
2
decides about entry and entry mode into a market with
two incumbents, one independent domestic ?rm and the early entrant orga-
nization. Firm F
2
pro?t under Green?eld Investment then is
?
2
(GF) = ?
V ar
(3) ?F =
t
9
?F ?O
2
In this case F
2
will not enter the market i?
t
9
?F ?O
2
< 0 or ?xed costs
of operation above
O
MA
2
=
t
9
?F (3.7)
Firm F
2
pro?t under M&A then is
?
2
(MA) = ?
V ar
(2) ??
V ar
(3) +O
D
?2O
2
=
5
36
t +O
D
?2O
2
The condition for F
2
to enter via M&A is then
?
V ar
(2) ??
V ar
(3) ??
V ar
(3) +F +O
D
?2O
i
=
t
4
?
t
9
?
t
9
+O
D
+F ?O
2
> 0
or ?xed costs of operation below
O
2
MA
=
t
36
+O
D
+F (3.8)
F
2
will therefore enter via Green?eld if its ?xed operational costs are such
that
57
O
2
MA
< O
2
< O
MA
2
(3.9)
As a side result, from the above cases it is obvious, that M&A entry in this
model in general is preferred to entry via Green?eld Investment if ?xed costs
of Green?eld entry as well as the operating ?xed costs of domestic banks are
large, as the latter leads to a reduction in the acquisition price to be paid.
Comparing contingent entry and entry mode probabilities
One can now compare the respective entry and entry mode probabilities for
sequential entrants for the cases of ?rm F
1
entering via Green?eld Investment
or via M&A. With O
2
uniformly distributed between 0 and 1 the probabilities
are easily matched with the respective ?xed costs threshold levels for the
respective entry mode
10
.
The following analysis is restricted to cases where parameters t, O
D
, F are
such that all probabilities are larger than zero and smaller than one. Intu-
itively that means we look at markets where all entry modes are possible in
general, so depending on operating cost levels of a respective potential entrant
market structure leads to heterogeneous pro?t-maximizing entry strategies,
with some ?rms preferring entry via M&A, others preferring Green?eld entry
and some maximizing pro?t by not entering at all.
Taking into account the threshold levels and the characteristics of the
uniform distribution, the probability of any sequential entry given that ?rm
F
1
entered via Green?eld is
P
1
= O
GF
2
=
t
16
?F (3.10)
The probability of any sequential entry given that ?rm F
1
entered via
M&A is
P
2
= O
MA
2
=
t
9
?F (3.11)
10
The probability of e.g. O
i
< O
i
for any distribution of O
i
simply equals the cumulative
distribution function from the lower bound of the distribution up to O
i
. For a uniform
distribution between bounds a = 0 and b = 1 the cumulative distribution then is
F(O
i
) =
O
i
?a
b?a
therefore F(O
i
) = O
i
58
Intuitively, a potential sequential entrant with overhead costs below these
threshold level will enter the market in some mode, while if the potential
entrant has higher costs he will not enter, as entry would result in negative
pro?ts then. The following proposition directly follows from comparing the
above probabilities.
Proposition 1
The probability of sequential entry is lower if ?rm 1 entered via Green?eld
Investment instead of M&A. Potential sequential entrants deterred from entry
in the former case are ?rms with ?xed operating costs of O
MA
2
< O
2
< O
GF
2
.
The probability of sequential entry is reduced by P
E
=
7
144
t.
The proposition is derived in appendix 3. Note that the reduction of
the probability of sequential entry is larger the larger, the degree of prod-
uct di?erentiation t in the market. The reasoning is, that the larger t, the
stronger the negative e?ect on pro?ts of an increase in the number of market
participants is for market participants
11
.
The absolute probability of sequential entry via M&A equals the oper-
ating cost threshold level for M&A entry of the potential sequential entrant
F
2
, therefore
P
3
=
O
2
GF
= ?
t
72
+O
D
+F (3.12)
if F
1
entered via Green?eld Investment, and
P
4
=
O
2
MA
=
t
36
+O
D
+F (3.13)
if F
1
entered via M&A.
Comparing these probabilities gives further insight into the e?ect of early
entrant’s decisions on successive entry modes, as stated in Lemma 2.
Lemma 2
The absolute probability of sequential entry via M&A is lower if ?rm 1
enters via Green?eld Investment instead of M&A . Potential sequential en-
trants with
O
2
MA
< O
2
<
O
2
GF
would enter via Green?eld Investment in the
former case and via M&A in the latter. The probability of sequential entry
via M&A is reduced by
t
24
.
11
This can also be seen technically, as the equilibrium price chargeable in the market is
p =
1
n
t, so
?p
?n?t
= ?n
?2
. This shows that an increase in the number of market participants
reduces prices in the market stronger, the larger the degree of product di?erentiation.
59
The Lemma is proved in appendix 4.
As we will discuss later on, the most important question for the early
entrant is, how its’ entry mode will a?ect the absolute probability of se-
quential entry via Green?eld Investment. This probability simply equals the
probability of any kind of sequential entry minus the probability of sequential
entry via M&A. So the absolute probability of sequential entry via Green?eld
Investment is
P
1
?P
3
=
t
16
?F ?[?
t
72
+F +O
D
] =
11
144
t ?2F ?O
D
for early entry via Green?eld Investment and
P
2
?P
4
=
t
9
?F ?[
t
36
+F +O
D
] =
t
12
?2F ?O
D
for early entry via M&A.
Again comparing these probabilities the following important Lemma can
be stated.
Lemma 3
Early entry via Green?eld Investment compared to entry via M&A reduces
the probability of sequential entry via Green?eld Investment by
t
144
.
The Lemma is derived in Appendix 5.
The e?ects of the early movers’ entry mode on F
2
’s entry decision can be
subsumed by the following graphic.
60
Figure 3.3: The E?ect of Early Entry Mode on Sequential Entry
To sum up, early entry via Green?eld Investment compared to early entry
via M&A has the following e?ects on sequential entry. For one, the absolute
probability of sequential entry is lower in the former case. This is however
not a su?cient result for our following analysis, as it is shown, that part of
the total reduction of entry also occurs via the reduction of the probability of
sequential entry via M&A. Further analysis however shows, that early entry
via Green?eld Investment is shown to de?nitely decrease the probability of
sequential entry via Green?eld Investment, which is the actual harmful type
of sequential entry from the point of view of the early mover in this model.
3.2.2 The optimal entry mode of the early mover
Now knowing how the entry mode of the early mover shapes sequential entry
probabilities, one can analyze the optimal entry mode for the early-moving
?rm F
1
.
To that end, pro?ts of ?rm F
1
under entry via Green?eld and entry via
M&A, given the respective e?ects on sequential entry, are simply compared.
Firm 1 pro?t when entering via Green?eld Investment can be written as
61
?
1
(GF) =
?
V AR
1
(3) which is the variable pro?t of F
1
in period 1
+P
3
×?
V AR
1
(3) which is the probability weighted variable pro?t of F
1
with F
2
entering via M&A in period 2
+(P
1
?P
3
) ×?
V AR
1
(4) which is the probability weighted variable pro?t of F
1
with F
2
entering via Green?eld in period 2
+(1 ?P
1
) ×?
V AR
1
(3) is the probability weighted variable pro?t of F
1
with F
2
not entering in period 2
?2O
1
?F is the ?xed costs of operation in both periods and of Green?eld entry
for F
1
Inserting variable pro?ts and rearranging the terms we can rewrite pro?t
under Green?eld as
?
1
(GF) = 2 ×
_
t
9
?O
1
?
F
2
_
?(P
1
?P
3
) ×
7
144
t (3.14)
where (P
1
?P
3
) ×
7
144
t constitutes the expected negative e?ect of sequen-
tial entry via Green?eld on ?rm 1 pro?ts, as (P
1
?P
3
) constitutes the ab-
solute probability of sequential entry via Green?eld Investment.
In the proposed setting, it is quite obvious that the only sequential en-
try mode a?ecting ?rm 1 pro?ts negatively is Green?eld Investment as this
increases the number of ?rms in the market
12
.
Firm 1 pro?t when entering via M&A can be written in similar style
12
Entry via M&A in this model does not change pro?ts of ?rm 1 as it is assumed that
all foreign entrants and local incumbents have the same marginal costs. Therefore sequen-
tial entry via M&A, from the point of view of ?rm 1, simply constitutes interchanging
identically behaving ?rms in the market.
62
?
1
(MA) =
?
V AR
1
(2) ??
V AR
A
(3) which is the variable pro?t of F
1
minus the ”variable part of
the acquisition price” in period 1
+P
4
×
_
?
V AR
1
(2) ??
V AR
A
(3)
_
is the probability weighted variable pro?t of F
1
minus the ”variable part of the acquisition price” with F
2
entering via M&A in
period 2
+(P
2
?P
4
) ×
_
?
V AR
1
(3) ??
V AR
A
(4)
_
is the probability weighted variable pro?t
of F
1
minus the ”variable part of the acquisition price” with F
2
entering via
Green?eld in period 2
+(1 ?P
2
) ×
_
?
V AR
1
(2) ??
V AR
A
(3)
_
is the probability weighted variable pro?t of
F
1
minus the ”variable part of the acquisition price” with F
2
not entering in
period 2
?4O
1
+ 2O
D
is the ?xed costs of operation in both periods and the e?ect of a
targets ?xed costs on the acquisition price.
Again inserting variable pro?ts and rearranging yields
?
1
(MA) = 2 ×
_
5
36
t ?2O
1
+O
D
_
?(P
2
?P
4
) ×
13
144
t (3.15)
with (P
2
?P
4
) ×
13
144
t again being the expected negative pro?t e?ect of
sequential entry via Green?eld Investment of ?rm 1 pro?ts.
Solving for respective operating ?xed costs pro?t levels like it has been
done for the sequential entrant in T=2 the following result can be derived.
Lemma 4
Firm 1 will enter via M&A, if its ?xed costs of operation are low, such
that
O
1
<
¯
O
1
=
t
36
+O
D
+
F
2
?(P
2
?P
4
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
via Green?eld Investment if ?xed costs of operation are of medium size,
such that
¯
O
1
< O
1
< O
1
=
t
9
?
F
2
?(P
1
?P
3
) ×
7
288
t
and will not enter the market if ?xed costs of operation are high, such
that
O
1
> O
1
63
The Lemma is derived in Appendix 6. The general pecking order of the
entry mode of the early moving ?rm 1 is not in?uenced by the threat of
sequential entry and therefore arising strategic e?ects of entry mode, which
is unsurprising due to the way the model is setup, as this order is determined
by assumptions on the in?uence of entry modes on the ex post operating
e?ciency of the foreign structure in the market.
The more interesting question is whether the threat of sequential entry
reduces the probability or amount of M&A and Green?eld entry in compar-
ison to a world, where there no such threat exists. In order to analyze this,
a two-period benchmark case without a sequential entry threat is derived.
For further use, the relative probability of early entry via M&A given
general entry is simply
¯
O
1
O
1
, which intuitively is simply the proportion of
?rms with operating ?xed costs such that they enter via M&A, divided by
the proportion of ?rms with operating ?xed costs such that they enter the
market at all.
3.2.3 Benchmark Case (No sequential entry)
The following benchmark case without a sequential entry threat is con-
structed. The benchmark would be a two-period market game with perfectly
symmetric periods concerning market structure from the perspective of the
early entrant.
In such a setting, cumulative pro?ts of ?rm 1 under Green?eld Investment
over both periods are then
?
BM
1
(GF) = 2 ×
_
t
9
?O
1
?
F
2
_
(3.16)
which is simply twice the pro?t of ?rm 1 in period 1 under Green?eld
Investment
13
.
Cumulative pro?ts of ?rm 1 under M&A entry are therefore
14
?
BM
1
(MA) = 2 ×
_
5
36
t ?2O
1
+O
D
_
(3.17)
13
With three players in the market in each pro?t each ?rm makes variable pro?ts of
1
9
t
per period. Fixed costs of entry are F, which occur only once, so per period ?xed costs of
Green?eld entry are
F
2
.
14
Entry via M&A leaves two players in the market generating per-period variable pro?ts
of
1
4
t. Foregone target pro?ts per period would have occurred in a market with three
players, so the acquisition price per period would be
1
9
t ?O
D
.
64
Solving for ?xed operation cost levels O
1
one ?nds that in the benchmark
case ?rm 1 will enter via M&A if
O
1
<
¯
O
BM
1
=
t
36
+O
D
+
F
2
(3.18)
will not enter the market if
O
1
> O
BM
1
=
t
9
?
F
2
(3.19)
and for ?xed overhead cost levels in-between ?rm 1 would enter via Green-
?eld Investment.
It can be immediately seen, that
¯
O
BM
1
=
¯
O
1
+ (P
2
?P
4
) ×
13
288
t ?(P
1
?P
3
) ×
7
288
t
and
O
BM
1
= O
1
+ (P
1
?P
3
) ×
7
288
t
The relative probability of entry via M&A given general entry is
¯
O
BM
1
O
BM
1
for
the benchmark case.
3.2.4 How does the threat of sequential entry change
the entry mode decision of early movers?
Comparing the benchmark case with the case of potential sequential entry
yields the following ?rst result.
Lemma 5
The threat of sequential entry leads to a lower probability of general entry
for early movers.
The result is derived in Appendix 7.
This Lemma is very intuitive. With some positive probability another
?rm will enter the market in period 2 via Green?eld Investment in the case of
potential sequential entry. Therefore, with this positive probability, variable
pro?ts of the early mover will be lower in period 2. So some potential early
moving ?rms able to make small positive pro?ts in a market with two
65
other competitors will in expectation make negative pro?ts, as there is the
possibility of having to compete with three other competitors in period 2.
Our main interest however concerns the non-trivial e?ect of introducing
the threat of sequential entry on the relative probability of the early entrant
choosing entry via Green?eld Investment over entry via M&A. Analyzing this
question gives the following main result of the basic model.
Proposition 2
The threat of sequential entry leads to entrants choosing Green?eld In-
vestment over M&A with a higher probability, formally
¯
O
BM
1
O
BM
1
>
¯
O
1
O
1
if the general structure of the market supports all available entry modes
with positive probability.
This proposition is proved in Appendix 8.
The main result is easily explained intuitively. With sequential entry,
Green?eld Investment, compared to early entry via M&A, reduces the prob-
ability of sequential Green?eld entry, which would reduce early mover pro?ts.
Due to expectations, acquisition prices for targets will also be lower for
the early entrants with the threat of sequential entry, as incumbent domestic
?rms expect lower pro?ts in period 2. However the variable pro?t-bene?t of
entry via M&A decreases disproportionately due to the non-linear relation-
ship between pro?ts and number of ?rms in the market.
Altogether the existence of a sequential entry threat makes M&A a less
interesting entry mode option compared to a static setting.
The di?erence between this model of sequential entry and the benchmark
case is illustrated by the following graphical example, where it is quite easy
to see that the relative probability of M&A entry decreases strongly when
modelling sequential entry.
66
Figure 3.4: Comparison of Benchmark and Sequential Entry Case
In the proposed model setup there are markets that would only be entered
via M&A by early movers, no matter sequential entry threat or not. This
would be the case if either ?xed costs of Green?eld entry and ?xed operative
costs of incumbents are su?ciently large
15
. One can easily see that O
1
?
O
1
would become negative in this case as well as P
2
> P
4
, such that entry would
only occur via M&A.
In general, however, neglecting the market dynamics e?ects of entry
modes underestimates the attractiveness of entry via Green?eld Investment.
15
Restricted to F being small enough for potential acquirers entering via Green?eld
Investment if takeover negotiations fail and operating ?xed costs O
D
such that no passive
consolidation takes place.
67
3.2.5 A note on completely endogenous market struc-
ture
For simplicity the model only discusses at 2 potential entrants. In a com-
plete equilibrium we would have sequential entry in period 2 until pro?ts for
potential entrants become zero.
However, if e.g. one would assume n potential symmetric sequential en-
trants with operating ?xed costs of O
2
the qualitative results would not
change at all. In fact entry deterrence considerations would become more
important for ?rm 1 as its pro?ts would reduce much more for n
Entry
> 1
?rms entering the market.
3.3 Welfare analysis
In this section it is shown how welfare e?ects of entry modes di?er between
a model without potential sequential entry and with such a threat. To that
end, the welfare e?ects of ?rm 1 entering via Green?eld and via M&A are
analyzed.
As a simple welfare measure, the sum of pro?ts of domestic ?rms (which
include acquisition prices if a domestic ?rm is acquired) and the consumer
rent in the market, is used.
Welfare in the case of ?rm 1 entering via Green?eld is
W
GF
= ?
GF
Dom
+CR
GF
= ?
GF
A
+?
GF
B
+CR
GF
(3.20)
where ?
GF
Dom
is the sum of pro?ts of domestic incumbent ?rms A and B in
both periods and CR is the consumer rent in both periods, which in general
per period is
_
s ?td ?p
_
×X, where s is consumer willingness to pay, td is
the average transport costs incurred by customers and p is the equilibrium
price in the market and X is the trade volume in the market. As we assumed
market size equal to 1 and s large enough, such that total demand in the
market equals market size, CR reduces to s ?p ?td
16
.
Total domestic pro?ts are then simply
17
16
It is common knowledge, that average transport costs for customers in the Salop setup
equal
t
4n
with n the number of ?rms in the market.
17
to reduce notational clutter in the following it is assumed that c = 0 without loss of
generality.
68
?
GF
Dom
=
4
9
t ?4O
D
?P
1
×
7
72
t +P
3
×
7
144
t
18
where it is straightforward, that
4
9
t ? 4O
D
= ?
GF
Dom
(BM) is the pro?t of
domestic ?rms without potential sequential entry. Compared to the bench-
mark case, one can observe that domestic ?rm pro?ts are lower in markets
that generally support all entry modes
19
, which is obvious as a potential
additional competitor reduces pro?ts for each market participant.
Inserting case dependent consumer rents and case probabilities and rear-
ranging we get the following consumer rent in case of early entry via Green-
?eld Investment
CR
GF
= 2s ?
7
18
t + (P
1
?P
3
) ×
5
72
t
where again it is straightforward, that 2s ?
7
18
t = CR
GF
(BM) is the
consumer rent without potential sequential entry. Again comparing to the
benchmark case we can see that consumer rent is higher in markets that gen-
erally support all entry modes, which is intuitive, as an additional sequential
entrant reduces both prices to be paid as well as average transport costs for
the consumers.
Welfare in the case of entry via Green?eld Investment by ?rm 1 is there-
fore
W
GF
= ?
GF
Dom
+CR
GF
=
t
18
?4O
D
+ 2s ?P
1
t
36
?P
3
t
48
(3.21)
or W
GF
= W
GF
(BM) ?P
1
t
36
?P
3
t
48
.
Deriving welfare under early entry via M&A works the same way.
The sum of pro?ts of domestic ?rms in this case is
?
MA
Dom
=
13
18
t ?4O
D
?P
2
27
144
t +P
4
7
144
t
18
Deriving these pro?ts is simple. If ?rm 1 enters via Green?eld Investment both local
?rms make pro?ts
t
9
? O
D
in period 1 respectively. In period 2, if a sequential entrant
enters via Green?eld both local ?rms make pro?ts
t
16
?O
D
each, if the sequential entrant
enters via M&A one local ?rm gets his outside option as the takeover target of
t
16
? O
D
and the other local ?rm makes pro?ts
t
16
?O
D
. If no sequential entry occurs, local ?rms
make pro?ts
t
9
? O
D
in period 2. Multiplying with respective probabilities of sequential
entry decisions yields the above domestic pro?t level.
19
So P
3
< P
1
69
where again ?
MA
Dom
(BM) =
13
18
t ?4O
D
is straightforward.
Consumer rent in the case of early entry via M&A is
CR
MA
= 2s ?
3
4
t + (P
2
?P
4
) ×
13
72
t
Therefore welfare in this case is
W
MA
= ?
MA
Dom
+CR
MA
= 2s ?
t
36
?4O
D
?P
2
t
144
?P
4
19
144
t (3.22)
where again 2s ?
t
36
?4O
D
= W
MA
(BM).
A comparison between the welfare e?ect of entry modes between the
benchmark case and the case of sequential entry can now be made.
The welfare di?erence between entry via Green?eld Investment and entry
via M&A in the benchmark case is
?W
BM
= W
BM
(GF) ?W
BM
(MA) =
t
36
so in the benchmark case entry via Green?eld Investment is the welfare
maximizing entry mode, yielding
t
36
higher welfare than entry via M&A.
In the case of sequential entry after some rearranging, the welfare di?er-
ence can be shown to be
?W = W(GF) ?W(MA) =
t
36
?
t
144
[?
31
72
t ?19F ?16O
D
]
. ¸¸ .
0
or
?W = W(GF) ?W(MA) = ?W
BM
?
t
144
[?
31
72
t ?19F ?16O
D
]
. ¸¸ .
0
(3.23)
Therefore the following propositions can directly be stated.
Proposition 3
Early entry via Green?eld Investment is the welfare maximizing mode of
entry, with or without potential sequential entry.
70
Proposition 4
With sequential entry, the welfare advantage of entry via Green?eld In-
vestment over entry via M&A is more pronounced, than without the threat of
sequential entry.
Whereas the ?rst proposition does not come as a surprise, the second
result is not so obvious. Intuitively entry via M&A in the case of potential
sequential entry accommodates further entry. However such accommodated
further entry, besides increasing expected consumer rent, also decreases ex-
pected pro?ts (including acquisition prices) for the domestic ?rms. In the
proposed setup, the second e?ect dominates the ?rst, not least because only
sequential entry via Green?eld Investment increases consumer rent, whereas
both types of entry reduce domestic pro?ts for at least one ?rm
20
.
3.4 Markets with restricted takeover possi-
bilities
Another interesting result can be derived from the model if one changes the
setup such, that only 1 out of 2 incumbents can be taken over. Assume that
only ?rm A can be taken over, while B, due to various possible reasons
21
can not be taken over.
Intuitively, one might at ?rst guess that in this setup the choice of entry
mode of the early mover shifts towards M&A, because by taking over ?rm A
it disables the sequential entrant to enter via M&A, as there are no targets
left in this case. So, by taking away this option from sequential entrants, one
might think ?rm 1 should be better o?. Therefore M&A might seem to be a
more appealing entry mode for ?rm 1 now.
However with symmetric ?rms concerning marginal costs this is not the
case as will be shown. The reason is, that, with symmetric ?rms, sequential
20
The reasoning runs through acquisition prices to be paid. If the sequential entrant
enters via M&A he pays the target its outside option, which is pro?ts the latter would
make if the sequential entrant enters via Green?eld Investment. Therefore the threat of
Green?eld Investment reduces the targeted ?rm’s pro?ts even when being taken over, so
the sequential entrant entering via M&A.
21
e.g. the ?rm might be state-owned with the state having a strategic interest in keeping
the ?rm state-owned, or the ?rm might be family-owned, where the family might not be
interested in selling the ?rm to some entrant due to non-monetary reasons.
71
entry via M&A is harmless to the early mover, whereas Green?eld entry re-
duces early mover pro?ts in period 2. Within the setup of restricted takeover
probabilities now ?rm 1, by choosing to enter via M&A, does not change the
entry consideration of sequential entrants that would have entered via Green-
?eld anyway, but it will push sequential entrants that would have entered via
M&A if a target had been available to enter via Green?eld Investment now.
We show this in the following.
If ?rm F
1
enters via Green?eld nothing changes compared to the base
model, as no additional limitation is introduced for the strategic entry deci-
sion of the sequential entrant.
Sequential entry (mode) probabilities still are P
1
= O
GF
2
=
1
16
t ? F and
P
3
=
O
2
GF
= ?
1
72
t +O
D
+F.
If ?rm F
1
enters via M&A the probability of any entry is still P
2
= O
MA
2
=
1
9
t ?F, but now P
4
= 0, as there is no available target for ?rm 2 left in the
market now.
Analyzing ?rm 1 behaviour there is no di?erence concerning the threshold
level of operative costs for general entry of F
1
.
A look at the threshold level for entry via M&A for ?rm 1 points out the
di?erence to the base speci?cation.
In the base model
¯
O
Base
1
=
1
36
t +O
D
+
F
2
?(P
2
?P
4
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
Without a target for sequential entrants this reduces to
¯
O
Extension
1
=
1
36
t +O
D
+
F
2
+ (0 ?P
2
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t (3.24)
It is straightforward that
¯
O
Extension
1
=
1
36
t +O
D
+
F
2
+ (0 ?P
2
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
<
¯
O
Base
1
=
1
36
t +O
D
+
F
2
+ (P
4
?P
2
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
as the inequality reduces to 0 < P
4
×
13
288
t, which is true for t > 0 and
P
4
> 0, which is true for all markets that in general support entry via M&A.
As O
Extension
1
= O
Base
1
it is also straightforward that the relative probabil-
ity of M&A entry of the early mover ?rm 1
¯
O
Extension
1
O
Extension
1
is lower with restricted
takeover possibility than in the basic model
¯
O
Base
1
O
Base
1
.
72
As the benchmark case is una?ected by the proposed change in the model
setup it then must be true that
Proposition 5
The negative e?ect of the threat of sequential entry on the relative proba-
bility of early entry via M&A is more pronounced, when takeover possibilities
in the market are limited, such that early entry via M&A eliminates the
option of sequential entrants to enter via M&A. In absolute terms such lim-
ited takeover possibilities increase the probability of early movers entering via
Green?eld Investment instead of entering via M&A.
This result is quite astonishing. Whereas general literature discusses ob-
vious e?ects of missing targets on entry mode choice, namely that there is
no possibility for a ?rm to enter via M&A if no target is available for this
?rm, we come up with an additional indirect missing target e?ect, due to
strategic considerations about potential sequential entry taken into account
by an early entrant.
3.5 Country-speci?c learning-by-doing e?ects
A ?nal interesting extension of the model proposed is incorporating the
availability of learning-by-doing e?ects in the market. With country-speci?c
learning-by-doing e?ects it is meant, that in the respective industry increas-
ing sales volume in other countries does not change the e?ciency of a ?rm in
the respective host country market, such that e?ciency of a ?rm in a market
only increases with the volume of former (sales) experience in this speci?c
market.
This extension is considered in a further simpli?ed version of the basic
model. We will see that, if learning-by-doing e?ects are strong for some
kind of industry, M&A is actually the sequential entry deterring mode for
early movers and therefore it is obvious that the threat of sequential entry
leads early movers to rather enter the market via M&A compared to the
benchmark case.
Assume for simplicity that in period T=1 all ?rms have marginal costs
of c = 1. Further assume that incumbents have been in the market for
such a long time that they have already used up all learning-by-doing e?ects
available, such that c
A
= c
B
= 1 in both periods. The potential sequential
entrant does by design not participate in the market inT = 1, so his marginal
costs in T=2 are also c
2
= c = 1.
73
For the early entering ?rm 1 assume that his marginal costs in period 2 are
a function of how much 1 has sold in the market in period T=1. Speci?cally
let us assume
c
1
(T = 2) =
1 if x
1
(T = 1) < 0, 5
0 if x
1
(T = 1) ? 0, 5
This is a very simple, special form of learning-by-doing e?ects, but the
results can be generalized
22
. By assuming this speci?c form we can abstract
from any strategic selling behaviour of ?rm 1 in period 1 as well as reduce
notational clutter.
3.5.1 Sales volumes in period T=1
With ?rms 1, A, B being symmetric in marginal costs and market size equal
to 1 it is straightforward that the sales volume of ?rm 1 in period 1 is x
1
(T =
1) =
1
3
if 1 enters via Green?eld Investment and x
1
(T = 1) =
1
2
if 1 enters
via M&A.
Therefore c
1
(T = 2) = 1 if 1 enters via Green?eld and c
1
(T = 2) = 0 if 1
enters via M&A.
3.5.2 Sequential entry probabilities in period T=2
If F
1
enters via Green?eld Investment results do not di?er from the base
model, as marginal costs of market participants are unchanged.
The probability of any entry as well as the absolute probability of entry
via M&A is still P
1
=
1
16
t ?F and P
3
= ?
1
72
t +O
D
+F respectively.
If F
1
now enters via M&A the sequential entry probabilities di?er from
the base model, due to the additional e?ect of M&A on marginal costs of
?rm 1.
The probability of any entry in T=2 is now
23
22
It is always true that ?rst period sales will be higher for ?rm 1, if it enters via M&A
compared to Green?eld Investment. Therefore its marginal costs in the second period will
always be lower if it enters via M&A compared to Green?eld Investment.
23
One has to stay simple in this analysis, therefore I keep assuming that ?rms will
locate equidistantly from each other, which is a strict assumption, but actually even coun-
terbalances our results, so the assumption does not drive our results. To be concrete, it
is additionally assumed that the sequential entrant locates farthest away from the early
entrant at the other side of the Salop circle with the two local incumbents in between.
74
P
2
=
1
t
_
1
3
t ?
1
5
_
2
?F (3.25)
and the absolute probability of entry via M&A is now
P
4
=
1
t
_
_
1
2
t ?
1
3
_
2
?2 ×
_
1
3
t ?
1
5
_
2
_
+O
D
+F (3.26)
Analyzing probabilities we come up with the main interesting result of
analysis of this extension.
Proposition 6
With learning-by-doing e?ects in the market and product di?erentiation
low early-mover entry via M&A compared to entry via Green?eld deters se-
quential entry. If product di?erentiation is high early-mover entry via M&A
still accommodates sequential entry.
The proposition is proved in Appendix 9.
The intuition is straightforward. If product di?erentiation is of low de-
gree, the base model-e?ect of entry deterrence via Green?eld through an
increase in the number of market participants faced by a sequential entrant
is small, as has been shown before. With learning-by-doing-e?ects entry via
M&A leads to ?rm 1 being able to price more aggressively in period 2, due to
lower marginal costs than under Green?eld Investment. This second ”‘e?-
cient competitor”’e?ect then dominates the ?rst ”‘more competitors”’ e?ect
for su?ciently low levels of transport costs, as a more e?cient competitor
hurts potential sequential entrants in expectations more than a larger number
of competitors.
The following ?gure shows how the respective probabilities of entry and
entry modes depend on ”‘transport costs”’ t in this extension.
75
Figure 3.5: Contingent Sequential Entry Probabilities and
Transport Costs
2 2,1 2,2 2,3 2,4 2,5 2,6 2,7 2,8 2,9 3
t
P
r
o
b
a
b
i
l
i
t
i
e
s
P1
P2
P3
P4
For F=0,01; O
D
=0,05
How the threat of sequential entry then in?uences the entry mode choice
of the early mover is obvious. As the threat of sequential entry already
reduces the acquisition price to be paid by the early mover as well as the
strategic entry deterring e?ect also favouring M&A entry it follows that
Proposition 7
In markets with low degrees of product di?erentiation and strong learning-
by-doing e?ects the threat of sequential entry increases the relative probability
of entry via M&A.
3.6 Conclusion
Modelling the threat of sequential entry makes it possible to discuss more
forward-looking entry mode strategies of potential foreign direct investors.
I come to the conclusion, that when the ”competition” e?ect, due to a
di?erence in the number of competitors in the market is the main di?erence
between entry modes, Green?eld Investment of early entrants compared to
entry via M&A has the pro?t-enhancing e?ect of reducing the probability of
harmful sequential entry by other foreign ?rms. Even though the former en-
try mode might additionally have the negative e?ect of pushing the sequential
entrant’s decision towards Green?eld Investment, the absolute entry deter-
ring e?ect is pro?t-enhancing, as the absolute probability of sequential entry
via Green?eld Investment is reduced.
76
And while acquisition prices will also be lower with the threat of sequential
entry, it is show, that early movers will choose Green?eld Investment with
a higher probability than in a static world without sequential entry, as the
entry deterring e?ect on pro?ts outweighs all other e?ects.
The above result holds true for all markets where both kind of entry
modes are chosen by the group of potential entrants with positive probability,
so if the degree of product di?erentiation, ?xed costs of Green?eld entry and
?xed cost structure of incumbent banks give support to both M&A as well
as Green?eld Investment. Obviously there are markets, where modelling
sequential entry does not change entry mode choices, due to incentives for
one speci?c entry mode being too dominant.
Additionally it is shown in the model, that with sequential entry, the
e?ect of early entry via M&A instead of entry via Green?eld Investment is
even more harmful to welfare than in the static case. The reason is, that
early entry via M&A accommodates further entry of foreign ?rms, which
decreases domestic ?rm pro?ts by more than it increases customer rent.
Slightly changing the setup of the model to account for ”scarcity” of
takeover targets, further shifts early entrants’ incentives towards Green?eld
Investment, due to the e?ect that entry via M&A in this setup takes away
the entry mode option for sequential entrants that is less/not harmful to
early mover pro?ts. I therefore ?nd a ”‘perverse”’ missing target e?ect on
entry mode choice, such that missing targets for other entrants pushes an
early entrant towards entering via Green?eld Investment.
All in all the proposed basic model helps explain the still very signi?cant
share of Green?eld Investment in total FDI (see e.g. Ra?, Ryan and Staehler
(2006)[113] via the market structure e?ect of entry modes and the latter e?ect
on entry dynamics.
Extending the model to account for country-speci?c learning by doing
e?ects in a stylized way, it is found that in such a setting the early mover
can become a ”terrifyingly” e?cient competitor in period 2 by entering via
M&A early on. The e?ect of becoming a stronger competitor via M&A
then dominates the ”competition” e?ect of reducing the number of market
participants by M&A concerning sequential entry probability for some kind
of markets. It can be shown then, that in this extended setup M&A is the
entry deterring early entry mode if the respective market shows a low degree
of product di?erentiation and if country-speci?c learning-by-doing e?ects are
su?ciently strong (like they are modelled here).
77
Referring to the general scope of this thesis, I deem this latter extension
to be a very good ?t for international retail and SME loan markets. Indeed
one can probably state, that product di?erentiation concerning ?nancial ser-
vices for these types of customers is rather low
24
. Also, due to heterogeneous
(banking) regulation in countries, banks (and also probably insurance com-
panies, for that matter) compared to ?rms in e.g. the real sector, might
be less able to transfer knowledge won by learning-by-doing in one market
to other markets. So e?ciency in a speci?c market in the former industries
should mainly depend on the cumulative volume of business in exactly this
market, and not so much on general world-wide level of experience. The fact
that heterogeneous regulation in part prohibits banks from making use of
other market experience in a respective host country market has been shown
by various studies of e.g. Berger et al.(2000)[12], who show that, except U.S.
banks in some speci?c markets, foreign banks are almost always less e?cient
than domestic banks in OECD countries. A similar story is discussed in the
case study, stating that general international market experience only helps
in conduction business in a speci?c country, if the other markets mimic this
speci?c market.
24
At least this should be true ex ante. Ex post there might be some product di?erenti-
ation stemming from an existing lending relation with a respective customer.
78
APPENDIX
Appendix 1: Deriving variable pro?ts
For starters I assume that willingness to pay s for the product is such,
that in equilibrium the whole market Y is served by the ?rms in the market.
In general sales volume of ?rm i can be derived by the indi?erence condition
of the marginal customer of the respective ?rm. The condition yields
x
i
=
p
j
+p
k
?2p
i
2t
+
Y
n
where p
j
,p
k
are the prices of the two closest competitors and n is the
number of ?rms in the market.
Inserting into ?rm i pro?t function yields
?
V ar
i
= (p
i
?c
i
) ×
_
p
j
+p
k
?2p
i
2t
+
Y
n
_
Solving for the optimal price and taking into consideration that all ?rms
will charge the same price due to same marginal costs c we get equilibrium
price charged by all ?rms
p =
Y
n
×t
Inserting into the sales volume we get equilibrium sales volume for all
?rms
x =
Y
n
Therefore variable pro?t for each ?rm is
?
V ar
=
_
Y
n
_
2
×t
With market size 1 and potential number of ?rms ranging between 2
(both foreign entrants enter via M&A or no entry at all) and 4 (both for-
eign entrants enter via Green?eld) the relevant variable pro?t levels for the
analysis are
?
V ar
(2) =
1
4
×t
?
V ar
(3) =
1
9
×t
?
V ar
(3) =
1
16
×t
79
Appendix 2: Proof of Lemma 1
The di?erence between M&A and Green?eld pro?ts for a respective bank
and a respective entry mode of another bank is of the form
= ?
V ar
i
(MA) ??
V ar
i
(GF) ?A +F ?2O
i
+O
i
=
?
V ar
i
(MA) ??
V ar
i
(GF) ?A +F ?O
i
So
d
dO
i
= ?1 < 0
which means the larger O
i
so M&A is preferred over Green?eld Investment
for low values (as ?
V ar
i
(MA) ??
V ar
i
(GF) ?A > 0 for the proposed setup) of
O
i
and vice versa for higher values of O
i
.
Green?eld pro?ts are of form
?
V ar
i
(GF) ?O
i
so
d
dO
i
= ?1 < 0 and entry becomes less likely if O
i
is high.
Appendix 3: Proof of Proposition 1
I want to show that
P
1
< P
2
so
1
16
t ?F <
1
9
t ?F
which reduces to
7
144
t > 0
which is true for t > 0,which is ful?lled by assumption.
The di?erence between P
2
and P
1
is
1
9
t ?
1
16
t =
7
144
t
Appendix 4: Proof of Lemma 2
I want to show that
80
P
3
< P
4
which is equal to P
3
?P
4
< 0. Inserting yields
?
1
72
t ?
1
36
t = ?
1
24
t < 0
true for t > 0, which is true by assumption.
The absolute di?erence between probabilities is P
4
?P
3
=
1
24
t.
Appendix 5: Proof of Lemma 3
We want to show that
P
2
?P
4
> P
1
?P3
equal to
1
12
t ?2F ?O
D
>
11
144
t ?2F ?O
D
which reduces to
1
144
t > 0
which is true for t > 0. which is given by assumption.
The di?erence [P
2
?P
4
] ?[P
1
?P3] =
1
144
t.
Appendix 6: Proof of Lemma 4
Solving pro?t inequality
?
1
(MA) = 2 ×
_
5
18
t ?2O
1
+O
D
_
+ (P
4
?P
2
) ×
13
144
t
>
?
1
(GF) = 2 ×
_
1
9
t ?O
1
?
F
2
_
?(P
1
?P
3
) ×
7
144
t
for O
1
yields
O
1
<
1
36
t +O
D
+
F
2
?(P
2
?P
4
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
Solving inequality ?
1
(GF) > 0 for O
1
yields
O
1
<
1
9
t ?
F
2
?(P
1
?P
3
) ×
7
288
t
81
Appendix 7: Proof of Lemma 5
It should be shown that an early mover with some overhead ?xed costs
O
1
would enter the market without the threat of sequential entry, but will
not enter if the threat of sequential entry exists. This is true i?
O
BM
1
> O
1
Inserting yields
(P
1
?P
3
) ×
7
288
t > 0
By the pecking order of entry modes we know that P
1
> P
3
as long as
the market structure in general supports both kinds of entry. As additionally
t > 0 it must hold true that RHS > 0.
Appendix 8: Proof of Proposition 2
I want to show that some potential entrant ?rm 1 with operating ?xed
costs of O
1
would enter via M&A without the threat of sequential entry and
via Green?eld if the threat of sequential entry exists. Therefore we got to
show that
¯
O
BM
1
O
BM
1
>
¯
O
1
O
1
Inserting in the RHS yields
¯
O
BM
1
O
BM
1
>
¯
O
BM
1
+
[
(P
4
?P
2
)×
13
288
t+(P
1
?P
3
)×
7
288
t
]
O
BM
1
+(P
3
?P
1
)×
7
288
t
which can be simpli?ed to
(P
3
?P
1
) ×7 ×
¯
O
BM
1
> [(P
4
?P
2
) ×13 + (P
1
?P
3
) ×7] ×O
BM
1
for markets where entry is feasible (O
BM
1
+ (P
3
?P
1
) > 0)
Inserting for probabilities and benchmark threshold levels after some ma-
nipulation yields the following inequality
239
18
t
2
?(428F + 290O
D
) t + (2016O
2
D
+ 5904FO
D
+ 3744F
2
) > 0
Analyzing the LHS yields
82
?LHS
?t
< 0 for t <
3852
239
F +
2610
239
O
D
?LHS
?t
= 0 for t =
3852
239
F +
2610
239
O
D
?LHS
?t
> 0 for t =
3852
239
F +
2610
239
O
D
as well as LHS(t = 0) = 2016O
2
D
+ 5904FO
D
+ 3744F
2
> 0
and LHS(t ??) ??.
So, the LHS has a minimum at t =
3852
239
F +
2610
239
O
D
.
Therefore it is su?cient to show that the inequality LHS > 0 holds for
this minimum.
Inserting the minimizing t-level yields
70488
239
F
2
+
294776
239
FO
D
+
103374
239
O
2
D
> 0
which is ful?lled by F > 0 and O
D
> 0.
Appendix 9: Proof of Proposition 6
I show that P
2
< P
1
for su?ciently low levels of t. P
2
< P
1
i?
P
2
=
1
t
_
1
3
t ?
1
5
_
2
?F <
t
16
?F
which reduces to inequality
7
144
t
2
?
2
15
t +
1
25
< 0
which is ful?lled for t < 2, 4. For t > 2, 4 it is still true that P
2
> P
1
.
83
Chapter 4
When do Banks Follow their
Customers Abroad?
4.1 Introduction
Among the often-stated motives for foreign direct investment (FDI) in the
banking sector, besides the classical market-seeking reasoning, is a bank’s
desire to follow its existing customer base abroad. This motive is well-
established both in the economic as well as the business literature (e.g. Aliber
(1984)[3],Casson (1990)[31], Williams(1997)[130] and Bain, Fung and Harper
(1999)[5]). As Casson (1990) states
”‘... US banks capitalize on their goodwill by following their
customers overseas; the multinationalization of manufacturing
?rms creates a derived demand for the multinationalization of
banks as well”’
Nolle and Seth (1996)[108] cite a study conducted by the U.S. General
Accounting O?ce
1
reporting that in the United States ”‘most foreign banks
serve customers of their home countries. An industry representative told us
that only a few banks are large enough to penetrate through home country
loyalties to attract other customer”’.
1
”‘Foreign Banks: Assessing their Role in the U.S. Banking System”’, Report to the
Ranking Minority Member, Committee on Banking, Housing and Urban A?airs, U.S.
Senate, GAO/GGD-96-26 (1996)[110]
84
Though this latter claim might be too strong
2
, the general statement
drives home the point, that follow your customer(FYC)-considerations are
an important reason for banks to establish a physical presence abroad.
In some instances existing clients even actively lobbied for their respec-
tive primary banks to follow ?rm expansion abroad. Well-documented and
often-stated early examples were U.S. multinational companies US Steel and
DuPont urging Citibank to establish a foreign presence in South America to
provide their local operations with banking services (e.g. Huertas (1990)[80]).
A large empirical and business/case study literature deals with this topic
and indeed ?nds strong indications that this motive plays a signi?cant role
in the multinationalization decision of banks.
In one of the earliest studies Fieleke (1977)[58], in his study on the de-
terminants of U.S. banks’ overseas expansion, found that ”‘?nancial need
of U.S. ?rms abroad”’ was the major factor in U.S. banks foreign location
choice. This early result for U.S. banks has been supported by numerous
studies, e.g. Nigh, Cho and Krishnan (1986)[106].
A similar role for follow your customer motives in the foreign direct in-
vestment decision of multinational banks has been found for foreign banks
entering the U.S. market by e.g. Hultmann and McGee (1989)[82] and Gold-
berg and Grosse (1994)[68] and for bank entry into less developed markets
(Sabi (1987)[116]). For example, Goldberg and Grosse (1994) study bank
sector foreign direct investment in respective U.S. states. They come up
with evidence, that states, that attract a large volume of real sector foreign
direct investment, also attract more bank FDI.
Nolle and Seth (1996)[108] also analyze the U.S. banking market. They
?nd indication that follow your customer strategies indeed seem prevalent
in foreign bank strategy in the U.S. banking market, as evidence points to
foreign banks devoting the dominant part of total extended loans to foreign
real sector ?rms. However their approach yields the some indication, that
this strategy might not be the main reason for foreign banks entering the U.S.
market. Then again, in contrast to some recent remarks in the industry, they
discover, that follow your customer strategies are still at least as important
in the entry decision of foreign banks as in previous times
3
.
2
As at least some banks seem able to attract local customers, see e.g. Berger et
al.(2000)[12].
3
The authors ?nd, that the share of loans to foreign ?rms as a percentage of total loans
of foreign banks in the U.S., after decreasing for some time, had reached and partly even
exceeded former levels again in the 1990s.
85
Even though the topic is has drawn a lot of attention in the empiri-
cal literature, there is a lack of formal theory on this subject. This seems
bothersome, especially when one considers the direction of argumentation
of a growing literature on market-seeking bank foreign direct investment.
This literature is strongly concerned with information asymmetries and prob-
lems of foreign banks when trying to serve local markets (e.g. Dell’Arricia,
Friedman and Marquez (1999)[47], Dell’Arricia and Marquez (2004)[48] and
Lehner (2007)[92]). These theories might not yield a su?cient explanation
for bank FDI that is induced by the follow your customer-motive.
Particularly the question arises, why a home bank, with an existing rela-
tionship to a home country multinational ?rm, does not simply provide bank-
ing services cross-border, or indirectly via providing the loan to the parent
company from its home base, to the multinational ?rm’s foreign subsidiary.
Physical transport costs seem negligible for loans and similar ?nancial ser-
vices, and informational requirements to provide a loan might already be
met by the bank’s general intimate knowledge about the client ?rm due to
previous and ongoing interaction.
Whereas there might still be motivation for following clients to assist in
local cash management and other services requiring face-to-face contact
4
, the
question is whether there also is such motivation concerning the provision of
loans.
In the following, a model is proposed to motivate such latter follow your
customer-behaviour by applying and re?tting a well-established theoretical
literature on the choice of type of ?nancing on this speci?c topic.
The following model setup is based on a view ?rst clearly laid out by Gert-
ner, Scharfstein and Stein (1994)[65], who discuss the bene?ts and costs of
debt ?nancing compared to internal ?nancing in a setup of a two-dimensional
moral hazard problem (managerial e?ort and managerial discretion in di-
verting project payo?s to himself) faced by a ?rm conducting a manager-run
investment project.
Marin and Schnitzer (2006)[97] broaden the scope of such analysis by
introducing a geographic dimension of debt ?nancing in the ?nancing decision
of a multinational ?rm setting up a manager-run subsidiary in a foreign
country. Letting liquidation e?ciency of banks di?er exogenously by their
proximity to the respective investment project, they ?nd that, depending
on project/?rm characteristics, multinational ?rms will either use host or
home/third country bank ?nancing (or ?nancing from internal sources) for
4
However the question then is, whether sales volume of such services are su?cient for
banks to have an incentive to enter a foreign country.
86
their FDI projects.
The trade-o? faced by the investing ?rm here is, that high liquidation e?-
ciency of the chosen bank type on the one hand allows the investing multina-
tional ?rm to capture a larger share of ex post project returns if the project is
successful, as well as in some cases being able to claim a larger excess liquida-
tion value (liquidation value minus debt repayment) in case the project fails.
On the other hand, however, managerial incentives to spend e?ort, therefore
increasing the expected size of the pie (larger expected project payo?s) to be
shared, are negatively a?ected by high liquidation e?ciency, as the manager
expects a lower share of project returns to be available to himself due to
a better outside option of the investing ?rm in negotiations about sharing
project continuation value.
I closely follow this main idea in this paper
5
. However, in order to in-
corporate strategic choice of domestic banks into the model, I allow for this
home bank to potentially make non-zero pro?ts as well as having discretion
in locational choice, or to put it simply, for home country banks to establish a
physical presence in the ?rm subsidiary’s host country. That way the respec-
tive bank has discretion in a pro?t-determining choice of the loan provision
mode, as the decision then faced by this bank is to whether establish a phys-
ical presence abroad or to serve the client from its home base. As the focus
will be on this conditional location decision faced by the bank, the model
abstracts from cases discussed in Marin and Schnitzer (2006)[97], where the
parent ?rm can potentially completely ?nance this FDI project from internal
cash ?ow or wealth.
Additionally the location-speci?c liquidation e?ciency of a bank is endo-
genized. This is ?rst done in a very simple fashion, where I lay out the fact
that physical transport costs might after all play a role in the decision of how
to supply a loan, as physical assets to be liquidated of a bank’s unsuccessful
client abroad might have to be transferred back to the bank’s home country
to sell these assets at a high price. With the help of a simple political econ-
omy story transport costs are then further endogenized by introducing the
ability of host country governments to (partially) keep assets to be liquidated
in the home country, restricting cross-border physical asset ?ow. In a sim-
pli?ed setting the respective government’s incentive to do so are discussed,
contingent on the country’s endowment in human capital. I do so by using
the notion of asset-embedded human capital, that therefore is immanent in
5
I’d really like to thank Prof. Marin for pointing out this paper as a possible starting
point for discussing bank location choice.
87
the liquidation value of the respective project
6
.
Introducing bank location choice as well as country-dependent liquida-
tion values, allows an analysis on why and under which circumstances banks
engage in foreign direct investment induced by the follow your customer-
motive.
The reason for engaging in follow your customer-bank FDI is then a po-
tentially di?erent liquidation e?ciency attained by the bank, in comparison
to the bank supplying the loan to its multinational client without a physical
presence in its client host market. The model therefore o?ers the possibility
to discuss project/?rm-speci?c and host country-speci?c optimal provision
modes for the respective domestic bank, showing that these two dimensions
of contingency are intertwined in shaping the location choice of the bank.
Using basic results attained from the model, the argumentation of Marin
and Schnitzer (2006)[97] concerning the link between foreign direct invest-
ment and international capital ?ows can be scrutinized.
The rest of this chapter is organized as follows. Section 2 lays out the
basic model, analyzing ?rm level, bank level and government level decision
making to at the end come up with the optimal loan provision modes from
the bank’s perspective for respective circumstances. Section 3 concludes.
First, I discuss the empirical observations the model can explain. Second, I
analyze how my results compare to those obtained by Marin and Schnitzer
(2006)[97]. Finally I point to the potential obstacles to exposing the theory
to an empirical test.
4.2 The Model
The following setup is considered. There are two countries, Home and For-
eign. A ?rm (investor) from Home considers entering the market in Foreign
via the establishment of a subsidiary in this host country. In order to do so,
the ?rm has to hire a manager to operate the subsidiary. Additionally the
?rm is cash-strapped in so far, as it is not able to ?nance this foreign direct
investment completely via internal funds. Therefore the ?rm has to take on
a loan of size K from a bank.
The foreign direct investment project of the ?rm analyzed is a two-period
project, yielding return X
1
in period 1 with probability p and 0 with proba-
bility 1?p and return X
2
in period 2 with certainty. The subsidiary manager
controls the success probability p in period 1 by his choice of e?ort level which
6
A similar notion is proposed in a chapter of the upcoming PhD. thesis by Yanhui Wu,
University of Munich.
88
equals the probability of success. The costs of e?ort are assumed to be of
the quadratic form C(p) =
1
2
zp
2
with z ?]0; ?].
The problem from the point of view of the parent ?rm is, that neither the
e?ort level of the manager, nor the project returns in the respective periods
are veri?able. Therefore the choice of e?ort level cannot be in?uenced by
the investor ?rm directly via an e?ort-based contract. Additionally the ?rm
needs to give indirect incentives to the manager to at least partially transfer
project payo?s back to the parent company.
Following Marin and Schnitzer (2006)[97] the two incentive problems can
be called the e?ort problem, and the repayment problem, where the former
must be solved by the parent ?rm by to maximize the expected return of
the FDI project, while the latter must be solved to maximize its share in the
expected return.
The combination of the other stakeholders in the subsidiary, bank and
parent ?rm, has two means to in?uence the actions of the manager. For one,
the parent has the ability to monitor the project closely, therefore being able
to make the payo? partially veri?able. That way it is able to capture a share
? of period 1 project payo?. However, monitoring comes at a cost which is
assumed to be of the quadratic type C(?) =
1
2
?
2
.
Additionally the bank has the right to liquidate the project after period
1, if the loan is not paid back. Precisely, the parent ?rm has to pay back the
demanded repayment D
i
to the bank, else the bank will liquidate. Following
(implicitly) Marin and Schnitzer (2006)[97] it must be the case, that only the
parent ?rm, not the local subsidiary manager has the right to pay back the
loan to the bank
7
.
A negotiation stage between manager and investing parent ?rm, after
period 1 payo?s are realized and before the credit repayment is due, is mod-
elled. If the parent ?rm does not repay the loan to the bank the project
is liquidated by the bank. As long as it is e?cient to continue the project
(X
2
> L
i
), as will be assumed in the following analysis, the parent ?rm and
the subsidiary manager can bargain over the continuation value, leaving both
stakeholders better o? ex post by continuing. Furthermore it is assumed, that
the bargaining power in such a renegotiation is exogenous, with the investing
7
When latter discussing the model it becomes obvious that when the parent ?rm could
decide on whether to repay the loan itself or to make the subsidiary a relatively independent
capital center it would always choose the former. As we will implicitly see, the reason is
that if the manager could repay the loan himself the investor would not be able to extract
anything more than net ? ×X
1
from the manager, therefore not participating in period 2
pro?ts at all.
89
parent ?rm able to claim a fraction 0 < ? < 1 of the continuation value and
the manager therefore able to claim (1 ??) of it.
The banking market structure is as follows. There is a bank monopolist
in the home market, which has an existing relationship with the above men-
tioned ?rm
8
. In the foreign market a large number of operating homogeneous
banks is assumed, such that there is perfect competition between these banks
leading to zero pro?ts for them in equilibrium.
The domestic bank can di?er from the foreign banks in terms of the ability
to e?ciently liquidate the ?rm’s subsidiary if the loan is not paid back. Bank
ex post liquidation value for the client ?rm’s subsidiary is denoted L
d
for the
domestic bank and L
f
for all foreign (local) banks
9
.
Within this setting the domestic bank can decide on whether to set up a
subsidiary abroad itself (follow its customer), or try to win the loan contract
and extend the loan without such a physical presence abroad
10
. The dif-
ference between the two modes of loan provision will be potentially di?ering
liquidation values the bank can achieve, which will be discussed in greater de-
tail below. Finally, if the bank decides to follow its customer abroad it incurs
small ?xed costs of setting up an agency or other organizational structure of
size F
11
.
The structure of the model can be subsumed in the following ?gure.
8
Actually one would not need to assume only one bank being active in the home market,
but from its existing relationship with the respective ?rm the bank analyzed might be able
to act as a monopolist concerning this ?rm due to various reasons.
9
As bank e?ciency should be shaped by characteristics of its main market, the assump-
tion, that banks stemming from the same country should, all else equal, be more similar
to each other concerning e?ciency than banks from other markets, seems reasonable.
10
The latter simply means, that the bank ?nances a project abroad without establishing
a local physical presence there, but does so by either actually giving the loan to the
domestic parent ?rm, which then internally transfers the capital to its subsidiary, or by
extending the loan to the subsidiary cross-border, where in the latter case the contract
still must call for repayment by the parent ?rm only as well as excess liquidation value
in case of failure falling back to the parent ?rm directly. Again this follows the implicit
setup by Marin and Schnitzer (2006)[97].
11
In the remaining paper these ?xed costs will not be discussed explicitly, only func-
tioning as a tiebreaker between loan provision modes when the bank is indi?erent else.
90
Figure 4.1: Time Structure of the Model
Firm
announces
interest in
investing
abroad
Domestic
bank
decides
upon going
abroad
Investor
chooses
bank, hires
manager
X
1
realized,
Negotiation
about return
sharing
Investor
chooses ?,
Manager
chooses p
Bank loan
due
Repayment
Liquidation
X
2
realized
4.2.1 Bank pro?t maximization under given liquida-
tion value
In a ?rst step the domestic banks pro?t maximization decision for a given
liquidation value it can achieve is analyzed.
Two possible cases have to be distinguished, the case of a moderately
cash-strapped ?rm and the case of a severely cash-strapped ?rm. From the
perspective of the bank, the ?rst case translates into a small loan (small K)
compared to the total size of the investment project, whereas in the latter case
the loan size would be large compared to the investment volume (large K).
As the liquidation value should depend on the total size of the investment
project, di?erent relative
K
L
i
result, which lead to completely di?erent risk
structures between these two types of ?rms from a bank’s perspective.
The case of a moderately cash-strapped ?rm
Let us ?rst consider the case of a relatively small loan, such that for all banks
L
i
> K, therefore bank i gets back at least the face value of the loan in all
circumstances, even when the project fails. The expected pro?t for bank i is
then
?
i
= ´ pD
i
+ (1 ? ´ p)min[L
i
; D
i
] ?K (4.1)
where ´ p is the expected success probability, equalling the e?ort level cho-
sen by the manager in equilibrium, of the investment project, depending, as
91
we will see, on a bank’s liquidation e?ciency L
i
.
Financing by a foreign (local) bank
As the local banks are symmetric concerning the liquidation value L
f
they
will demand the same repayment D
f
in equilibrium
12
. As the liquidation
value in this case of a small loan always su?ces to repay at least the loan
size, the project is riskless to all banks. Note also, that even in the case the
liquidation value is very large (L
i
> D
i
), the bank will only be allowed to
keep D
i
and give the excess liquidation value (L
i
?D
i
) to the investing parent
?rm. Therefore the local banks will compete themselves down to demand a
repayment of D
f
= K, such that demanded repayment equals the size of the
loan and these banks make zero pro?ts in expectation
13
.
In order to analyze the e?ort and monitoring choices inside the ?rm,
which determine rents to be distributed, the problem is solved by backward
induction.
If the project is not liquidated before, the project yields a payo? of X
2
in period 2. Due to the non-veri?ability of the payo?, the manager can keep
the whole payo? to himself. Also, there is no more control right the other
stakeholders can use to give the manager an incentive to hand over part of
this payo? ex post.
At the end of period 1, at the negotiation stage between parent ?rm and
subsidiary manager, two possible cases are to be distinguished. If the project
failed in period 1, yielding a return of 0, which happens with probability
(1 ? p
i
) no payo?s can be transferred from the manager to the parent ?rm.
Therefore the loan can not be repaid to the bank by the parent ?rm and the
bank will liquidate the foreign subsidiary. In this case the manager does not
get any rent from the project and bears his e?ort costs of C(p
i
). The invest-
ing parent ?rm’s payo? in this case is the excess liquidation value (L
i
?D
i
)
it gets from the bank minus its cost of monitoring C(?
i
) =
1
2
?
2
i
. Obviously,
monitoring goes to waste, if the project is not successful. If however the
project is successful in period 1, which happens with probability p
i
, yielding
a payo? of X
1
, the manager can pay the investor part of period 2 payo?s ac-
cruing to him from his share of the ?rst-period payo?, to prevent liquidation
12
This symmetric equilibrium is a standard result in such a game of Bertrand Compe-
tition with perfectly symmetric ?rms.
13
As L
i
> K it is straightforward that D
i
= K leads to zero pro?ts in expectation. If
any foreign bank would demand repayment K+ another foreign bank could win the loan
contract and make positive pro?ts by e.g. demanding repayment K +
2
.
92
of the project by enabling the investor to repay the loan. If the negotiation
is not successful, leading to liquidation of the subsidiary, the parent ?rm gets
?
i
X
1
+ (L
i
? D
i
) ?
1
2
?
2
i
and the manager’s payo? is (1 ? ?
i
)X
1
?
1
2
zp
2
i
. As
e?ort and monitoring levels have already been chosen at the stage of nego-
tiation, therefore also costs of e?ort and monitoring sunk at this stage, the
negotiation-relevant outside option of the parent ?rm is therefore (L
i
? D
i
)
and 0 for the subsidiary manager, respectively. As continuation is assumed
to be e?cient we will have the ?rm and the manager getting their outside
option plus their share of the continuation value via renegotiation
14
.
Altogether the respective expected payo?s given ?nancing via a foreign
bank are then
E
I
(f) = K+p
f
[?
f
X
1
+(L
f
?D
f
)+?(X
2
?L
f
)]+(1?p
f
)[L
f
?D
f
]?
1
2
?
2
f
(4.2)
for the parent ?rm and
E
M
(f) = p
f
[(1 ??
f
)X
1
+ (1 ??)(X
2
?L
f
)] ?
1
2
zp
2
f
(4.3)
for the subsidiary manager.
Foreign banks will compete themselves down to a required require repay-
ment
D
f
= K. Inserting for D
f
into (3.2) then yields
14
Note however, that two additional constraints have to be ful?lled to prevent ine?cient
liquidation. For one, the parent ?rm has to be able to extract enough repayment from
the manager to be able to repay the loan, formally ?
i
X
1
+ (L
i
?D
i
) + ?(X
2
?L
i
) ? D
i
or ?
i
X
1
+ ?X
2
+ (1 ? ?)L
i
? 2D
i
, which is ful?lled for the investment project being
su?ciently pro?table and parent ?rm bargaining power vis-a-vis its manager su?ciently
high (note that with e?cient continuation X
2
> L
i
, so
??X
2
+(1??)L
i
??
> 0). Additionally,
the manager’s share of period 1 pro?ts must be su?ciently high to be able to transfer
the above payment to the parent ?rm, formally (1 ? ?)X
1
> (L
i
? D
i
) + ?(X
2
? L
i
) or
D
i
> (1??)L
i
+?X
2
?(1??)X
1
. Note that the latter constraint does not present an upper
bound to required repayment levels of banks! I assume both of these constraints to be
ful?lled, because else the project would never be continued (if these constraints would not
at least hold for one type of bank i). I check all upcoming results on whether they interfere
with these constraints but only report when they do so. The implicit constraints will not
play any role in the qualitative results but are just stated for the sake of completeness.
93
E
I
(f) = p
f
[?
f
X
1
+L
f
+?(X
2
?L
f
)] + (1 ?p
f
)L
f
?
1
2
?
2
f
(4.4)
To ?nd the equilibrium payo? for the parent ?rm one has to analyze the
e?ort and monitoring choice of manager and parent ?rm respectively.
In appendix 1 the equilibrium e?ort level chosen by the manager under
?nancing by a foreign (local) bank is derived.
p
f
=
X
1
+(1??)(X
2
?L
f
)
z+X
2
1
which is larger zero if project continuation is e?cient, and
?
f
=
X
1
[X
1
+(1??)(X
2
?L
f
)]
z+X
2
1
is the equilibrium monitoring level chosen by the parent ?rm
15
.
The intuition behind the above e?ects of some right-hand side variables
are quite obvious. (1 ? ?) denotes the bargaining power of the manager
in negotiating sharing the second period payo?. The higher the manager’s
expected share in second period payo?, the higher is his willingness to spend
e?ort, which leads to a higher probability of the project to continue into
the second period. From the perspective of the parent ?rm, this leads to
the choice of a higher monitoring level. The reason is that the investor ?rm
can now take a large share of period 1 payo? for himself without harming
the managers’ incentive to spend e?ort too much, as the latter is su?ciently
incentivised to do so by his high share of second period payo? when the
project is continued
16
. Also, a higher continuation value, due to similar
reasoning, induces the manager to spend more e?ort and the parent ?rm to
15
Before discussing these intermediate results, one should note, that in the following
the analysis for moderately cash-strapped ?rms (due to the structure of the analysis this
restriction does not play a role for the analysis of severely cash-strapped ?rms) has to be
restricted to levels of variables, such that p
i
< 1 and ?
i
< 1,as neither pro?t shares nor
probabilities can be allowed to exceed 100%. Evaluating the inequality at the equilibrium
levels of the former variables, we can state this restriction e.g. via the de?nition of su?cient
levels of e?ort cost parameter z. So the analysis is restricted to e?ort costs and project
payo? characteristics such that
z ? Max[(1 ??)(X
2
?L
i
)X
1
; X
1
(1 ?X
1
) + (1 ??)(X
2
?L
i
)].
16
In absolute terms ?
f
is still restricted, as the manager must still have su?cient liquidity
(1 ??
f
)X
1
to induce the parent ?rm to not let the project get liquidated by the bank.
94
monitor more. A higher ”‘marginal”’ e?ort cost z leads both the manager
to spend less e?ort and the parent to monitor less. Whereas the former
is obvious, the latter stems from the fact that, with higher e?ort costs, the
parent ?rm would destroy e?ort incentives of the manager too much by taking
away too large a share of ?rst-period pro?ts.
Financing by the domestic bank
The domestic bank is allowed to generally di?er from the local foreign
banks by the liquidation value L
d
it is able to generate. As the payo? of
the parent ?rm, who chooses the bank it wants to work with, positively
depends on the equilibrium monitoring and e?ort level ?
i
and p
i
, parent
?rm payo? directly and indirectly depends on the liquidation value of the
bank used in ?nancing. Intuitively a higher liquidation value increases the
expected payo? of the parent ?rm when the project fails in the ?rst period,
due to a higher excess liquidation value falling back to the parent ?rm, and
also enhances the parent ?rm’s position in bargaining over second period
pro?ts. Additionally, a higher liquidation value however reduces the e?ort
the manager will spend for the project, as the outside option of the parent
?rm in bargaining over second period payo?s increases, leaving less payo? of
continuing for the manager in expectation.
As the domestic bank can have a liquidation value di?ering from the local
foreign banks it might be able to demand a higher loan repayment, therefore
generating positive pro?ts for the bank. For this bank it need not be true
that D
d
= K.
The expected payo? of the parent ?rm under ?nancing by a domestic
bank is
E
I
(d) = p
d
[?
d
X
1
+(L
d
?D
d
)+?(X
2
?L
d
)]+(1?p
d
)(L
d
?D
d
)+K?
1
2
?
2
d
(4.5)
which reduces to
E
I
(d) = p
d
[?
d
X
1
+?(X
2
?L
d
)] + (L
d
?D
d
) +K ?
1
2
?
2
d
(4.6)
and the expected payo? of the manager is
E
M
(d) = p
d
[(1 ??
d
)X
1
+ (1 ??)(X
2
?L
d
)] ?
1
2
zp
2
d
(4.7)
The payo? maximizing e?ort and monitoring levels are of identical struc-
ture to the case of ?nancing by a local foreign bank, the only di?erence being
the di?erent liquidation values, formally:
95
p
d
=
X
1
+(1??)(X
2
?L
d
)
z+X
2
1
and
?
d
=
X
1
[X
1
+(1??)(X
2
?L
d
)]
z+X
2
1
Note that ?
d
and p
d
are (negatively) depending on the respective banks’
liquidation e?ciency
17
, but not on the size of repayment demanded by the
domestic bank.
The optimal choice of the repayment size for the domestic bank
One can now derive what the optimal and feasible repayment for the
domestic bank looks like. As the size of repayment D
d
does not in?uence the
equilibrium e?ort and monitoring level, it is true that the optimal demanded
repayment D
d
equals the maximum feasible (contract-winning) repayment
D. When choosing the level of required repayment the domestic bank has to
take into account that, if it requires too large a repayment, the parent ?rm
will instead choose ?nancing by a local bank. So the domestic bank has to
choose D
d
such that it leaves the parent ?rm with its outside option, which is
the expected payo? for the latter when ?nancing the FDI project via a local
bank. Therefore the maximum feasible required repayment D can be derived
from the condition under which the parent ?rm is indi?erent between using
the domestic or a local bank, formally
p
d
[?
d
X
1
+?(X
2
?L
d
)] + (L
d
?D) +K ?
1
2
?
2
d
=
p
f
[?
f
X
1
+?(X
2
?L
f
)] +L
f
?
1
2
?
2
f
Solving the indi?erence condition for D yields
D =
p
d
[?
d
X
1
+?(X
2
?L
d
)] ?p
f
[?
f
X
1
+?(X
2
?L
f
)]
. ¸¸ .
I
+(L
d
?L
f
)
. ¸¸ .
II
?
1
2
(?
2
d
??
2
f
)
. ¸¸ .
III
+K
(4.8)
17
One can easily see that both ?
d
and p
d
negatively depend on L
d
, as
??
d
?L
d
= ?
(1??)X
1
z+X
2
1
<
0 and
?p
d
?L
d
= ?
1??
z+X
2
1
< 0.
96
The ?rst part (I) of the RHS captures the e?ort and repayment e?ect
of choosing a domestic instead of a local foreign bank. The second part (II)
captures the direct e?ect of excess liquidation value when choosing a domestic
over a foreign bank. The third part (III) shows the e?ect on monitoring costs,
due to di?erent equilibrium monitoring levels chosen by the ?rm, for di?erent
bank types.
Obviously if the domestic bank generates the same liquidation value as
the local banks, it can at the maximum require a repayment that leads to
zero pro?ts for the bank in expectations D = K
18
.
As I later want to discuss the choice of the domestic bank between follow-
ing its customer or not, which will a?ect the liquidation e?ciency L
d
of the
bank, one ?rst has to understand how the maximum requirable repayment
for the bank depends on its liquidation value
19
.
Di?erentiating D with respect to L
d
, as is done in appendix 2, one ?nds
dD
dL
d
= 1 ?? (4.9)
with ? = p
d
? ?
?p
d
?L
d
[?
d
X
1
+?(X
2
?L
d
)]
Intuitively, the maximum repayment premium D?D
f
over foreign banks
the domestic bank can demand equals the investing ?rm’s willingness to pay
for the domestic bank’s di?erent liquidation e?ciency. The ?rm in general
looks for a bank generating a relatively high liquidation value if the resulting
bene?t of excess liquidation value and an induced higher feasible share in
payo?s exceeds the disadvantage of reducing the manager’s incentive to spend
e?ort. This is the case if ? < 1. If the higher liquidation value reduces the
manager’s e?ort too much, the parent ?rm would rather like to work with
a bank that achieves a relatively lower liquidation value. This is the case if
? > 1.
18
This can be checked easily: If L
d
= L
f
, then ?
i
and p
i
are also the same for both
types of banks. The equation boils down to D = K = D
f
.
19
The constraint that the parent ?rm has to get a transfer su?ciently high to be able to
pay back the ?rm could potentially be binding here, however only in a quantitative way
and only in the case of very large di?erences in the liquidation e?ciency of the domestic
bank and its foreign competitors. Qualitatively, given the assumption, that the liquidity
constraint holds for zero-pro?t making banks requiring D
f
, the constraint tendentially also
holds for D = D
f
+ . The constraint that the manager has su?cient funds from period
1 pro?ts to transfer back to the parent ?rm can be neglected here, as, as seen above, this
constraint just adds a lower bound to the required repayment D
d
.
97
Technically, the e?ect of the liquidation value on the e?ort problem is
captured by
?p
d
?L
d
< 0. The larger this negative e?ect of a high liquidation
value L
d
on the manager’s e?ort level p
d
, the larger (positive) ?
?p
d
?L
d
[?
d
X
1
+
?(X
2
? L
d
)] > 0 and therefore the larger Omega. So ? < 1 becomes less
likely and therefore it is more likely that the ?rm will prefer a bank with low
liquidation value.
What is going to be the situation in the market for this loan?
If the e?ort problem is su?ciently small, the parent ?rm has a positive
willingness to pay for a higher liquidation e?ciency of a bank. Therefore
the domestic bank can win the loan contract with a demanded repayment
of D > K and make positive expected pro?ts, if it can generate a higher
liquidation value L
d
> L
f
from the investment project than the local banks.
If, in that case, the domestic bank generates a lower liquidation value than
the local banks K < L
d
< L
f
, it could only charge D = D
f
= K at
the maximum, which would lead to zero pro?ts for the bank, whether the
contract is won or not (where the domestic bank will straightforwardly not
win the contract.).
If however the parent ?rm, due to the e?ort problem faced, has a positive
willingness to pay for a lower liquidation e?ciency, the domestic bank wins
the contract and makes positive pro?ts (D > K), only if it creates liquidation
value lower than the local foreign banks. With higher or equal liquidation
e?ciency the domestic bank will make zero pro?ts
20
.
Next up the case of a ?rm that needs to take up a large loan relative to
the size of the investment project is discussed.
The case of a severely cash-strapped ?rm
The di?erence between this and the former case is, that the size of the loan
K the ?rm has to take up to ?nance the project compared to project size is
now so large that K > L
i
, so the liquidation value of the complete project
for any bank does not cover the loan. Therefore, from the perspective of the
banks, the loan is risky now. In case of a project failure the bank will make
a loss K?L
i
. As all banks will only be willing to supply the loan if expected
pro?t is non-negative, they will all ask for a repayment D
i
> K to make
positive ex post pro?ts in the case the project is successful.
20
Note, that the discussion is restricted to cases where all banks have su?ciently high
immanent liquidation e?ciency, such that L
i
> K holds true.
98
Again the expected pro?t of the bank is ?
i
= ´ pD
i
+ (1 ? ´ p)L
i
?K.
The payo?s for parent ?rm and subsidiary manager can intuitively be
derived again. We will immediately see the di?erence to the small loan case.
If the project fails, it is again liquidated by the loan providing bank. As
L
i
< K there is no excess liquidation value left for the parent ?rm. So its
payo? in this case is now 0. If the project is successful, the outside option of
the investing ?rm in bargaining with the manager is 0 and the continuation
value is now (X
2
?D
i
).
The expected payo? of the investing parent ?rm in general therefore is
now
E
I
= K +p
i
[?
i
X
1
+?(X
2
?D
i
)] + (1 ?p
i
)0 ?
1
2
?
2
i
(4.10)
and the expected payo? of the manager is
E
M
= p
i
[(1 ??
i
)X
1
+ (1 ??)(X
2
?D
i
)] ?
1
2
zp
2
i
(4.11)
One again needs to check for the maximum repayment that can be de-
manded by the domestic bank. In this case, the analysis is very straightfor-
ward.
Financing by a foreign (local) bank
As local banks are symmetric, they will again compete themselves down
to zero pro?ts in expectations, so the demanded repayment D
f
is
D
f
=
1
p
f
[K ?(1 ?p
f
)L
f
] (4.12)
The payo? of the parent ?rm when choosing a local bank is now
E
I
(f) = K +p
f
[?
f
X
1
+?(X
2
?D
f
)] ?
1
2
?
2
f
One can again I derive equilibrium e?ort and monitoring levels (see ap-
pendix 3)
p
f
=
X
1
+(1??)(X
2
?D
f
)
z+X
2
1
and
?
f
=
X
1
[X
1
+(1??)(X
2
?D
f
)]
z+X
2
1
99
Note that the equilibrium e?ort and monitoring level are now directly
negatively depending on the demanded repayment of the respective loan-
providing bank.
Financing by the domestic bank
It is straightforward that under ?nancing by the domestic banks expected
parent ?rm payo?, e?ort and monitoring levels respectively, are simply
E
I
(d) = K +p
d
[?
d
X
1
+?(X
2
?D
d
)] ?
1
2
?
2
d
with
p
d
=
X
1
+(1??)(X
2
?D
d
)
z+X
2
1
and
?
d
=
X
1
[X
1
+(1??)(X
2
?D
d
)]
z+X
2
1
The optimal choice of the repayment size for the domestic bank
As
?p
i
?D
i
< 0 and
??
i
?D
i
< 0 it is obvious that the payo? of the parent ?rm
is strictly decreasing in the demanded repayment D
i
irrespective of a bank’s
liquidation e?ciency.
Therefore the maximum repayment the domestic bank can demand equals
the demanded repayment of local foreign banks D = D
f
.
As the local foreign banks compete themselves down to zero pro?ts, D =
D
f
=
1
p
f
[K ? (1 ? p
f
)L
f
] is such that the domestic bank would make zero
pro?ts in expectations, if it generates liquidation value L
d
= L
f
, as expected
pro?t of the domestic bank then is
p
d
D
d
+ (1 ?p
d
)L
d
?K = p
f
D
f
+ (1 ?p
f
)L
f
?K = 0
Therefore the domestic bank would make negative pro?ts if it generates
liquidation value lower than the local foreign banks. In that case the domestic
bank, in order to break even, would have to demand a larger repayment
than feasible D
d
> D
f
= D, therefore not winning the loan contract. If it
generates higher liquidation value it can demand repayment D
d
= D
f
and
make positive expected pro?ts from winning the loan contract as (1?p)L
d
>
(1 ?p)L
f
.
The results of subsection 2.1 can be subsumed by the following Lemma
that can be used in advance.
100
Lemma 1
The domestic bank will choose the liquidation value-maximizing loan pro-
vision mode, if its client is severely cash-strapped or moderately cash-strapped
without facing too large a problem of incentivising its subsidiary manager to
spend e?ort (? < 1). If the bank’s client is moderately cash-strapped and
faces a very large e?ort problem by its manager (? > 1), the bank will choose
the liquidation value-miminizing loan provision mode, however restricted to
L
d
> K.
4.2.2 Loan provision mode choice and endogenous liq-
uidation value
I now want to discuss, what the liquidation value of the domestic bank might
look like, depending on how the bank provides the loan and where it therefore
liquidates the project assets.
For starters, the ”‘basic bank-inherent liquidation value”’ of the project
is discussed. As the domestic bank already has an existing relationship with
the respective ?rm, all else equal, due to the bank already having extended
loans to the ?rm and therefore possibly having generated information about
potential asset takers in the industry in the home country, the domestic bank
should have a higher ”‘inherent liquidation”’ e?ciency in the home country
than the local foreign banks in the foreign country, formally L
0
d
> L
0
f
.
Let us assume that the domestic bank can not sell the liquidated project
assets in the foreign market, if it has no physical presence in this country, as it
can not ?nd potential takers in this market
21
. Therefore, if the bank wants to
provide the loan from its home o?ce, it has to transfer the liquidated assets
back to its home market, where it has knowledge about potential takers of
the assets. However, the bank occurs transportation costs by shipping the
asset cross-border, such that only a fraction of the value L
d
= (1 ? t)L
0
d
arrives in the home country, with 0 < t < 1.
If the domestic bank follows the customer abroad, establishing a physi-
cal presence, it is assumed, that the domestic bank does get to know about
potential asset takers in the foreign market aided by its general industry
knowledge. However, it does not have as good an idea about potential de-
21
This rather strict assumption is made for simplicity only. As long as it holds true, that
liquidation in a market can be done more e?ciently if the respective bank has a physical
presence, and therefore e.g. managers that can locally screen the market for asset-takers,
the following qualitative results hold true.
101
mand in the foreign market compared to the home market only generating
L
d
= ?L
0
d
with 0 < ? < 1 when selling the asset abroad.? can be interpreted
as a proxy for the local information requirement to ?nd the asset-taker with
the highest willingness to pay for the liquidated assets.
The domestic bank’s liquidation value of the project is therefore max-
imized by following its customer, establishing a physical presence abroad,
if
?L
0
d
> (1 ?t)L
0
d
or ? > (1 ?t)
If ? < (1 ?t), the liquidation value is maximized by shipping liquidated
assets back home. Therefore the maximum attainable liquidation value is
independent of the bank following its customer or not, as the advantage of
being present in the local market is not made use of in equilibrium.
Lemma 2
The domestic bank’s liquidation value for the project is maximized by
establishing a physical presence abroad, if transport costs t of repatriating the
physical assets to be liquidated are high relative to the bank’s disadvantage
when selling the assets in the local market (? > (1 ? t)). If ? > (1 ? t) the
maximum liquidation value achievable by the bank is independent of its loan
provision mode choice.
This result does not add too much to a discussion about bank’s contin-
gent liquidation e?ciencies, compared to the exogenous assumptions about
liquidation values made by Marin and Schnitzer (2006). Rather the above
setup just acts as a starting point for the further analysis.
Let us now discuss a further interesting endogenization of the transport
costs of shipping assets abroad from a political (foreign government) perspec-
tive.
Retaining ”‘local”’ assets: Political interference in cross-border as-
set transfer
Assume now that the project assets include human capital. Imagine the
assets of the project that are liquidated to be for example high-tech produc-
tion machinery lines. Therefore an amount of human capital H
FDI
is asset-
embedded. This notion is not wide-spread in the theoretical literature yet,
but a closer look at the intuition shows this concept makes sense. Consider a
subsidiary which operates an assembly line. It buys the single machines for
102
the line. However, to make the line work e?ciently the engineering produc-
tion manager has to align the machinery properly and specify them, so they
operate together in an optimal way. By doing so the engineering production
manager leverages his human capital onto the physical assembly line. Even
if this engineer now is strapped from these assets, part of his human capital
in form of his speci?cation/alignment skills are still present in the physical
assembly line. When a bank after liquidation decides to transfer these assets
back home, therefore the total stock of human capital in the foreign economy
decreases by this level.
Foreign’s economy-wide production function is assumed to be of the sim-
ple form
Y = H
?
K
1??
= (H
FDI
+H
0
)
?
K
1??
(4.13)
where H
0
is foreign’s own endowment in human capital, H
FDI
is the
human capital stock embedded in the assets of the discussed FDI project,
and K is the country’s capital stock
22
.
Assume now that the foreign government can restrict physical asset-
transfer out of the country such that a fraction tL
0
d
and therefore tH
FDI
can
be kept in the country. The government can actively choose t by de?ning
physical asset export restrictions. However, if the government does so, it faces
costs of C(t) = ?t, for example because it has to invest in border patrols, or
more generally because it looses reputation among potential following foreign
investors, or due to other reciprocal actions by other governments.
We assume that the government’s objective is simply to maximize gross
domestic product Y
23
over t, taking the costs of implementing t into consid-
22
Let us abstract from the fact that FDI also increases the capital in the country as it
does not yield additional insight, and taking this into account could also interfere with the
analysis of whether the subsidiary manager can, from a political point of view, actually
transfer period 1 pro?ts back to the mother company. Alternatively, one could distinguish
between portfolio capital (pro?t streams) and physical capital (liquidated assets), such
that the host country government would be able to block transfers of physical capital,
but not of portfolio capital. This assumption would also be absolutely sensible, when you
think about trying to smuggle a suitcase across border, compared to trying to smuggle an
assortment of large machines across border. In this case the following results on human
capital scarcity in the FDI host country would just as well apply to capital scarcity of this
country.
23
This objective function of the government could e.g. be rationalized if the government
levies a withholding tax on all value-added in the economy. In this case, tax income, all
103
eration. So the government maximizes over t
Y = (tH
FDI
+H
0
)
?
K
1??
?C(t) (4.14)
As shown in appendix 4 the optimal asset export restriction t then is
t
?
=
(
?
?K
1??
H
FDI
)
1
??1
?H
0
H
FDI
(4.15)
The simple point to be made is, that
dt
?
dH
0
= ?
1
H
FDI
< 0, so that due to
decreasing marginal productivity of human capital a country with a low hu-
man capital endowment will be less permissive concerning asset-repatriation
than a country with high human capital endowment.
So another intermediate result can be stated.
Lemma 3
”‘Transport costs”’ for repatriation of project assets from the FDI host
country to the (parent ?rm and bank) home country will be larger, the lower
the endowment of the host country in human capital.
So, taking into account Lemma 2, the domestic bank’s liquidation value
of the project is maximized by following its customer, establishing a physical
presence abroad, if the host country is endowed with little human capital,
such that (1 ? t
?
(H
0
))L
0
d
< ?L
0
d
. In this case the bank attains maximum
liquidation value by selling the assets in the host country market. If the re-
spective target country for the ?rm investment is richly endowed with human
capital, such that (1 ?t
?
(H
0
))L
0
d
> ?L
0
d
the domestic bank’s ex post liquida-
tion value is maximized by shipping the assets cross-border and sell them in
the home market. The threshold value for the human capital endowment in
this evaluation is
H
0
= (
?
?K
1??
H
FDI
)
1
??1
?(1 ??)H
FDI
(4.16)
4.2.3 When do banks follow their customer abroad?
Having derived the qualitative optimal liquidation value to be implemented
by a domestic bank, as well as the partially endogenized relationship between
type of loan provision and respective liquidation value for the domestic bank,
these results can now be combined to discuss when a domestic bank should
else equal, would be maximized by a maximization of country GDP.
104
follow its customer abroad
24
. Fixed costs of entry only act as a tiebreaker in
the model if, concerning the liquidation value, the bank is indi?erent whether
to follow the customer or provide the loan without a physical presence abroad.
The below results directly follow from the previous analysis. For starters,
the general structure of the bank decision to follow its customer abroad or not
depends on whether the respective loan project favours banks with high or
low achieved liquidation values, and how these liquidation values are shaped
by the bank’s decision to engage in FYC foreign direct investment itself or
not. If a maximized liquidation value is pro?t-maximizing for the bank, it
will choose the provision mode that maximizes the liquidation value it can
generate. If a (relatively) low liquidation value is pro?t-maximizing, the bank
will choose the loan provision mode that leads to the lowest liquidation value
it can generate. Additionally, the relative liquidation value for the domes-
tic bank, compared to the characteristics of local foreign banks, determines
whether the loan contract can be won by the domestic bank.
For the latter consideration, two other following threshold levels play a
role in the analysis, namely the threshold level for human capital endowment
H
0
, such that the liquidation value for the domestic bank, when selling the
liquidated asset in its home market, equals the liquidation value for the local
foreign banks, formally (1?t
?
(H
0
))L
0
d
= L
0
f
, and ?, such that the liquidation
value for the domestic bank when selling the liquidated asset in the host
country equals the liquidation value for the local foreign banks, formally
?L
0
d
= L
0
f
.
Solving for the respective threshold levels yields
H
0
= (
?
?K
1??
H
FDI
)
1
??1
?(1 ?
L
0
f
L
0
d
)H
FDI
(4.17)
and
? =
L
0
f
L
0
d
(4.18)
Taking above results and Lemmas 1-3 together, and taking into account
the ?xed costs of establishing a physical presence abroad as a tie-breaker,
the following ?nal summarizing proposition can be made
25
.
24
I focus on the ”‘political”’ interpretation of cross-border transport costs for assets.
Of course, one could just as well discuss results when having t depending on geographical
distance.
25
It is implicitly assumed that, even when the foreign government blocks away a rela-
tively large portion in a cross-border asset ?ow, the loan for a moderately cash-strapped
105
Proposition
The domestic bank will follow its customer abroad, establishing a physical
presence in the foreign market, and provide the loan to the ?rm if
a) its customer is either severely cash-strapped (K > L
i
) or moderately
cash-strapped (K < L
i
),while facing a not too large e?ort problem of the
subsidiary manager (? < 1), and the target country for the ?rm’s investment
is poorly endowed with human capital (H
0
< H
0
). Additionally the domestic
bank must not face too high an informational disadvantage in ?nding local
asset-takers, formally ? > ?.
b) its customer is moderately cash-strapped (K < L
i
),the ?rm faces a large
e?ort problem for the subsidiary manager (? > 1), and the host country is
richly endowed with human capital (H
0
> H
0
). Additionally the domestic
bank must face a relatively high informational disadvantage in ?nding local
asset-takers, formally K < ?L
0
d
< L
0
f
.
The domestic bank will supply the loan to its customer without a physical
presence in the subsidiary’s host market if
c) its customer is either severely cash-strapped or moderately cash-strapped,
while facing a small e?ort problem, and the host country is richly endowed
with human capital ((1 ?t(H
0
))L
0
d
> max[?L
0
d
; L
0
f
]).
d) its customer is moderately cash-strapped while facing a large e?ort
problem, and the host country is relatively poorly endowed with human capital
(K < (1 ?t(H
0
))L
0
d
< min[?L
0
d
; L
0
f
]).
The domestic bank will not supply the loan to its customer, leaving the
provision to a local foreign bank if
e) its customer is either severely cash-strapped or moderately cash-strapped,
while facing a small e?ort problem, the host country is poorly endowed with
human capital and the local information disadvantage for selling assets in the
host market is high for the domestic bank (L
0
f
> max[(1 ?t(H
0
))L
0
d
; ?L
0
d
]).
?rm stays non-risky for the domestic bank. This simply restricts the cases to be analyzed,
not changing the general qualitative following results. Allowing t to change the risk type
of a loan from a non-risky to a risky loan would simply add the result, that the lower the
human capital endowment of a host country, the more likely the FDI project is risky from
the point of view of the banks. Then unambiguously the optimal strategy of the domestic
bank is to follow its customer abroad, if this bank is able to overcome the informational
disadvantage of selling the potentially liquidated assets in the host country. This ?nding
simply reinforces the notion that a domestic bank’s physical presence in the FDI host
country maximizes its liquidation e?ciency on its client’s assets, if the host country is
characterized by human capital scarcity.
106
f ) its customer is moderately cash-strapped, while facing a large e?ort
problem, the host country is relatively richly endowed with human capital
and the local information disadvantage for selling assets in the host market
is low for the domestic bank (K < L
0
f
< min[(1 ?t(H
0
))L
0
d
; ?L
0
d
]).
Concerning cases where liquidation value should be minimized by the do-
mestic bank two remarks have to be made. For one, these cases are restricted
to moderately cash-strapped ?rms (K < L
i
), so e.g. minimizing liquida-
tion value by selling assets abroad without a physical presence is ruled out
(L
d
= 0). The domestic bank only wants to reduce liquidation value to the
lower bound L
d
= K in these cases. Below this threshold level the project
would be ex post (after choosing the provisioning mode) risky, which changes
the objectives of the bank.
Additionally it is straightforward that, down to this threshold level L
d
,
it is not ex post ine?cient for the bank to liquidate the assets in the low-
est value-yielding way, as ex post all excess liquidation value (L
d
? K) is
transferred to the client ?rm.
To sum up, in general the loan provision mode choice of the domestic
bank can be mapped by the type of client (?, K), the type of host country
(H
0
, ?, L
0
f
), and bank-speci?c characteristics L
0
d
.
The two following illustrations summarize the qualitative provision mode
outcomes, showing the equilibrium provisioning outcomes given client ?rm
and country/market characteristics.
107
Figure 4.2: Equilibrium Bank Strategies - The Case of Severely
Cash-Strapped or Moderately Cash-Strapped/Small E?ort
Problem Firms
?
H
0
1
FOLLOW YOUR CUSTOMER
STRATEGY
LOAN PROVISION
WITHOUT
LOCAL STRUCTURE
LOAN PROVIDED BY
LOCAL FOREIGN BANK
H
0
FDI
d
f
FDI
o H
L
L
H K
H ) 1 ( ) (
0
0
1
1
1
? ? =
?
?
?
?
?
?
0
0
d
f
L
L
= ?
Figure 4.3: Equilibrium Bank Strategies - The Case of
Moderately Cash-Strapped Firms with a Large E?ort Problem
?
H
0
1
LOAN PROVISION WITHOUT
LOCAL STRUCTURE
LOAN PROVIDED BY LOCAL
FOREIGN BANK
FOLLOW YOUR CUSTOMER
STRATEGY
H
0
FDI
d
f
FDI
o H
L
L
H K
H ) 1 ( ) (
0
0
1
1
1
? ? =
?
?
?
?
?
?
0
0
d
f
L
L
= ?
Basically the optimal provision mode (Follow your customer FDI ver-
sus staying at home) of the domestic bank is a non-unique function of the
108
provision mode-speci?c liquidation value attainable. The optimal and the liq-
uidation value-maximizing provision mode are identical in two out of three
project/?rm-characteristics cases. However, in the case of a moderately cash-
strapped ?rm facing a severe e?ort problem concerning the manager, this
result does not hold true. The liquidation-value maximizing mode is shown
to be a function of country-speci?c characteristics.
So, for the majority of cases discussed, the domestic bank will follow its
customer abroad if the latter enters a country relatively poorly endowed with
human capital.
4.3 Conclusion
I proposed a model yielding insight into the decision of banks to enter a
foreign market physically to follow an existing customer abroad or to serve
this customer from home.
The building stone for the analysis is the paper by Marin and Schnitzer
(2006)[97], who discuss the choice of ?nancing mode for an investor/parent
?rm engaging into foreign direct investment, taking into account a double
moral hazard problem on the side of the host country subsidiary manager
concerning e?ort and repayment. Building on their model, the possibility
of pro?t-making domestic banks, having the choice to follow their customer
abroad, has been added, as well as, in a stylized way, an endogenous provision
mode-dependent arising di?erence in this bank’s liquidation e?ciency. The
former remodelling is done to bring in a dimension of general strategic choice
for banks to be analyzed, while the latter uses a simple ”‘political”’ story to
specify the e?ect of host country characteristics on bank follow your customer
FDI.
Through these additions to the model I am able to discuss bank-,
project/?rm- as well as country speci?c determinants of whether a bank
follows its customer abroad or not.
It is ?rst shown that banks unambiguously choose the liquidation-value
maximizing strategy if the respective customer is severely cash-strapped,
making the loan risky from the perspective of banks. Then the liquidation-
value maximizing strategy given country characteristics is described.
I ?nd that, in this case, follow-your customer induced bank FDI tenden-
tially takes place if the host country is poorly endowed with human capital.
Principally it could also be shown, that follow your customer FDI takes place
if home and host country are distant from each other (high t exogenously)
or, in a slightly twisted setup, if the host country is poorly endowed with
109
capital. Additionally the client-following bank must have a su?ciently high
immanent liquidation advantage over local foreign banks to overcome its in-
formational disadvantage (? < 1) in the local market, or stated in another
way, this informational disadvantage must not be too high.
As could be expected from the basic model by Marin and Schnitzer
(2006)[97], the bank will not necessarily choose the liquidation value-maximizing
provision mode, if the respective customer is only moderately cash-strapped,
such that its FDI project can be considered non-risky from a bank’s perspec-
tive.
The intuitive di?erence to the former case, from the point of view of
banks, is, that an increased liquidation value does not directly a?ect bank
pro?ts. Ex post, the loan providing bank will always get K if the ?nanced
multinational ?rm subsidiary fails, if the loan is small compared to the size
of the project and its assets. Therefore the liquidation value only indirectly
determines the maximum repayment that a bank can require in case of a
successful project outcome, as well as whether the bank can actually win the
loan contract without making negative expected pro?ts.
In this case one additionally needs to distinguish between ?rms that, due
to the exogenous relative bargaining power of parent ?rm and subsidiary
manager, the structure of e?ort costs for the subsidiary manager and payo?
characteristics of the project, face a high or low problem in implementing
su?cient e?ort spent by the manager.
In the subcase of a small e?ort problem the domestic bank will choose
the liquidation value-maximizing provision mode, but not so if the e?ort
problem of its customer is large. In the latter subcase, the bank will choose
the liquidation value-minimizing mode (restricted to L
d
> K), as in this
case the negative e?ect of a high liquidation value on the manager’s e?ort
outweighs the positive e?ects on excess liquidation value and on the fraction
of pro?ts transferred from the manager to the parent ?rm.
I think the analysis helps understand what might be deemed to be a
small puzzle of follow your customer FDI by banks. Considering the bank-
ing sector, there does not seem to be a su?ciently satisfying reason why
banks should follow their customers abroad to extend loans to them and not
simply provide the loan cross-border (directly or indirectly) from home, as
”‘transport costs”’ in the classical sense do not really seem to apply to such
cross-border transactions. Also, considering the literature on multinational
banking, which discusses information asymmetries between local and po-
tential entrant banks (e.g. Dell’Arricia, Friedman and Marquez (1999)[47],
Dell’Arricia and Marquez (2004)[48]) and how to solve the problem of in-
formational disadvantages by physically entering the market compared to
110
cross-border lending (e.g. Lehner (2007)[92]), such considerations should
only play a minor role in supplying a loan to a customer the respective bank
already has an existing loan relationship with, therefore already possessing
knowledge about the ?rm.
Follow your customer-behaviour by banks here is explained by the incen-
tive to alter the liquidation e?ciency of the respective bank on a customer
project abroad in a pro?t-enhancing way. One argument in this reasoning is,
that even though transport costs should not play a large role in the provision-
ing of loans directly, they do play a role in the loan provision mode decision
when it comes to the liquidation of assets, if these assets are of physical na-
ture, e.g. an assembly line or similar machinery. Transport costs in various
forms might then occur when shipping these assets to the market where they
generate the highest market value for the liquidating bank.
Introducing country characteristics in the model might help explain why
we observe bank foreign direct investment in countries that at ?rst sight do
not seem to be too attractive for market-seeking bank FDI. Indeed, as shown
in the introduction of this thesis, bank foreign direct investment in developing
countries has increased signi?cantly over the last decade. In the proposed
model, underdeveloped countries with low human capital endowment would
attract follow your customer-induced bank FDI following some real sector
investment projects.
One could however argue, that these countries should also tendentially
not be likely targets for horizontal real sector FDI. But if these countries e.g.
are attractive targets for outsourcing, they do attract vertical real sector FDI.
Within my story, these FDI projects would then yield an incentive for banks
to engage in follow your customer bank FDI in this country themselves, if
the respective real sector ?rms would (weakly)
26
prefer to ?nance the project
with a ”‘high liquidation e?ciency”’ bank type.
Another implicit result, that is not openly discussed in this paper, stem-
ming from the model structure, is, that a parent ?rm starting up a subsidiary
abroad would always demand to be the only party allowed to repay the loan.
If the subsidiary manager itself would be allowed to repay the loan, the par-
ent ?rm would never be able to extract more than D from the manager,
because the latter could else simply repay D to the bank himself to prevent
liquidation of the project. In contrast, with exogenous bargaining power, the
parent ?rm could extract more than repayment R > D if it can threaten to
break o? renegotiation resulting in the liquidation of the project.
26
In the case of a severely cash-strapped ?rm, the ?rm is really indi?erent between banks
concerning their liquidation value.
111
Introducing follow your customer strategies in this model world also sheds
some light on the discussion of Marin and Schnitzer (2006)[97]. By introduc-
ing cases where debt ?nancing would be conducted by local banks, they dis-
cuss settings in which Foreign Direct Investment by ?rms does not constitute
a capital ?ow into the respective host country. They also show empirically
that the ?nancing of inward FDI by local banks is a signi?cant source of
?nancing for entering ?rms.
However, in their model, they do not allow for bank FDI. Empirically,
their dataset unfortunately restricts them to discuss ex post-local banks, so
that they can not distinguish between local foreign banks and local sub-
sidiaries of banks from other countries.
The above theory suggests, that a signi?cant part of these local inward
FDI ?nanciers might actually be subsidiaries of entrant ?rms’ home coun-
try banks. If these subsidiaries primarily re?nance themselves from home
country deposits and/or other domestic sources of ?nance, real sector FDI
?nanced by these subsidiaries, even when the former does not constitute a
direct capital ?ow, leads to an indirect capital ?ow via the bank structure.
So, even though in a strict sense, indeed some FDI projects would not con-
stitute a direct capital ?ow, they still might lead to capital ?ows via local
bank subsidiaries of home country banks ?nancing these projects.
The proposed model yields a variety of testable hypotheses. However,
numerous obstacles for empirical testing of the model exist. For one, bilat-
eral in-depth bank-?rm data for each loan project involving a multinational
?rm subsidiary is needed to observe the pattern of bank provision mode
for di?erent ?rm-/project characteristics. Even country characteristics can
not be discussed without information about ?rm characteristics, due to the
interaction of both types of characteristics.
Also, concerning the discussion of country characteristics on the level of
bank foreign direct investment, one would need explicit information about
whether an observed bank foreign direct investment project is motivated by
follow your customer-considerations, or whether the project is conducted for
local market-seeking reasons. In the latter case, one would expect a positive
in?uence of the level of host country development on the total volume of
bank FDI in?ows, whereas in the above follow your customer-story the e?ect
works in the opposite direction.
As I at this time do not have access to such a data set, empirical testing
of the proposed theory has to be left to future research.
112
APPENDIX
Appendix 1: Equilibrium e?ort and monitoring levels in case of
small loan and ?nancing by a local bank
The parent ?rm maximizes its expected payo? E
I
over the choice of the
monitoring level ? given the ?nancing is conducted by a local foreign bank.
Max
?
E
I
= p
f
[?
f
X
1
+L
f
+?(X
2
?L
f
)] + (1 ?p
f
)L
f
?
1
2
?
2
f
Simultaneously the subsidiary manager maximizes his expected payo?
E
M
over the choice of e?ort level p given the ?nancing is conducted by a
local foreign bank.
Max
p
f
E
M
= p
f
[(1 ??
f
)X
1
+ (1 ??)(X
2
?L
f
)] ?
1
2
zp
2
f
The reaction functions for the parent ?rm and manager respectively are
?
f
= p
f
X
1
and p
f
=
(1??
f
)X
1
+(1??)(X
2
?L
f
)
z
Inserting reaction functions into each other then yields
p
f
=
X
1
+(1??)(X
2
?L
f
)
z+X
2
1
and
?
f
=
X
1
[X
1
+(1??)(X
2
?L
f
)]
z+X
2
1
Appendix 2: Maximum feasible required repayment D and bank’s
liquidation value
I start with the equilibrium maximum feasible required repayment D,
which is
D = p
d
[?
d
X
1
+?(X
2
?L
d
)]?p
f
[?
f
X
1
+?(X
2
?L
f
)]+(L
d
?L
f
)?
1
2
(?
2
d
??
2
f
)+K
The derivative
?D
?L
d
is then
?D
?L
d
= 1 ??p
d
+
?p
d
?L
d
[?
d
X
1
+?(X
2
?L
d
)] +
??
d
?L
d
[p
d
X
1
??
d
]
113
As we know from the payo?-maximization problem of the parent ?rm
?
d
= p
d
X
1
the ?nal term is zero, yielding
?D
?L
d
= 1 ??p
d
+
?p
d
?L
d
[?
d
X
1
+?(X
2
?L
d
)]
Appendix 3: Equilibrium e?ort and monitoring levels in case of
severely cash-strapped ?rm and ?nancing by a local bank
The parent ?rm again maximizes its expected payo? E
I
over the choice
of the monitoring level ? given the ?nancing is conducted by a local foreign
bank.
Max
?
E
I
(f) = p
f
[?
f
X
1
+?(X
2
?D
f
)] ?
1
2
?
2
f
Simultaneously the subsidiary manager maximizes his expected payo?
E
M
(f) over the choice of e?ort level p given the ?nancing is conducted by a
local foreign bank.
Max
p
E
M
= p
f
[(1 ??
f
)X
1
+ (1 ??)(X
2
?D
f
)] ?
1
2
zp
2
f
The reaction functions for the parent ?rm and manager respectively are
?
f
= p
f
X
1
and p
f
=
(1??
f
)X
1
+(1??)(X
2
?D)
z+X
2
1
Inserting reaction functions into each other then yields
p
f
=
X
1
+(1??)(X
2
?D
f
)
z+X
2
1
and
?
?
=
X
1
[X
1
+(1??)(X
2
?D
f
)]
z+X
2
1
Appendix 4: The foreign government’s choice of t
The governments maximization problem is
max
t
Y = (tH
FDI
+H
0
)
?
K
1??
?C(t)
The First Order Condition then is
K
1??
?(tH
FDI
+H
0
)
??1
H
FDI
?? = 0
Solving for t then yields
t
?
=
(
?
?K
1??
H
FDI
)
1
??1
?H
0
H
FDI
114
Chapter 5
The E?ect of Bank Sector
Consolidation through M&A
on Credit Supply to Small and
Medium-Sized Enterprises
5.1 Introduction
In recent years the evolution of market structure in banking can be de-
scribed by two characteristics: An overall consolidation of the banking sector,
strongly driven by Mergers and Acquisitions, and an internationalization of
banking institutes driven by Foreign Direct Investment
1
. This evolution was
made possible by improvements in information technology, ?nancial deregula-
tion, globalization of real and ?nancial markets and, for Europe, the abolition
of exchange rate risks.
International consolidation in the 1990s was partly driven by multina-
tional banks establishing subsidiaries in foreign markets through the acqui-
sition of local incumbent banks. The volume of cross-border M&A involving
target banks in emerging economies for example rose from about 6 billion
USD in the period 1990-1996 to about 50 billion USD in the period 1997-
2000[63].
1
As is discussed in the introductory chapter of this thesis.
115
Still, most of M&A activity in the banking sector was on a national
level, possibly as a reaction to the threat of multinational bank entry, as the
following graph (taken from Berger et al.(2000)[12] illustrates.
Figure 5.1: Volume of M&A Activity in the Banking Sector
Source: Securities data company
In these surroundings two widely-voiced concerns about the changing
structure of the banking sector have come up in public discussion in recent
years.
For one, there is growing concern in OECD countries, that the consol-
idation of the banking sector might lead to reduced credit availability to
small ?rms. The BIS Group of Ten report on Consolidation [23] identi?es
two possible underlying hazardous processes. First, close to the theoretical
story by Stein (2002)[123], larger and more complex credit institutions arising
through consolidation might have a lower propensity to lend to small ?rms.
Second, as it is widely accepted that ”‘relationship lending”’ is an important
characteristic of credit contracts between banks and small ?rms (e.g. Berger
and Udell (2002)[19], a problem on the market level may arise. Relationship
banking is often described to be on the basis of soft, non-veri?able informa-
tion such as management character and this type of information by design is
hardly transferable between banks in contrast to hard, veri?able information
116
like balance sheets or other audited statements. Therefore, small ?rms loos-
ing their old loan relationship with a bank due to consolidation might face
di?culties ?nding new credit partners.
Indeed, Sapienza (2002)[117] ?nds indication that smaller borrowers are
at the loosing end of bank consolidation, as long-standing credit relations
are disrupted and client information transmission to other banks is hard to
process.
If this negative e?ect is for real, it constitutes a serious challenge to devel-
oped economies. First of all, SMEs should be more harmed by a contraction
of credit available to them than large ?rms, as the former tend to be predom-
inantly ?nanced by bank credit (see BIS G10 report)[23]. From the point of
view of the whole economy, small and medium-sized companies accounted
for 66% of total employment in Europe on average and above 50% in the
U.S. and Canada in 1996 (Source: Eurostat). Also SMEs are supposed to
be more ?exible than large ?rms making them key drivers of innovation and
sectoral change. So if possible concerns about small business ?nancing are
right, the problem is of severe economic magnitude.
At the same time there has been growing concern in Less Developed
Countries (LDCs), that foreign bank entry is non-bene?cial or even harmful
to small ?rms in host countries. Stiglitz
2
stated that
‘Foreign bank entry in Argentina............failed in terms of providing ade-
quate ?nancing for small and medium-sized enterprises”’.
One important thing to note here is, that ”‘cross-border market penetra-
tions are often performed via M&As, rather than via opening new branch
o?ces”’(Berger et al.(2000)[12].
Therefore one could argue that both concerns are tightly linked. As a
matter of fact, market entry through M&A is not fundamentally (qualita-
tively) di?erent from in-market or in-country M&A in the banking sector, at
least not in the sense proposed in the following. However, as will be brie?y
discussed in an informal extension of the following model, due to home coun-
try e?ects the impact of national versus multinational consolidation on SME
credit availability may di?er in its strength.
I follow the literature by Diamond (1984)[53] in that small ?rms in respect
to banking really di?er from large ?rms in that they are ”‘informationally
opaque”’. Like Stiglitz and Weiss (1981)[124] point out, ”‘the informational
wedge between insiders and outsiders tends to be more acute for small compa-
nies, which makes the provision of external ?nance particularly challenging”’.
In this chapter a theory is developed, incorporating both the notion of
2
in: El Pais, 10.1.2002
117
the in?uence of organisational characteristics of banks on their lending be-
haviour, using a twisted version of a general model on ?rm organization and
information processing proposed by Stein (2002)[123], as well as the idea of
relationship banking between banks and small ?rms, to give an explanation
for the potentially adverse e?ects of bank sector consolidation through M&A
on credit availability for SMEs. To that end, a fraction of the setup proposed
by Stein (2002)[123], who models the in?uence of internal capital markets on
management decision making, is used. I twist the model by changing its
objectives, introducing managerial choice in a bank on which type of loan
to specialise on, which allows for a direct discussion of the impact of or-
ganizational change on lending strategies, which also enables me to extend
the model to incorporate the idea of relationship banking. In contrast to
Stein (2002)[123], the model is therefore able to yield results concerning the
impact of organizational change within one bank on small ?rm credit avail-
ability on the market level. Through the notion of relationship banking one
can indirectly incorporate third bank behaviour into the model. Additionally,
the chapter discusses rather informal extensions usable to analyze optimal
(small) ?rm policy towards the banking sector, as well as the possible di?er-
ence between national and international M&A from the perspective of small
?rms.
Agreeing with Stein (2002)[123], that the simple notion of a technological
disadvantage of large banks in dealing with small ?rms in an overall sense is
rather too vague, this disadvantage is traced back to the organizational setup
of banks. Also I follow his idea ”‘that the key distinguishing characteristic
of small-business lending is that it relies heavily on information that is soft
..... that cannot be veri?ed by anyone other than the agent who produces
it.”’[123]. In contrast, banks extend credit to large ?rms based on hard-
information such as detailed income statements, balance sheets, etc.., which
in the following model can be learned about by other agents (CEO) inside the
bank but not by the bank’s outside investors. So, to sum up, it is assumed,
that transaction-based lending
3
is predominant in large ?rm credit supply,
while relationship-based lending dominates with small ?rm credit.
Large institutions typically show more layers of hierarchy than small ones.
Assuming that at least some of the decision power lies within higher levels
of the hierarchy, the importance of being able to pass on information to
the next level is more critical to the bottom (loan) manager than in small
institutions. This leads to the manager in a small bank being more likely to
3
Each transaction stands on its own such that information from the relationship is
irrelevant. Transaction-based lending can be further di?erentiated in ?nancial statement
lending, asset-based lending and credit scoring.
118
focus on projects generating soft information, small-?rm credit, than one in
a large bank. In the following model the e?ectiveness of soft information as
a means to get to know loan projects’ outcomes additionally depends on the
length of the bank-borrower relationship capturing the notion of relationship
banking. Due to this aspect, it is also less likely that small ?rms attain credit
from third banks, after their old relationship is cut. I will show that in the
following model, even if the dropped small ?rms attain credit from a third,
non-consolidated, bank, this will most likely come at the expense of other
small ?rms.
The chapter proceeds as follows. Section 2 gives a brief overview about
recent developments in the banking sector, as well as an overview of the
empirical and theoretical literature about possible motives of (cross-border)
consolidation through M&A. In section 3 the basic model about managers’
specialization decisions in small and large banks is laid out. Section 4 ana-
lyzes the market level e?ects of M&A on small ?rm ?nancing for a variety of
cases. Section 5 brie?y discusses the possible di?erence between inter- and
intranational bank sector consolidation. Section 6 deals with consequences
of M&A as described in the existing literature and other related facts about
banking that can get some new formal explanation from my model. Section
7 concludes.
5.2 Bank Sector Consolidation: An Overview
Before analyzing possible consequences of (cross-border) consolidation through
M&A, I ?rst want to give an overview about possible motives and underlying
factors for this kind of consolidation.
Why has consolidation in this industry picked up steam recently? For
sure, changes in the institutional environment like the Riegle-Neal and the
Gramm-Leach-Bliley Act in the U.S.
4
and the Single Market Program as well
as the monetary union in the EU, enabled banks to consolidate in a variety
of new ways and at lesser costs.
Concerning bank’s incentives to take an active part in consolidation, prac-
titioners interviewed in a study by the bank for international settlements
(BIS) mention revenue enhancement and cost savings as the primary motives
4
The Riegle-Neal Act lifted restrictions on interstate banking for U.S. banks, enabling
the industry to consolidate across U.S. states. The Gramm-Leach-Bliley Act allowed banks
to operate in both the commercial and investment banking segment, therefore making
consolidation across these segments legally feasible.
119
for such activity (BIS 2001)[23].Motives unmentioned, quite understandably,
are managerial motives for M&A.
5.2.1 Revenue enhancement as a motive for bank sec-
tor M&A
Revenue enhancement through M&A could come through the following forms
• Economies of Scale and (Geographic) Scope
• Increase in market power
Revenue economies of scale and (geographic) scope
One of the possible motives for/drivers of international bank expansion
overall, and for international M&A as a means of that, is the increased
demand for international ?nancial services by multinational corporations.
Trade in goods increased from 21% of world GDP in 1987 to 40% by 1997
(see World Bank (2004)[7]. Besides that, the volume of FDI and therefore the
geographic separation of production also increased. Such internationally op-
erating ?rms may be in need of an established banking partner in each of the
places they produce or sell their goods. Several empirical studies underline
this ”‘follow your customer”’-strategy of banks. For example, Goldberg and
Gross (1994) found, that foreign direct investment in a U.S. state was strongly
positively linked with foreign banking assets in this state[68](see chapters 2
and 4 for further studies). Revenue economies of geographic scope therefore
arise in the sense, that ?rms might be willing to pay premia for a bank’s
services if the same bank can provide services to the ?rm in other regions of
operation, too.
Special to universal banks providing a large variety of ?nancial services,
scope economies could be at hand through consumers’ willingness to pay a
premium for one-stop shopping, maybe also driven by the consumer’s un-
willingness to share his private information with more than one ?nancial
institution
5
, and through ”‘reputation economies”’. The latter may arise if
a universal bank is able to transfer its superior reputation in one banking
service to another by collective branding (Rajan, 1996)[114].
5
A simple line of reasoning would be transaction costs of documenting information each
time, another reasoning could be along the line of sharing market information with banks
that also provide services to competitors of the respective ?rm.
120
However there might also be diseconomies of scope in the (banking) in-
dustry, arising from less specialisation leading to less tailor-made products
and therefore lower prices chargeable, or due to customer worries that com-
bining services might lead to con?icts of interest within the bank (e.g. Berger
et al. (2000)[12].
Increase in market power
This motive is most probably a prevalent one for national or even regional
bank M&A, to a lesser degree for cross-border consolidation. In general, the
market level price e?ect of M&A depends on the induced increase in local
market concentration and the general demand structure.
In-market M&As in small (local) markets, with the demand side having
few outside options (e.g. small businesses who seem to strongly depend on
local banks for ?nancing)
6
, could most probably enable a consolidating insti-
tution to charge higher prices from their customers through e.g. lower deposit
rates and higher small business loan rates (see e.g. Berger et al.(2000))[12].
Empirical studies back up this conjecture. Banks tend to have better and
more permanent margins in more concentrated markets
7
.
From a market perspective cross-border consolidation, or even the threat
of it, could however decrease the exercise of market power because of in-
creased market contestability in any given country. However, the empirical
results on this are mixed (e.g. Molyneux et al. (1994)[101], Bikker and
Groeneveld (1998)[22] and Cerasi et al.(1998)[32]).
5.2.2 Cost saving as a motive for bank sector M&A
Possible cost savings theoretically arise through
• Economies of Scale
• Economies of Scope
• Increase in Cost X-E?ciency
6
Kwast, Starr-McCluer and Wolken (1997) ?nd that households and small businesses
almost always choose a local ?nancial institution[90].
7
They charge higher rates on small business loans and pay lower deposit rates (Berger
and Hannan (1997)[13] and react slower to changes in open-market interest rates (Jackson
(1997)[83].
121
Cost economies of scale
Practitioners often mention scale as an important means of reducing av-
erage costs in the banking industry [12]. However several empirical studies
for the U.S. found a rather U-shaped relationship between scale and aver-
age costs, suggesting that medium-sized banks ($100 million – $10 billion in
assets) are slightly more cost e?cient than either large or small banks
8
.
But as Berger et al.(2000)[12], as well as several other studies (see The
Economist (2006)[127], mention, most of this empirical literature relies on
data from the 1980s. The former authors argue, that, due to both techno-
logical process (Automated Teller machines (ATMs, Internet Banking, Risk
Management IT) as well as new dimensions of ?nancial engineering (inter-
national placement of bonds, larger scale economies may have arisen. In
retail business, the emergence of internet banking is a classic example for
conducting banking services in an environment of high ?xed costs (such as
the development of the portal) and low variable costs due to less sta?ng
required per transaction. Other possible sources of cost scale economies are
call centers and payment processing.
However from a broader perspective these latter scale economies, due to
the fact that these processes hardly function as USPs
9
of a bank, could prob-
ably be just as well be achieved by small banks outsourcing parts of these
processes or building networks
10
. If banks in general outsource the parts of
their value chain, that feature cost economies of scale, such e?ency consid-
erations concerning the size of the respective bank vanish. One example for
such outsourcing in Germany is Postbank taking over transaction services
for both Deutsche Bank and Dresdner Bank.
Cost economies of scope
Theoretically there are two main contradicting arguments concerning cost
economies of scope in the banking sector. On the one hand re-usability of
customer information for many products may lead to scope e?ciency (e.g.
Greenbaum et al. (1989)[72]), as duplication of e?ort in information research
is impeded. On the other hand a shift away from core competencies always
may lead to additional administrative costs as well as foregone cost reductions
along the learning curve (see Winton (1999)[132]).
8
see e.g. Bauer, Berger and Humphrey (1993)[9] and Clark (1996)[36].
9
Unique selling positions
10
Therefore one could expect banks to rather downsize by vertical disintegration of
operations, outsourcing e.g. call centers to specialised provider.
122
Cost X-E?ciency
Improvements in X-e?ciency through M&A can be achieved, if the ac-
quiring bank has superior managerial skill or organizational practice which
spills over to the target bank.
Simulations by Savage (1991)[118] and Sha?er (1993)[120] lead to the
conclusion, that X-e?ciency can signi?cantly be raised if ine?cient targets
are restructured by X-e?cient acquirers.
However empirical research yields weaker results on whether actual M&As
increased costs X-e?ciency
11
.
5.2.3 E?cient Risk diversi?cation as a motive for bank
sector M&A
Scale and scope economics in risk management
One very plausible motive for (international) consolidation via M&A for
a bank is to improve the risk-expected return tradeo?.
Under the modern theory of ?nancial intermediation (e.g. Diamond
(1984)[53], Diamond (1991) [54]) this argument holds against the traditional
view of capital markets, that investors optimize their portfolio in the risk-
return dimension themselves.
It is hard to di?erentiate however, whether observed risk-adverse behav-
iour of large U.S. banks (see e.g. Hughes and Mester (1998)[81] is for the
bene?t of the shareholder or due to managerial objectives.
Literature does not ?nd that scale plays a very important role for the
tradeo? (except when you think about banks getting so big that they ”‘can
not fail”’ because of state intervention (see Berger et al. (2000)[12]), but en-
hancing scope through geographic and service portfolio diversi?cation might
very well reduce risk without a similar decrease in expected returns.
The possibility for international risk diversi?cation in the banking indus-
try can be observed in Figure 5.2, which is taken from Berger et al.(2000)[12].
One striking example for such diversi?cation possibilities is the correlation
between banks’ Return on Equity (ROE) in France and the U.S. in the span
between 1979 and 1996 which is -0,815.
11
For the U.S. most studies ?nd positive, but small e?ects (e.g. De Young (1997)[50]).
Vander Vennet (1996,1998)[128][129] for Europe found that cross-border consolidation
increased X-e?ciency, but national M&A often failed to do so.
123
Figure 5.2: Correlation Analysis of Bank ROE among Nations
5.2.4 Managerial motives for bank sector M&A
The existing principal-agency literature gives a broad range of reasons why
managers might want to pursue M&As, both nationally and internationally,
as acquirers.
So-called empire-building tendencies of managemers
12
have received am-
ple interest in theoretical literature. For all kinds of M&As, the sphere of
control of the acquiring institution’s manager becomes larger which increases
manager’s utility mostly through reputational e?ects as well as to a lesser
extent through possible increased compensation (e.g. Chevalier and Avery
(1998)[33]).
Cross-border consolidation especially might be motivated from a manage-
rial point of view by two additional points.
First, with shareholders on average still mostly stemming from a ?rm’s
home country
13
, establishing additional business abroad instead of at home
12
e.g. Jensen and Murphy (1990)[84]
13
This home bias in equity has ?rst been ?rst discussed by French and Poterba
124
might enable a manager to enjoy more perquisites or slack o?. The rea-
soning would be, that home country shareholders normally know less about
economic conditions abroad than at home adding more uncertainty about
payo?s which the manager may use to increase his perks or reduce his e?ort.
Shareholders would then hardly know whether reduced pro?t (e?ciency) is
due to conditions in the new host country or due to changed behaviour of
management.
Also, cross-border consolidation might be comfortable for managers in
that it can reduce risk (see subsection above) increasing their job security
14
,
even if this risk reduction is ine?cient from the shareholder’s point of view,
in that yields too low a level of risk-adjusted expected return.
After discussing some motives for banks engaging in (cross-border) M&As,
the focus is now on the main topic of the chapter, namely how such activity
e?ects a speci?c ?rm segment in the market.
5.3 The Model
The setup di?erentiates between large and small banks by their organiza-
tional characteristics. A small bank is assumed to consist of a single loan
manager, whereas a large bank consists of two such loan managers plus a
CEO on top of the organization.
Both kind of banks are exclusively funded by risk-neutral outside investors
with an outside option of zero. Whereas in a small bank the ?nancing relation
is directly between loan manager and outside investor, in a large bank the
capital runs through the hands of the CEO, who receives capital from the
outside investor and subsequently allocates it among his two loan managers.
After a round of lending activity the investor gets back his initial investment
plus all monetary return on investment.
In both banks loan managers have to decide whether to specialise on sup-
plying credit to large ?rms (L) or small ?rms (S) ex ante. Let us assume that
the only di?erence between these two types of ?rms is the type of information
the bank loan manager can extract from them. To be precise, loan managers
in both type of banks can only extract ”‘soft information”’ from small ?rms,
which is non-veri?able to investors and other agents inside the bank, whereas
dealing with large ?rms yields ”‘hard information”’ that is still non-veri?able
to outside investors, but veri?able to other agents inside the bank.
(1991)[64].
14
see e.g. Morck, Shleifer and Vishny (1990)[102]
125
No matter what kind of credit the loan manager specialises on, he is
assumed to always be able to choose between two potential loan projects.
Each loan can take on size K = {0, 1, 2}. The projects can either be in a
good state of the world (G) yielding return g(K) or in a bad state of the
world (B) yielding return b(K). Both states have ex ante probability 1/2.
The respective projects’ states are non-correlated. Total returns for loan
projects controlled by manager i are denoted c
i
.
To structurize the problem and later reduce notational clutter, the fol-
lowing assumptions about project returns are made
15
.
1. ?2 < b(2) < ?1 < 0 < g(1) < 1
2. g(2) = 2g(1)
3. b(2) < min[3b(1) ?g(1); 2b(1)] = 3b(1) ?g(1)
4. g(1) > ?b(1)
Each loan manager can learn about the actual state of either both of his
possible projects or none of them before deciding how to allocate his capital
under control. The probability of the manager learning about the projects’
actual state of the world is
µ =
_
?
1/t
for small ?rms
? for large ?rms
_
with 0 < ?, ? < 1 (5.1)
? is a general e?ciency parameter for a loan manager generating informa-
tion about a small ?rm project. It is assumed, that this signal becomes more
informative, though with decreasing marginalities, over the length of rela-
tionship t ? [1; ?[ between the bank and the respective small ?rm client
16
.
This seems to be a very intuitive setup as soft facts like management charac-
ter usually take some time to explore. To further simplify the problem it is
assumed that there are always two possible small ?rm loan projects at hand
15
Assumption 1 gives us a well-constrained problem to work on. Assumption 2 is made
just to reduce notational clutter. Assumption 3 leads to simpli?ed equilibria later on and
will be discussed in advance. However, this assumption doesn’t change the qualitative
results. Intuitively it states that funding a bad project with 2 units of capital yields very
bad results, therefore such funding is tried to be ruled out by decision makers in the model.
Assumption 4 is necessary for investors funding the bank with, as it guarantees positive
expected pro?ts for investors.
16
I assume without a prior relationship between bank and ?rm that t = 1.
126
that have same relationship length with the respective bank. In the basic
model analysis each small ?rm has only on existing banking relationship.
The chapter discusses the multiple relationship case in an extension.? is the
respective e?ciency parameter for a manager generating information about
a large ?rm loan. As hard information is mostly passed on through stan-
dardized ?nancial statements the signal has a constant value of information
independent of the speci?c relationship.
To make this problem interesting, ? is assumed to increase over time (or
for that matter, ? to have fallen), else there will hardly have any interesting
equilibria on changing loan strategies in banks
17
.
Finally, as in contrast to Stein (2002) this model is less one of mechanism
design, but rather of a combined Nash-Bayesian game, one needs to charac-
terize prior believes of agents in the model. It is assumed, that all players
in the game ex ante always believe that other players will not be successful
in their research ex ante. However, except the extreme case, that players
believe with probability 1 that others will be successful in their research, the
qualitative results are unchanged.
All agents, loan managers and a possible CEO, have utility functions
U
i
= K
i
+c
i
(5.2)
where K
i
is the amount of capital and c
i
the net cash ?ow of projects under
control of agent i. Agents therefore act like ”‘e?cient empire-builders”’, so
agents’ are interested in both getting as much capital as possible under their
control as well as use the allocated capital e?ciently. These preferences,
together with the non-veri?ability of project information to outside investors,
leads to ?nancing constraints for banks in equilibrium, as discussed in Lemma
1.
Lemma 1
Investors will ?nance small banks with two units of capital in equilibrium
if g(1) + b(1) > 0 and b(2) < 2b(1), where the latter is ful?lled by assump-
tion.(see Appendix for proof)
Lemma 2
If Lemma 1 holds investors will ?nance large banks with four units of
capital
17
The rise of ? can very intuitively be explained by e.g. progress in auditing technology
over time. For transition economies a fall in ? could easily be explained by a loss of social
capital in the transition process.
127
I will only discuss this intuitively, as this Lemma obviously follows from
Lemma 1 in combination with assumed prior beliefs of players in the model.
In such a setup expected payo?s for outside investors’ funding a large bank
with K units are exactly like funding two small banks with K/2 units, there-
fore Lemma 2 holds.
Having laid out the general problem, one can now discuss the specialisa-
tion decision in both types of banks.
5.3.1 Inside the small bank
I start out with the much easier case of a small bank. The single loan manager
of this bank will choose his specialisation based on his expected utility level.
As shown above he will always have two units of capital to allocate on his
projects. His expected payo?, depending on allocative action and state of
information, can be summarized by the following payo? matrix (with chosen
allocation on the left and information on project 1 and 2 respectively received
by the manager at the top).
A — I {GG} {GB} {BG} {BB} {None}
(2;0) 2g(1) 2g(1) b(2) b(2) g(1) +
b(2)
2
(1;1) 2g(1) g(1) +b(1) g(1) +b(1) 2b(1) g(1) +b(1)
(0;2) 2g(1) b(2) 2g(1) b(2) g(1) +
b(2)
2
With the assumptions on project return structures made, one can straight-
forwardly see that the manager will choose
• Allocation (1;1) if he does not receive any information or if information
is {BB}
• Allocation (2;0) if he receives information {GB}
• Allocation (0;2) if he receives information {BG}
• Any feasible allocation if he receives information {GG}
His expected utility specialising on small ?rm loans will then be
EU(S) = ?
1/t
×
_
3g(1)
2
+
b(1)
2
_
+ (1 ??
1/t
) ×[g(1) +b(1)] + 2 (5.3)
His expected utility specialising on large ?rm loans will then be
128
EU(L) = ? ×
_
3g(1)
2
+
b(1)
2
_
+ (1 ??) ×[g(1) +b(1)] + 2 (5.4)
By comparing the utility levels one can come up with a causal relation
between the specialisation decision and the length of established relationships
at the point of decision between the loan manager and his existing small ?rm
client base.
Proposition 1
A small bank loan manager will specialise on small ?rm loans if he has
long-standing relationships with his existing small ?rm base. Else he will
specialise on large ?rm loans. The less e?cient general research about small
?rms is relative to research e?ciency about large ?rms, the longer the critical
length of relationship between small ?rm and bank to ensure further ?nancing
of the small ?rm. The critical length of relationship is t = ln ?/ ln ?. (Proof:
See Appendix)
The following graph illustrates the specialisation decision.
129
Figure 5.3: Relationship Length and Specialisation Decision
18
1 2 3 4 5 6 7 8 9 10
Relationship lengths t with small firms
A
c
c
u
r
a
c
y
o
f
I
n
f
o
r
m
a
t
i
o
n
?
1/t
?
Specialisation on
large firms
Specialisation on
small firms
In this case small banks that have existing relationships with small ?rms
with length t > 5 will specialise on small ?rm credit and small banks with
t < 5 will specialise on large ?rm credit.
5.3.2 Inside the large bank
The case of the large bank is far more interesting, as it includes strategic
interaction.
There are four di?erent stages to be analyzed to ?nd the equilibrium
specialisation decision.
1. Unit manager specialisation decision
2. Unit manager decision on whether to report information to CEO or not
3. CEO decision on capital allocation to loan managers
4. Manager decision on capital allocation to loan projects
18
For ? = 0, 1 and ? = 0, 6
130
The model can be solved by backward induction.
Stage 4
From the small bank case we already know the equilibrium allocation strategy
when a loan manager controls two units of capital (Case 1).
However in the case of a large bank managers might have anywhere from
1–4 units of capital to work with.
Case 2) Manager has one unit of capital
The manager’s problem can be summarized by the following payo? matrix
A — I {GG} {GB} {BG} {BB} {None}
(1;0) g(1) g(1) b(1) b(1)
g(1)
2
+
b(1)
2
(0;1) g(1) b(1) g(1) b(1)
g(1)
2
+
b(1)
2
So maximizing his expected payo? the manager will choose allocation
(1;0) in case he receives information {GB} and allocation (0;1) in case he
receives information {BG}. For any other information he will be indi?erent
between feasible allocations.
Case 3) Manager has three units of capital
19
A — I {GG} {GB} {BG} {BB} {None}
(2;1) 3g(1) 2g(1) +b(1) b(2) +g(1) b(2) +b(1)
3g(1)
2
+
b(2)
2
+
b(1)
2
(1;2) 3g(1) b(2) +g(1) 2g(1) +b(1) b(2) +b(1)
3g(1)
2
+
b(2)
2
+
b(1)
2
Maximizing his expected payo? the manager will choose allocation (2;1)
in case he receives information {GB} and allocation (1;2) if he receives infor-
mation {BG}. For any other information he is indi?erent between feasible
allocations.
Case 4) Manager has four units of capital
Here the only feasible allocation is (2;2),as loan volume per project is
restricted to K = 2.
131
Stage 3
Given that the CEO knows the optimal allocation decision of managers at
the loan level he will assign capital to the loan managers such as to maxi-
mize his expected utility. His allocation will therefore be determined by the
information he receives from his loan managers. Remember that the large
bank will be funded with four units of capital in equilibrium. Due to the
non-veri?ability of soft information, the manager’s preference structure and
the research success of loan managers being insecure, the CEO will not be
able to distinguish between good and bad soft information and ”‘silence”’ on
the side of the manager due to ?nding bad hard information
20
.
With the CEO prior probability belief about managers’ research success
being zero
21
, his contingent payo? matrix can be subsumed as in the table
in Appendix A
22
.
We can then identify the optimal contingent capital allocation strategy
for the CEO, as shown in the following table.
20
Non-documentable information as non-veri?able to the CEO will not be taken into
consideration by the CEO, as he understands the manager to always claim to have found
his projects to be in the good state of the world.
21
Therefore, if the manager does not report any documented information, the CEO will
assume that the manager failed in learning about the projects.
22
We do not need to di?er between cases (GB) and (BG) so the later is subsumed in
the former.
132
Information received CEO utility maximizing allocations
{GG}{GG} (4;0);(3;1);(2;2);(1;3);(0;4)
{GG}{none} (4;0)
{GG}{GB} (4;0);(3;1);(2;2)
{GG}{BB} (4;0)
{GB}{GG} (2;2);(1;3);(0;4)
{GB}{none} (2;2)
{GB}{GB} (2;2)
{GB}{BB} (3;1);(2;2)
{BB}{GG} (0;4)
{BB}{none} (2;2)
{BB}{GB} (2;2);(1;3)
{BB}{BB} (2;2)
{none}{GG} (0;4)
{none}{none} (2;2)
{none}{GB} (2;2)
{none}{BB} (2;2)
These optimal allocations follow directly from the analysis of expected
payo?s of the CEO given the information received.
23
Stage 2
One can now analyze the optimal information strategy of a manager towards
the CEO. At this stage it must be di?erentiated between managers who have
chosen to specialise (at stage 1) on small ?rms, possibly generating soft in-
formation, and those who have focused on large ?rms, possibly generating
hard information. It is of no matter whether soft information gets passed on
to the CEO because the latter won’t value this information at all. Also the
manager can not pass on any kind of information if he hasn’t gathered any.
So we can restrict analysis to managers who have gathered hard information
from large ?rms. These managers can choose whether to report their docu-
mented information to the CEO or keep quiet/just state that their projects
are in a good state. The capital allocation to this manager i, conditional on
23
Here is where assumption 3 plays a quantitative role. If not for this assumption, there
would be unclear optimal allocations for information tuples {none}{BB} and {BB}{none},
as the optimal allocation in these cases depends on the relative size ofb(1), b(2) and g(1).
However, the qualitative results are not changed by the assumption.
133
reporting or not, can be analyzed with the help of the following table
24
.
The following table gives the expected capital allocation for each manager
contingent on the information he and the other manager j passes on to the
CEO.
i — j {GG} {GB} {BG} {BB} {None}
{GG} (2;2) (3;1) (3;1) (4;0) (4;0)
{GB} (1;3) (2;2) (2;2) (2,5:1,5) (2;2)
{BG} (1;3) (2;2) (2;2) (2,5:1,5) (2;2)
{BB} (0;4) (1,5;2,5) (1,5;2,5) (2:2) (2;2)
{none} (0;4) (2;2) (2;2) (2:2) (2;2)
The resulting optimal information strategy towards the CEO has the
following main characteristics. First, passing on information to the CEO
when information about both projects is bad is weakly dominated by not
reporting this documented information (”‘none”’). Therefore the manager
will always not report the documented information in that case. Second, for
all other successfully gathered information reporting the actual information
to the CEO always weakly dominates not reporting.
One obvious advantage of specialising on large ?rm loans potentially yield-
ing hard information is evident here. Positive hard information passed on to
the CEO can increase the capital available to the respective loan manager to
2+ units. At the same time negative hard information can be hidden from
the CEO (not passed on to him), therefore no counteracting negative e?ect
of generating hard information exists.
Stage 1
With the help of results from stages 2-4 a manager’s decision on whether to
specialise on small or large ?rms can now be derived. Note that in this ?nal
step one not only has to worry about how the specialisation decision a?ects
capital allocation among managers but also about the way the specialisation
decision changes the projects’ expected net cash ?ows due to di?erences in
research e?ciency and capital allocation.
I analyze a normal game between the two loan managers in a large bank
to ?nd conditional Nash Equilibria for specialising on small ?rms. In order
to do so, ?rst each manager’s expected utility contingent on his own and the
24
For multiple allocation equilibria (see stage 3).I assume that the manager assigns same
probability to all of those optimal allocations and furthermore use the expected allocation
for the analysis.
134
other managers specialisation decision is analyzed
25
.
Given that manager j chooses to specialise on large ?rms, manager i’s
expected utility specialising on small ?rms is:
EU
i,j=L
(S) = E(K
i
|i = S, j = L) +E(c
i
|i = S, j = L) (5.5)
Using the results from stages 2 and 3 (and weighing them with respective
probabilities) yields
E(K
i
|i = S, j = L) = 2 ?
?
2
(5.6)
Note that the expected amount of capital received in this case depends
negatively on research e?ciency for hard information. This is very intuitive:
The more likely the other manager, specialising on large ?rms, attains hard
information about his project the more likely he will get more funding from
the CEO at the cost of the manager specialising on small ?rms.
Adding the results from stage 4 yields
EU
i,j=L
(S) = 2 ?
?
2
+(1 ?
?
4
)
_
(g(1) +b(1) +?
1/t
(0, 5g(1) ?0, 5b(1))
_
(5.7)
Given that manager j chooses to specialise on large ?rms ,manager i’s
expected utility specialising on large ?rms is:
EU
i,j=L
(L) = E(K
i
|i = L, j = L) +E(c
i
|i = L, j = L) (5.8)
Again, using results from stages 2-4 yields
EU
i,j=L
(L) = 2 + (?
2
?6?) [1/8(b(1) ?g(1))] +g(1) +b(1) (5.9)
Once more this shows two intuitive characteristics. For one, capital allo-
cation to each manager is independent of research e?ciency, as both man-
agers specialise on hard information with same e?ciency. Expected utility as
a whole is increasing in ? as allocation among loan projects becomes better,
therefore the expected cash ?ows of the pool of both projects under control
are higher
26
.
25
Note that by the structure of the game the specialisation decision is made before the
manager learns about the then-available loan projects’ state of the world.
26
When e.g. one project is in the good state and the other in the bad state, c
i
will be
g(1) + b(1) when the manager does not get information about the project states and has
2 units of capital to work with and 2g(1) when the manager learns about the states and
can therefore allocate all capital to the ”‘good”’ loan project.
135
The other two conditional expected utilities are constructed in the same
way and are
EU
i,j=S
(S) = 2 +g(1) +b(1) +?
1/t
(0, 5g(1) ?0, 5b(1)) (5.10)
and
EU
i,j=S
(L) = 2 +
?
2
+g(1) +b(1) +?[g(1) ?0, 5b(1)] (5.11)
We can put these results in the following simpli?ed standard game form
De?ning
X ? EU
i,j=S
(S)
Z ? EU
i,j=L
(S)
V ? EU
i,j=S
(L)
Y ? EU
i,j=L
(L)
yields the following normal game form of the problem.
Manager i – Manager j Small Firms Large Firms
Small Firms X;X Z;V
Large Firms V;Z Y;Y
One can now analyze the critical length of relationship t which leads to
both managers specialising on small ?rms.
Obviously, specialisation on small ?rms (S; S) is a Nash Equilibrium if
X > V and it is unique if Z > Y . As managers in large banks are homogenous
in this model it is actually pretty intuitive that the only reasonable equilibria
are (S; S) and (L; L).
Inequality 1 (X > V ) is ful?lled for
2 +g(1) +b(1) +?
1/t
(
1
2
g(1) ?
1
2
b(1))
>
2 +
?
2
+g(1) +b(1) +?(g(1) ?
1
2
b(1))
Rearranging yields
?
1/t
> ?
1 + 2g(1) ?b(1)
g(1) ?b(1)
(5.12)
136
Solving for t yields two case-dependent results:
t >
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.13)
for ? <
g(1)?b(1)
1+2g(1)?b(1)
and
t <
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.14)
for ? >
g(1)?b(1)
1+2g(1)?b(1)
Case 1 (? not too large) yields a feasible threshold level for relationship
length
t
1
=
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.15)
above which (S; S) is a Nash-Equilibrium. Case 2 would yield a negative
threshold level meaning that for large values of ? (S; S) is never a Nash-
Equilibrium. In this case ?
1+2g(1)?b(1)
g(1)?b(1)
is larger than one, so no relationship
length, leading to an increased allocation e?ciency of the manager for a given
volume of capital to work with, is su?cient to o?set the possible capital
allocation advantage (getting more capital from the CEO to work with) of
choosing to specialise on hard information projects.
Inequality 2 (Z > V ) is ful?lled for
2 ?
?
2
+ (1 ?
?
4
)
_
(g(1) +b(1) +?
1/t
_
1
2
g(1) ?
1
2
b(1)
__
>
2 + (?
2
?6?)
_
1
8
(b(1) ?g(1))
_
+g(1) +b(1)
Rearranging yields
?
1/t
> ?[1 +
1 +g(1)
1 ?
?
4
(g(1) ?b(1))
] (5.16)
Again solving for t, we got to distinguish between two cases, RHS < 1
yielding feasible relationship lengths and RHS > 1, where in the latter case
no t can be su?ciently large.
I ?nd, that for ? > ??
_
(??)
2
?4 with ? =
4+9g(1)?5b(1)
2(g(1)?b(1))
no t exists such
that inequality 2 holds.
137
For ? > ? ?
_
(??)
2
?4 we get
t >
ln?
ln[?[1 +
1+g(1)
1?
?
4
(g(1)?b(1))
]
(5.17)
So inequality 2 is ful?lled for t > t
2
(and low levels of ?) with
t
2
=
ln?
ln[?[1 +
1+g(1)
1?
?
4
(g(1)?b(1))
]
(5.18)
Summing up the two results, for low enough levels of e?ciency in gener-
ating hard information and for su?ciently high relationship lengths to small
?rms at hand, loan managers in large banks will specialise on loan provision
to small and medium-sized enterprises
27
.
The following proposition summarizes the above ?ndings on stage 1.
Proposition 2
Loan managers in large banks will always specialise on large ?rms, if
research e?ciency when dealing with hard information is su?ciently large
(? >
1
1+
1+g(1)
g(1)?b(1)
), independent of available relationships to small ?rms.
They will also de?nitely specialise on large ?rms, even if ? is small,
if they do not have very long standing relationships with small ?rms (t <
ln ?
ln
[
?(1+
1+g(1)
g(1)?b(1)
)
]
).
Specialising on small ?rms in a large bank is an equilibrium strategy i?
? <
1
1+
1+g(1)
g(1)?b(1)
and t > t
1
=
ln ?
ln
[
?(1+
1+g(1)
g(1)?b(1)
)
]
. It is an unique equilibrium i?
? < ? ?
_
(??)
2
?4 and t > t
2
=
ln?
ln[?[1+
1+g(1)
1?
?
4
)(g(1)?b(1))
]
.
5.3.3 Small business ?nancing in small and large banks
To sum up, small banks will lend to small ?rms if they have existing rela-
tionships with small ?rms with length at least
t =
ln ?
ln ?
(5.19)
27
This is a results that di?ers from the line of arguing of Stein (2002)[123], where from
his theory one can derive, that large banks would always choose to ?nance large ?rms.
138
In comparison, even in the best possible case for small ?rms (low ? and
managers coordinate on equilibrium (S; S)), large banks will only lend to
small ?rms with which they have at least relationship length
t
1
=
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.20)
It can easily be shown that t
1
> t (see proof in Appendix 3). This means,
that small ?rms having a relationship with their respective bank of length t,
with t < t < t
1
, will get a positive expected loan volume
28
in the period of
interest, if their bank stays small, but zero loan volume for sure if their bank
becomes part of a larger bank structure.
Proposition 3
Small banks are more likely to extend credit to small ?rms than large
banks. Large banks are more likely to extend credit to large ?rms than small
banks. Small business clients, in the period of interest, with relationship
lengths t < t < t
1
with their respective banks will attain an expected loan
volume of 1 if their bank is small, but no loan if their bank is large.(Directly
follows from proof in Appendix 3)
In the following, let us focus on ?rms with such relationship lengths t
with the consolidated institution as the basis of further analysis, as these are
the interesting cases to be studied.
5.4 The Role of Consolidation
Now when one thinks of consolidation as a merger or acquisition between two
small banks, leading to the evolution of a large bank with the respective loan
managers still in place but now headed by an additional CEO, one can easily
discuss the consequences of consolidation on small ?rm credit on a bank and
market level. In the proposed basic model setup the organisational change
has no impact on bank technological parameters ? and ? and all banks share
the same technology.
28
If a client of a small bank, the small ?rm with respective t as above will get an expected
loan volume of (1 ??
1/t
) ×1 +?
1/t
[0, 25 ×4 +0, 25 ×0 +0, 5 ×1] = 1, which can directly
be seen from the expected capital allocation of the manager as laid out in the section on
small banks.
139
There is one, though hardly interesting, ?rst result, namely consolidation
will never have any adverse e?ect on small ?rms if ? = 0. Let us rule out
that possibility for now.
5.4.1 Changing credit supply within the merged bank
Proposition 4
Small ?rms with relationship length t with t < t < t
1
with their respective
small bank will not be supplied with loans in the analyzed period, if their bank
merges or is acquired, while they would have been ?nanced with an expected
loan volume of 1 by this bank if it had stayed independent.
This result directly follows from Proposition 3, as consolidation via M&A
simply changes the structure of the respective banks from small to large. As
was shown in the case of a small bank, small ?rms with relationship length
t < t would not have received credit from their respective small bank even
if it stayed independent. Even in the best possible case small ?rms with
relationship length t, with t < t < t
1
, will not receive credit from the new
banking structure their old bank is consolidated into, but would have done
so, if the small bank had stayed independent.
The following ?gure subsumes, which bank-?rm relationships are a?ected
by the respective bank being involved in M&A activity.
Figure 5.4: Bank-Firm Relationships A?ected by M&A
t
No Loan,
unaffected
by M&A
No Loan,
negatively effected
by M&A
Positive expected
loan volume,
unaffected by M&A
t t
1
140
5.4.2 SME credit supply at the market level
Whether small ?rms are in the end adversely a?ected by consolidation de-
pends on whether other small banks make up for the lost loan supply from
the consolidated bank.
For starters let us focus on the case where small ?rms only have an existing
relationship with just one bank.
If small ?rms have only one standing relationship with a small bank, then
the ?rms de?nitely adversely a?ected by consolidation will be those who had
relationship length t with t < t < t
1
with one of the consolidated small banks.
Again this is straightforward. The small ?rms identi?ed are those who had
a relationship with one of the consolidated banks but will not get ?nanced
by it again in the period of interest.
With any other bank, small or large, they have relationships with length
t = 1.
As a direct consequence from the results of the basic model above, it must
be true, that these ?rms will never receive credit from another large bank.
Whether they stand any chance to receive credit from another non-consolidated
small bank depends on whether this small bank has existing relationships to
small ?rms and how research e?ciency parameters look like.
Small ?rms in general attain credit from small banks in the period of
interest, if they have at least relationship length t = ln ?/ ln ? with this
bank. As t = 1 for the small ?rm it will only possibly get funded by this
third small bank if ? > ?, whereas it would have received credit if its old
relationship bank had stayed independent if ?
1/t
> ?, which is a less binding
constraint for t > 1 with the old bank.
If this third small bank had formerly specialised on small ?rms it will have
relationship length t > 1 with at least two small ?rms. In this case the small
?rm set free by a consolidated institution will not receive credit from the
third bank, even if ? > ?, as the loan manager in this bank achieves higher
expected utility from sticking with supplying loans to its incumbent small
?rm clients, due to better knowledge about them leading to more e?cient
allocations of loans.
So one can conclude that small ?rms will not only be a?ected by the
merger through the direct e?ect, that it is less likely that the newly merged
bank, in which at least one former part was their credit partner, will extend
credit to them, but also by the fact that it is a lot less likely that they receive
credit from a third non-consolidated bank, which constitutes a real problem
at the market level
29
.
Proposition 5
29
To be precise the only potential source of credit for the analyzed ?rm are new entrant
141
Small ?rms with a single relationship with a consolidating bank will su?er
in overall credit availability in expectations through this consolidation. They
will be less likely to get credit from the consolidated institution, as well as be
less likely to receive credit from other banks in the market, due to their lack
of a relationship with other banks. The only potential source of ?nance are
small new entrant banks in a setting where ? > ?.
5.4.3 Small ?rms with multiple bank relationships –
The odd ?rm out
The impact of consolidation on small ?rms above was concerned with small
?rms only having an existing relationship with one bank.
Let us now consider the case where these ?rms have n > 1 relationships
with at least one of the banks not being part of a M&A process, so the
respective ?rm had taken up a loan from 2 banks in its history at some
di?ering times.
In this case there exist di?erent possible scenarios for the overall credit
availability to small ?rms. The easiest case in thinking about the problem is
an economy with three banks i,j,k where i and j are small banks that merger
while k stays independent and the loan market situation for small ?rm s is
analyzed
30
. It should be clear, that a small ?rm s with relations to banks i
and j will not fare any better than if the ?rm only had a relationship with
either i or j, as both banks merge together.
The respective length of relationships between the small ?rm and the
banks are t
i,s
,t
j,s
and t
k,s
.
Already knowing that, if bank k is a large bank it will likely not supply
credit to ?rm s, we focus on the more interesting case of bank k being small.
The benchmark for the following analysis is a single relationship small
?rm who will not attain credit from the consolidated institution. So, for
additional insights compared to the case of a single ?rm-bank relationship,
the focus is on the case t
i,s
< t
1
.
Case 1: Bank k has been specialising on large ?rms so far and t
k,s
> t
banks. Because if the other type of incumbent small banks, that have specialized on large
?rms to date, exists, these will not supply positive expected loan volume to the analyzed
small ?rm in the period of interest, given the assumption of rising ? over time. This is
shown in appendix 4.
30
I do not discuss bank entry here.
142
Due to the assumption of rising ? over time this case does not exist
31
.
Case 2: Bank k has been specialising on large ?rms so far and t
k,s
< t <
t
i,s
If so, the small ?rm will de?nitely not receive any credit from any incum-
bent bank whereas it would with positive probability, if bank i had stayed
independent.
Case 3: Bank k has specialised on small ?rms serving ?rms l,m and
t
k,s
< t
k,l
= t
k,m
.
Even if t
k,s
> t the small ?rm will not receive credit from the third bank
k because the loan manager of the bank has longer standing relationships
at hand, therefore better knowledge about these ?rms and therefore higher
expected utility serving ?rms l,m instead of s.
Case 4:Bank k has specialised on small ?rms serving ?rms l,m and t
k,s
>
t
k,l
= t
k,m
> t
This is the case, where the ?rm su?ering from M&A activity is not the
one set free by a consolidated institution, but other small ?rms are negatively
a?ected by such a development.
Here ?rm s will attain positive expected loan volume from bank k but
only at the expense of either ?rm l or m. Remember that small banks in
the model are restricted to only screen two possible loan projects. So even
though in this case ?rm s still possibly attains credit, at least one other small
?rm will be adversely a?ected by the ripple e?ect of consolidation through
M&A.
The case analysis can be subsumed in the following proposition.
Proposition 6
Keeping up multiple relationships with banks decreases the probability of a
small ?rm being negatively a?ected by consolidation through M&A. However,
as a group, small ?rms overall su?er from consolidation. Small ?rms nega-
tively a?ected by consolidation need not be direct clients of the consolidated
institution.
31
With rising ? if t
k,s
> t in the period of interest it must be true that t
k,s
> t in
the period before. Therefore bank k would have specialised on small ?rms in the earlier
period.
143
5.5 National versus Multinational Consolida-
tion through M&A and heterogeneous coun-
tries
Up to this point I have established a model giving insights into potential
e?ects of overall M&A in the banking sector on small ?rm credit availability.
Let us now discuss an international perspective namely whether in the eye
of small ?rms consolidation involving a foreign bank is better or worse than
pure intranational active consolidation.
How could these two types di?er in a non-trivial way
32
?
One potential di?erence might come in the form of a home country spe-
ci?c bank heterogeneity in research e?ciency. As stated in the basic setup,
the research e?ciency on hard information should intuitively be rather in-
dependent of length of relationship between bank and large ?rm, due to
the standardization of research on hard information. But one factor of how
good managers inside a bank do research on hard information should be the
amount of times they have done that which means the population of large
?rms the manager has dealt with. This is simply the idea of ”‘learning by
doing”’(e.g. Krugman (1987)[89]. As the globalization of ?nancial services is
a rather new phenomenon the size of the home country large ?rm population
should therefore have a positive e?ect on the respective banks’ research e?-
ciency concerning hard information
33
. One possible functional form, similar
to the one used for learning in an international environment by Krugman
(1987)[89], for research e?ciency of bank i is
?
i
= min(?
P
T=1
[X
i,T
+?X
j,T
] ; 1) (5.21)
with 0 < ? < 1;X ? 0;0 < ? < 1
32
Of course intranational M&A, besides changing the organizational structure of banks,
also reduces the number of banks in the market whereas international M&A only changes
the structure of one local bank, leaving the number of banks operating in the market
constant. Let us abstract from this simple di?erence.
33
One could of course also argue, that the di?erence between national and foreign banks
might also depend on the e?ciency ? of generating soft information. As Berger and
Udell (2002)[19] put it ”‘Cross-border consolidation may create additional problems for
relationship lending because a foreign-owned bank may come from a very di?erent market
environment, with a di?erent language, culture,....”’.
144
where X
i,T
denotes the size of the large ?rm population in the bank’s
home country at time T, X
j,T
is the former in all other countries, ? gives the
level of learning by international spillovers and P is the period of interest.
At the maximum the manager always learns about the actual state of
large ?rm credit projects, so learning is bounded at maximum 1.
If bank research e?ciency on hard information depends on the size of
the population of large ?rms in their respective home country, banks from
”‘larger”’ countries will have higher research e?ciency on hard information
than banks from ”‘small”’ countries, as long as ? = 1 in all the respective
countries, as
? theta
i
?X
i,T
> 0 in that case.
What follows from this for the analysis on bank sector consolidation?
If a country is small in the above sense, all else equal, national banks will
have a rather small degree of research e?ciency on hard information ?
medium
.
A foreign bank from an even smaller country would show even smaller ?
low
. In contrast a foreign bank stemming from a larger country would feature
a large ?
high
. Now assume that any of these banks take over or merge with
another domestic bank, with hard information research e?ciency equal to
the maximum of these e?ciencies of both merging banks.
Remember, that in the best case from the point of view of small ?rms
this bank will only lend to small ?rms if
t > t
1
=
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.22)
It can easily be shown for bank i that
?t
1
?X
i
> 0 for ?
i
< 1
34
.
So even if a higher ?, induced in the consolidated institution by a merger
with a foreign bank from a large country, does not lead to the bank not
providing loans to any small client ?rm anymore, no matter the lengths
of relationships at hand, it does increase the minimum relationship lengths
required to keep up ?nancing of small enterprise customers.
Summarized, the higher the research e?ciency of the consolidated bank
the less likely the bank keeps serving its existing small ?rm client base.
Proposition 7
Result 1: Bank sector consolidation should have worse e?ects on small
?rm credit supply, the larger the respective country’s large ?rm population is.
34 ?t
1
?X
i
=
?t
1
??
×
??
?X
1
.
It is easy to show that
?t
1
??
=
ln ?
??(ln ?)
2
> 0.
As
??
? X
1
> 0 it follows that
?t
1
?X
i
> 0.
145
Result 2: Consolidation through M&A involving foreign banks is more
harmful to small ?rms than pure national consolidation, if the foreign ac-
quiring/merging bank stems from a country ”‘larger”’ than the host country,
whereas national consolidation is more harmful if the foreign banks come from
a ”‘smaller”’ country.
5.6 The Empirics of Small Business Lending
and Consequences of Bank M&A
What does existing literature have to say about the characteristics of bank
lending to SMEs and consequences of M&A activity in the banking sector?
5.6.1 Characteristics of Small Business Lending
As is assumed in the setup of the model, one very important feature of
small business lending is that it is quite substantially relationship-based.
Numerous studies employing a variety of methods to measure relationship
strength between bank and borrower (e.g. length of relationship, exclusivity
of credit relationship, service scope of relationship) ?nd, that strength of
relationships has positive implications for the respective borrower, in that the
latter for example pay lower interest rates (e.g. Berger and Udell (1995)[17])
and are more likely to get funded by a respective bank (e.g. Scott and
Dunkelberg (1999)[119]).
5.6.2 Consequences of (Cross-border) Consolidation on
Small Business Lending: The Existing Literature
What does the existing literature have to say about the consequences of
(cross-border) M&A in the banking sector on small ?rms?
For in-market M&A, leading to increased market concentration, loan in-
terest rates and service fees might increase due to the increased market power
of banks in this market. Theoretically, small ?rms might be most a?ected by
this development, due to the fact that they are found to be rather immobile
concerning their source of funding (Kwast et al.(1997)[90].
146
Indeed, studies (e.g. Akhavein, Berger and Humphrey (1997)[2]) ?nd that
in-market M&As, substantially increasing market concentration, empower
banks in this market to charge higher prices.
In terms of cross-border consolidation, this is typically not a kind of in-
market M&A from the perspective of small ?rms, even though it might be for
large wholesale customers in investment banking, so it seems that national
M&A in this respect is more likely to be harmful to small ?rms.
Besides the market power implications, M&As also change three charac-
teristics of the involved banks, which, after reviewing the literature, might
a?ect small business credit supply (see Berger et al.(2000)[12], where the
second and third characteristics are the one the proposed model deals with:
• Increase in consolidated bank scale (and scope)
• Increase in consolidated bank organizational complexity
• External e?ects on the lending behaviour of other banks in the market
In general an increase in bank size might lead banks to shift their strategic
focus from small business loans to wholesale services such as underwriting and
other investment banking activities. An explanation could be, that capital
market services can only be provided by large banks due to e.g. a critical
mass of depositors. If banks face an upward sloping supply curve of capital
this possible new market, now feasible to be served, would reduce the capital
of this bank allocated to small business ?nancing.
For bank scope e?ects, as Berger et al.(2000)[12] mention Williamson-
type organizational diseconomies of scale, large banks might be ill-equipped
to conduct relationship-based lending.
Empirical studies concerning the U.S., like Berger and Udell (1996)[18]
and Berger et al.(1998)[16], indeed ?nd a negative relationship between a
bank’s size and the proportion of its assets employed in providing small
business credit.
35
Raising the topic of large banks’ relationship lending-capabilities, Cole,
Goldberg and White (1999)[38] and Berger et al.(2002)[15] ?nd, that large
banks tendencially engage transaction-based lending to small ?rms, whereas
relationships play a much larger role in lending between small banks and
their respective small business clients.
However, whereas ex ante bank scale seems to be a negative indicator on
whether small ?rms receive credit from the respective bank at all, quite a
35
The latter study observed, that small banks (below $100 million assets) use 9% of
assets on small business lending, very large banks (above $10 billion assets) only 2%.
147
few studies show evidence, that large banks seem to be safe havens for the
small ?rms they after all serve in times of ?nancial distress. Hancock and
Wilson (1998)[75] for the U.S. ?nd that ?nancial distress reduced small ?rm
credit in large banks way less than in small banks. DeHaas and van Lelyveld
(2002)[46] come up with the same results for Eastern Europe for large foreign
banks compared to relatively smaller national banks. For the latter it is quite
unclear whether this in an e?ect of bank scale or rather (geographic) scope.
Houston et al.(1997)[79] for the U.S. ?nd that loan growth of banks in U.S.,
who are part of a Bank Holding Company (BHC) was less depending on
those banks own than on the holding’s ?nancial health, cementing the idea
that these holdings serve as internal capital markets.
The e?ect of an increase in organizational complexity on small business
credit is theoretically analyzed in my model. Numerous studies investigate
this relation also looking on the di?erence between pure national and multi-
national banks.
For example out-of-state ownership of a bank in the U.S. predominantly
had a negative e?ect on small business credit in the respective banks (e.g.
Berger et al.(1998)[16]. Being part of a bank holding had an e?ect in the
same direction (e.g. Berger and Udell (1996)[18].
In order to understand the e?ect of consolidation on the market level, one
has to analyze how other banks react to M&As in their market in respect to
small business lending.
The empirical literature gives mixed results. Berger et al.(1998)[16] point
to indication, that other banks more than o?set the negative e?ect on small
business lending in a consolidated institution by expanding their business in
this ?eld. However, contrary to this point, Berger et al.(1999)[11] ?nd a ne-
glectable external e?ect of M&A on other banks’ small business lending, with
only mature small banks being a?ected positively in their lending activity in
this segment.
Furthermore, Berger et al.(1999)[11] discover a positive e?ect of consoli-
dation on market entry in the banking industry. Combined with the generally
found empirical results, that de novo banks tend to have the highest share
of assets invested in small business loans
36
, this induced (additional) entry
might help o?set bank-level negative e?ects of consolidation on the credit
availability for small and medium sized enterprises.
In a recent contribution, Bonaccorsi di Patti and Gobbi (2007)[52] study
the e?ect of bank mergers and acquisitions on credit supply to Italian ?rms,
36
This has been found by several studies e.g. Goldberg and White (1998)[91], Berger,
Bonime, Goldberg and White (1999)[11] and DeYoung, Goldberg and White (1999)[115].
148
with the ?rm dataset including mainly small and medium sized enterprises.
Supporting the above model’s propositions the authors ?nd, that these ?rms
are at least in the short-run negatively a?ected by such change in the market.
To be precise, such reorganization within the banking sector leads to a drop
in credit of 9%. However, not captured by our model, they ?nd that credit
levels for these ?rms revert to old volume after three years.
5.6.3 Bank sector consolidation and relationship man-
agement strategies of small businesses
One indirect result of the model is, that for the single small company the
threat of bank sector consolidation, and therefore the fear of loosing its credit
partner, should give an incentive to have lending relationships with multiple
banks. Even though such multi-sourcing is theoretically expensive (even
more so than for transparent ?rm) as soft information has to be transferred
a number of times
37
, the empirical literature does support the idea that ?rms
use multiple banks as sources of ?nance (e.g. Detragiache et al.(2000)[49].
Berger et al.(2001)[14] argue more or less exactly to my point, that ”‘in-
formationally opaque ?rms are more likely to have multiple lenders......This
is because after being cut o? by the primary bank, opaque ?rms are likely
to encounter more di?culty in ?nding additional lenders and/or have less
favourable loan terms until their new relationships mature.”’
This is very close to the theoretical story suggested above, as the major
problem of small businesses in ?nding new credit partners is here, that the
former do not have an existing relationship with the latter.
5.7 Findings and Shortcomings
In this paper I have developed a model taking up concerns about the e?ect
of consolidation in the banking sector through M&A on credit availability
for small ?rms, combining the notion that banks di?er in their lending strat-
egy because of di?erent organizational setup with the idea that relationship
lending plays a large role in small business lending. I come up with the result
that consolidation via M&A indeed reduces the availability of small business
37
Costs could come in the form of simple transaction costs, duplicated e?ort and free-
rider problems. Also, ?rms might be reluctant to share con?dential information to multiple
banks who also have relationships with their competitors (e.g. Chiesa et al.(1995)[21].
149
lending in the economy. A reasoning on why ?rms might want to keep up
lending relationships with several banks can be derived in consequence.
Unfortunately a discussion on welfare e?ects of such changes can not
be made, as the loss in potential loan provision for SMEs is the gain in
potential loan provision for large informationally non-opaque ?rms in this
model. Welfare discussion would therefore need a further assumption on
which type of ?rm should be the favoured main recipient of bank ?nance. As
it is often argued, that small, new ?rms are the backbone of technological
development, whereas large ?rms tend to operate in saturated markets, it
might be true form an economy-perspective that loan provision to SMEs
should be the priority. This would then tentatively suggest that welfare is
reduced by such a shift in loan supply from small to large ?rms.
The model probably gives a good ?rst formal market-level insight into
the topic, but might fail to fully explain the reality because of a few short-
comings
38
.
One might be, that consolidated banks might often organize in a decen-
tralized way, for example by having each unit perform like a ?nancial center,
which would not lead to the evolution of an internal capital market. This
in turn would keep the manager’s specialisation decision in the old incen-
tive surroundings of a small bank. Also banks might organize in a way to
completely disentangle large and small ?rm credits personnel-wise.
So it looks like their might be scope for organizational strategy reducing
the hazardous e?ect of consolidation on small ?rm lending. It seems like
the main point of bank organization in the view of small ?rms would be
decentralized ?nancing negating the e?ect of the existence of internal capital
markets on lending strategy.
Also, I implicitly employ the idea, that soft, relationship information is
rather bank-speci?c than manager-speci?c. If, realistically the information
is with the manager, the manager of a merged bank might simply leave the
consolidated bank taking his relationships with him and starting up a new
small bank. Whether they do so would then depend on their expected utility
in the new bank compared to the consolidated institution.
Sticking with the manager’s objectives, one though vague argument might
also be a misspeci?cation of manager’s preferences. Managers might have
38
Of course, the result that there simply will be no more provision of loans to some
SMEs in some circumstances has to be seen as rather qualitative. Realistically small ?rms
might not be completely excluded from loans, but rather would have to pay higher interest
rates than without consolidation. Still the result that consolidation in the way discussed
in my model is harmful to SMEs would hold.
150
a really strong preference to keep up existing relationships due to person-
nel friendships with small ?rm clients’ management, social standing in the
community, etc. o?setting potential empire-building tendencies. Changing
preferences that way would lead to the conclusion that small business lend-
ing might actually be of larger volume than e?cient, as even non-performing
loans in expectations could be handed out due to these kind of preferences.
Another point to notice is that I implicitly assume that both types of
banks, small and large, are ex ante capable of supplying both loans to large
?rms and small ?rms. However due to risk-regulation and the possibly dif-
fering nature of bank relationships with small and large ?rms this might not
be the case
39
.
For all of the above reasoning, a large variety of possible avenues for fur-
ther theoretical research exists. Also, discussing entry motives of multina-
tional banks or motives for intranational M&A and how they are intertwined
with consequences of such actions should be very interesting, because motives
should shape strategies which should determine the e?ects of such actions.
Concerning empirics, the literature on the e?ects of in-market consolidation
or foreign bank entry via M&A on SME loan supply still needs further work,
as results seem ambiguous on quite a lot of questions about the topic. One
important point is an international comparison of e?ects of M&A on small
business lending. At the moment most of the research is based on U.S. data,
so there is not much leeway to understand whether and how the underlying
processes di?er between countries.
One ?nal important line of research to us would also be how the Basel II
rules change the picture on small ?rm lending. On the one hand, Basel II is
not in accordance with relationship lending threatening small ?rms relying
on this type of credit. On the other hand with the clear knowledge about
information requirements through the communication of Basel II small ?rms
will be pushed to increase their level of hard information, which should enable
them to be more competitive with large ?rms in the market for loans.
Within this thesis one ?nal point that can be made is that, stemming
from the above analysis, one would expect foreign bank entry via Green?eld
Investment to be the optimal entry mode from the perspective of small and
medium-sized enterprises, because this mode of entry, contrary to entry via
acquisition, does not directly lead to one small bank in the market becoming
integrated in a large bank structure, therefore the negative e?ect of such ?rms
39
For example a large ?rm might need multiple services from one bank such as wholesale
banking. Small banks might simply not have the scale to compete in these lines of business
thereby loosing businesses like credit procurement from large ?rms as well.
151
set free by the new organizational structure and not ?nding a new bank for
loan provision, does not apply to entry via Green?eld Investment.
152
APPENDIX
Appendix 1: Proof of Lemma 1
The expected net payo? of investor ?nancing a small bank with K units
of capital is respectively
• EP(1) =
g(1)
2
+
b(1)
2
• EP(2) = g(1) +b(1)
• EP(3) =
3g(1)
2
+
b(1)
2
+
b(2)
2
Overall, a ?nancing volume K is feasible if EP(K) > 0 (outside option)
For g(1)+b(1) > 0, which is ful?lled by assumption 4, banks are therefore
able to obtain two units of capital.
For b(2) < 2b(1) expected net payo? of investing three units of capital is
negative:
To show:
3g(1)
2
+
b(1)
2
+
b(2)
2
< 0 ?b(2) < ?b(1) ?3g(1)
for g(1) +b(1) > 0 follows
?b(1) ?3g(1) < 2b(1)
by assumption 3 b(2) < 2b(1) therefore
b(2) < ?b(1) ?3g(1) ?EP(3) < 0
Appendix 2: Proof of Proposition 1
The small bank loan manager will strongly prefer specialisation on small
?rm loans over specialisation on large ?rm loans i?
EU(S) > EU(L)
or
153
?
1/t
×
_
3g(1)
2
+
b(1)
2
_
+ (1 ??
1/t
) ×[g(1) +b(1)] + 2
>
? ×
_
3g(1)
2
+
b(1)
2
_
+ (1 ??) ×[g(1) +b(1)] + 2
Rearranging yields
?
1/t
> ?
and, using the log
t >
ln?
ln?
Therefore the minimum relationship length inducing the manager to again
lend to small ?rms in the period of interest is
t >
ln?
ln?
Appendix 3: Proof of di?erence in SME loan supply small
banks/large banks
I want to show t < t
1
Inserting yields
t =
ln ?
ln ?
< t
1
=
ln ?
ln
[
?(1+
1+g(1)
g(1)?b(1)
)
]
which can be simpli?ed to
1+g(1)
g(1)?b(1)
> 0
which is ful?lled as by the assumption necessary for bank operations to
be funded g(1) ?b(1) > 0/g(1) > 0.
Appendix 4: Third banks behaviour formerly specialised on
large ?rms
I proof by contradiction that banks formerly specialised on large ?rms
will not supply a positive expected loan volume to the analyzed ?rm set free
by a consolidated institution.
De?ne ?
?1
as the hard information research e?ciency in the former period
and ?
0
as the parameter in the period of analysis, with by assumption ?
?1
<
?
0
.
A bank will only supply positive expected loan volume to a ?rm with it
has relationship lengths t = 1 in the period of interest i?
154
? > ?
0
The bank specialises on large ?rms in the earlier period i?
? < ?
?1
So a bank specialised on large ?rms will provide positive expected loan
volume to the small ?rm in the analyzed period i?
?
0
< ? < ?
?1
which is a contradiction to
?
0
> ?
?1
155
A
p
p
e
n
d
i
x
A
:
C
E
O
p
a
y
o
?
s
g
i
v
e
n
i
n
f
o
r
m
a
t
i
o
n
r
e
c
e
i
v
e
d
a
n
d
s
u
b
s
e
q
u
e
n
t
c
a
p
i
t
a
l
a
l
l
o
c
a
t
i
o
n
I
—
A
(
4
;
0
)
(
3
;
1
)
(
2
;
2
)
(
1
;
3
)
(
0
;
4
)
{
G
G
}
{
G
G
}
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
{
G
G
}
{
n
o
n
e
}
4
g
(
1
)
7
g
(
1
)
2
+
b
(
1
)
2
3
g
(
1
)
+
b
(
1
)
b
(
2
)
2
+
5
g
(
1
)
2
+
b
(
1
)
2
2
g
(
1
)
+
b
(
2
)
{
G
G
}
{
G
B
}
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
3
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
b
(
2
)
{
G
G
}
{
B
B
}
4
g
(
1
)
3
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
2
b
(
1
)
b
(
2
)
+
g
(
1
)
+
b
(
1
)
2
b
(
2
)
{
G
B
}
{
G
G
}
2
g
(
1
)
+
b
(
2
)
3
g
(
1
)
+
b
(
1
)
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
{
G
B
}
{
n
o
n
e
}
2
g
(
1
)
+
b
(
2
)
5
g
(
1
)
2
+
3
b
(
1
)
2
3
g
(
1
)
+
b
(
1
)
b
(
2
)
2
+
5
g
(
1
)
2
+
b
(
1
)
2
2
g
(
1
)
+
b
(
2
)
{
G
B
}
{
G
B
}
2
g
(
1
)
+
b
(
2
)
3
g
(
1
)
+
b
(
1
)
4
g
(
1
)
3
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
b
(
2
)
{
G
B
}
{
B
B
}
2
g
(
1
)
+
b
(
2
)
2
g
(
1
)
+
2
b
(
1
)
2
g
(
1
)
+
2
b
(
1
)
g
(
1
)
+
b
(
1
)
+
b
(
2
)
2
b
(
2
)
{
B
B
}
{
G
G
}
2
b
(
2
)
b
(
2
)
+
b
(
1
)
+
g
(
1
)
2
g
(
1
)
+
2
b
(
1
)
3
g
(
1
)
+
b
(
1
)
4
g
(
1
)
{
B
B
}
{
n
o
n
e
}
2
b
(
2
)
b
(
2
)
+
3
b
(
1
)
2
+
g
(
1
)
2
3
b
(
1
)
+
g
(
1
)
b
(
2
)
2
+
3
g
(
1
)
2
+
3
b
(
1
)
2
2
g
(
1
)
+
b
(
2
)
{
B
B
}
{
G
B
}
2
b
(
2
)
b
(
2
)
+
b
(
1
)
+
g
(
1
)
2
g
(
1
)
+
2
b
(
1
)
2
g
(
1
)
+
2
b
(
1
)
2
g
(
1
)
+
b
(
2
)
{
B
B
}
{
B
B
}
2
b
(
2
)
b
(
2
)
+
2
b
(
1
)
4
b
(
1
)
b
(
2
)
+
2
b
(
1
)
2
b
(
2
)
{
n
o
n
e
}
{
G
G
}
2
g
(
1
)
+
b
(
2
)
b
(
2
)
2
+
5
g
(
1
)
2
+
b
(
1
)
2
3
g
(
1
)
+
b
(
1
)
7
g
(
1
)
2
+
b
(
1
)
2
4
g
(
1
)
{
n
o
n
e
}
{
n
o
n
e
}
2
g
(
1
)
+
b
(
2
)
b
(
2
)
2
+
2
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
2
b
(
1
)
b
(
2
)
2
+
2
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
b
(
2
)
{
n
o
n
e
}
{
G
B
}
2
g
(
1
)
+
b
(
2
)
b
(
2
)
2
+
5
g
(
1
)
2
+
b
(
1
)
2
3
g
(
1
)
+
b
(
1
)
5
g
(
1
)
2
+
3
b
(
1
)
2
2
g
(
1
)
+
b
(
2
)
{
n
o
n
e
}
{
B
B
}
2
g
(
1
)
+
b
(
2
)
b
(
2
)
2
+
3
g
(
1
)
2
+
3
b
(
1
)
2
g
(
1
)
+
3
b
(
1
)
b
(
2
)
+
g
(
1
)
2
+
3
b
(
1
)
2
2
b
(
2
)
156
Chapter 6
Conclusion and Outlook
The ongoing globalization of the banking industry has raised a number of
questions on the entry motives and strategies of banks going abroad, as well
as the impact of the change of market structures in the banking industry,
that has accompanied this process.
This thesis has started out with giving an overview about what existing
economic literature has to say about a large array of questions arising in
the discussion about multinational banks and then has o?ered an insightful
practitioners’ view on strategic decisions within multinational banks via the
case study of Bank Austria Creditanstalt in Central and Eastern Europe.
In the following this body of work has focused on, directly and indirectly,
shedding new light on three speci?c issues that have arisen in the analysis of
multinational bank behaviour and the consequences of associated changes in
banking markets’ structure.
Chapter 2, after giving an overview about the existing literature on multi-
national banking, yielded the following main insights via the case study con-
ducted on the Austrian market leader in one of the core regions of multina-
tional bank expansion world-wide, Central and Eastern Europe. In general,
a highly complex relationship between a bank’s core business strategy and its
home market, its mode and timing for entry into foreign markets, its strat-
egy in building business in these markets as well as the geographic pattern
of a banks geographic expansion, was found. Retail-oriented and wholesale-
oriented banks seem to di?er signi?cantly in their timing and choice of entry
mode into foreign markets, with wholesale-oriented institutions seeming far
more inclined to grow foreign business organically via Green?eld Investment,
whereas successful retail banking in foreign markets seems to require non-
mobile assets only available via the acquisition of local incumbents. Addi-
157
tionally, having a su?ciently large commercial client customer base to follow
abroad seems to enable wholesale-oriented banks to enter markets at earlier
stages of economic development, as entry into such a market can be pro?table
almost from the beginning with such sources of revenue at hand. Concerning
empirical identi?cation strategies of factors underlying the location decision
of multinational banks, it is concluded from the ?ndings, that, at least for
regions economically integrating and with similar law and bank regulation,
focusing on single host country characteristics might be ine?cient, as location
strategies might a broader regional, not country-speci?c, type
1
.
Chapter 3 theoretically discussed the choice of entry mode of multina-
tional institutions into foreign markets in general, not restricted to the bank-
ing industry. Whereas the existing literature discusses the choice of com-
panies whether to enter a foreign market via the setup of a completely new
structure (Green?eld Investment) or via the acquisition of an incumbent
company in the host market in a completely static way, my proposed model
incorporates a more ”‘dynamic”’ view. In a market that is entered sequen-
tially by foreign companies the choice of entry mode of the early mover
a?ects the entry decision of potential subsequent entrants, where early entry
via Green?eld Investment has the strategic advantage compared to entry via
acquisition, that pro?t-reducing sequential entry via Green?eld Investment
becomes less likely. The basic model thus yields a reasoning, why Green?eld
Investment makes up such a substantial share of foreign direct investment
in general, especially concerning the number of occurrences, whereas recent
theoretical literature has focused on adding further explanations for entry
via M&A, especially focusing on the analysis of asset-complementarity of ac-
quirer and target. It is also found, that a perverse e?ect of limited takeover
possibilities in markets to be entered further increases the attractiveness of
Green?eld Investment for early market entrants. Whereas such a basic model
can explain why Green?eld Investment should be an important entry path
for e.g. the wholesale banking industry, an extension is additionally pro-
posed that yields the tentative prediction, that entry via M&A should be
a favoured entry mode into foreign markets in the retail banking industry.
Incorporating the notion of market-speci?c learning by doing e?ects we ?nd
M&A to be the early entry mode deterring harmful sequential entry, if the
respective industry is characterized by a low degree product di?erentiation
between ?rms in the market. Due to generally assumed low ex ante product
di?erentiation and heterogeneous regulation across countries, we deem this
1
A variety of further results, new or con?rming existing literature, can be found in the
conclusion of the case study.
158
extension to be a good ?t for the analysis of the retail banking industry.
The theory proposed in Chapter 4 o?ered new insights into a type of
banking foreign direct investment, that has not received su?cient formal
treatment in the existing literature though being heavily discussed both in
general business literature as well as empirically, namely banks going abroad
to provide services locally to its existing multinational ?rm customer base.
In my model, which bases on a recent theoretical contribution by Marin
and Schnitzer (2006)[97], a bank’s decision whether to follow its customer
abroad or not for loan provision is discussed. The model yields that this
location decision of banks is shaped by an interaction of client-speci?c and
host country-speci?c characteristics. Tendentially
2
banks will follow their
customers abroad if their multinational customer enters what might be called
a non-fully developed country characterized by a low endowment in human
capital. The argument runs along the line of collateral, where it is derived,
that governments’ of such countries might be more inclined to block out?ow
of valuable asset-embedded human capital, creating the need for banks to
sell such assets of a non-performing loan project locally in the host country.
Doing so e?ciently however requires a physical presence abroad to be able
to identify potential local asset-takers. This line of argument is close to
concerns of practitioners that mentioned, that issues arising about the cross-
border liquidation of collateral are an important reason for banks to follow
customers abroad
3
.
Chapter 5 again did not primarily focus on multinational banking per se,
but uses a theoretical model to analyze the e?ect of general re-organization of
the banking industry on loan supply for small and medium-sized enterprises
characterized by their informational opacity. To be precise the e?ect of M&A
in the banking sector on the former is analyzed theoretically, where the gen-
eral process applies to in-market consolidation as well as foreign bank entry
via M&A. It is shown, that the likelihood of loan provision to small ?rms is
reduced by M&A activity in the banking sector due to two modelled reasons.
For one, a small ?rm is likely to loose its loan relationship with its small
bank when this bank integrates into a large bank structure via either being
a target or an acquirer in an M&A deal, due to loan managers incentives
for specialising on large ?rm loans being larger in large, hierarchical banking
institutions. Incorporating the notion of relationship-based banking for the
SME sector, such that soft, non-veri?able, information becomes better the
2
As the following argument holds true for two out of three types of clients discussed.
3
I’d like to thank Jana Schwarze (Commercial Clients, Stadtsparkasse Muenchen) and
Christoph Schropp (formerly Dresdner Bank AG) for pointing out this issue to me.
159
longer the respective bank-client relationship, I ?nd that these ?rms set free
by the consolidated institution are unlikely to attain loans from other small
banks in the market due to a missing relationship with these institutions.
Concerning the general focus of this thesis, it can tentatively be concluded,
that foreign bank entry via Green?eld Investment should not have an adverse
e?ect on the availability of loan ?nancing for SMEs, whereas entry via M&A
should be harmful to this client group.
Numerous unanswered questions concerning future developments in the
multinationalization of the banking industry persist.
For one, the basic business strategies of banks are undergoing a major
change. Banks increasingly focus on o?-balance sheet activity, abandoning
their role as risk-takers in their ?nancing business by passing on securitized
debt to other ?nancial market participants such as hedge funds. Increas-
ingly leaving such ?nancial services of the balance sheet banks do not act
as ?nancial intermediaries in the classic sense anymore, rather acting as in-
formation brokers and sales channels for other ?nancial market participants.
As a result of this development, banks are in less need of long-term re?nanc-
ing via deposits, shifting their retail strategies to act as a sales channel for
higher-margin investment products such as certi?cates and investment funds.
Instead of re?nancing via deposits banks therefore increasingly re?nance via
the capital market as well as more short-term the interbanking market. As
the classic intermediation role between depositor and borrower therefore be-
comes less important, and services such as asset management and investment
banking services such as bond underwriting seem to require su?cient scale,
an arising specialisation of banks on subsegments of the value chain is proba-
bly just in its beginning. Such specialisation is already found in Central and
Eastern Europe in a recent study by Dinger and von Hagen (2007)[55], who
?nd that old established banks in the region use their large existing branch
network to focus on raising deposits and subsequently transfer these assets to
new (mostly foreign) banks via the interbanking market, whereas the latter
new banks focus on the provision of loans, backed by re?nancing through
the interbanking market. How such changing business strategies of banks
will in?uence the further globalization of the banking industry remains to be
seen.
At the same time new political discussions about restricting entry (espe-
cially via the acquisition of incumbent national ?rms) in key industries have
gained momentum, with examples being the prevented takeover of Span-
ish energy national champion Endesa by German E.ON or recent discussion
about prohibiting majority stakes of foreign investment funds in core in-
dustries in Germany. With the banking industry still perceived to be of
160
strategic importance for the economy, further globalization of the industry
might be made infeasible by political interventionism
4
. Even without such
political barriers, according to the economic literature (e.g. see Berger et
al.(2000)[12]) the globalization of the banking industry might be restricted
to an incomplete level compare to other industries, as informal barriers to
entry in the commercial SME and retail banking sector coupled with a multi-
national banks’ potential partial inability to successfully integrate acquired
smaller local incumbents, might leave at least some banking services markets
in the hands of smaller local players.
Finally, the internationalisation of the banking industry might also come
under scrutiny by typical home countries of such multinational banks, as the
international exposure of home banks might be cause for concern of home
country governments. The latest development in the sub-prime mortgage
loan crisis in the U.S. has left numerous foreign ?nancial institutions
5
in
bad condition. As a result of this exposure of foreign banks to U.S. market
risks, indication has pointed to these institutions also restricting loan business
in their home countries. Such development might lead to political claims
towards domestic institutions to decrease their scope of business in (risky)
foreign markets.
4
Which, at least for the case of foreign bank entry via M&A, could possibly be ratio-
nalized by the ?ndings in Chapter 5 of this thesis.
5
Banks that have publicly been heavily discussed to be severely a?ected by this crisis
are British Northern Rock and German Banks IKB and Sachsen LB.
161
Bibliography
[1] Erste Bank AG. Corporate website.http://www.sparkasse.at/erstebank, 2007.
[2] Jalal Akhavein, Allen Berger, and David Humphrey. The e?ects of
bank megamergers on e?ciency and prices: Evidence from the pro?t
function. Review of Industrial Organization, Vol. 12:95–139, 1997.
[3] Robert Aliber. International banking: A survey. Journal of Money,
Credit and Banking, Vol. 16 (4):661–678, 1984.
[4] Thomas Andersson and Roger Svensson. Entry modes for direct in-
vestment determined by the composition of ?rm-speci?c skills. Scan-
dinavian Journal of Economics, 96(4):551–560, 1994.
[5] Elisa Bain, Justin Fung, and Ian Harper. Multinational banking: His-
torical, empirical and case perspectives. Melbourne Business School
Working Paper 99-05, 1999.
[6] Cli?ord Ball and Adrian Tschoegl. The decision to establish a foreign
bank branch or subsidiary: An application of binary classi?cation pro-
cedures. Journal of Financial and Qualitative Analysis, 17 (3):411–424,
1982.
[7] World Bank. Integration with the global economy. Technical report,
World Development Indicators, 2004.
[8] P.G. Baudino, G. Caviglia, E. Dorrucci, and G. Pineau. Financial FDI
to the eu accession countries. ECB Directorate General International
and European Relations, 19 March 2004, 2004.
[9] Paul Bauer, Allen Berger, and David Humphrey. E?ciency and produc-
tivity growth in u.s. banking. In Harold Fried, C.A. Knox Lovell, and
Shelton Schmidt, editors, The measurement of productive e?ciency:
162
Techniques and Applications, pages 386–413, Oxford, 1993. Oxford
University Press.
[10] William J. Baumol. Business behaviour, value, and growth. Mac Mil-
lan, New York, 1959.
[11] Allen Berger, Carl Bonime, Lawrence G. Goldberg, and Lawrence
White. The dynamics of market entry: The e?ects of mergers and
aquisitions on de novo entry and small business lending in the banking
industry. Technical report, Board of Governors of the Federal Reserve
System, 1999.
[12] Allen Berger, Robert DeYoung, Hesna Genay, and Gregory Udell.
Globalization of ?nancial institutions: Evidence from cross-border
banking performance. Brookings-Wharton Papers on Financial Ser-
vices, Vol. 3:23–158, 2000.
[13] Allen Berger and Timothy Hannan. Using measures of ?rm e?-
ciency to distinguish among alternative explanations of the structure-
performance relationship. Managerial Finance, 23:6–31, 1997.
[14] Allen Berger, Klapper, and Gregory Udell. The ability of banks to
lend to informationally opaque small businesses. Journal of Banking
and Finance, Vol. 25:2127 – 2167, 2001.
[15] Allen Berger, Nathan Miller, Michael Petersen, Raghuram Rajan, and
Jeremy Stein. Does function follow organizational form? evidence from
the lending practice of large and small banks. NBER working paper
No. 8752, February 2002.
[16] Allen Berger, Anthony Saunders, Joseph Scalise, and Gregory Udell.
The e?ects of bank mergers and aquisitions on small business lending.
Journal of Financial Economics, Vol. 50(2):187–229, 1998.
[17] Allen Berger and Gregory Udell. Relationship lending and lines of
credit in small ?rm ?nance. Journal of Business, 68:351–382, 1995.
[18] Allen Berger and Gregory Udell. Universal banking and the future
of small business lending. In Saunders and Walter, editors, Universal
Banking: Financial System Design Rediscovered, pages 558–627, Burr
Ridge, IL, 1996. Irwin Publishing.
[19] Allen Berger and Gregory Udell. Small business credit availability and
relationship lending: The importance of bank organisational structure.
Economic Journal, 112 (477):32–53, 2002.
163
[20] D. Bernheim. Strategic entry deterrence of sequential entry into an
industry. Rand Journal of Economics, 15:1–11, 1984.
[21] S. Bhattacharya and G. Chiesa. Proprietary information, ?nancial in-
termediation and research incentives. Journal of Financial Intermedi-
ation, 4:328–357, 1995.
[22] Jacob Bikker and Johannes Groeneveld. Competition and concentra-
tion in the eu banking sector. De Nederlandsche Bank, 1998.
[23] BIS. Group of ten report on consolidation in the ?nancial sector. Tech-
nical report, Bank for International Settlements, 2001.
[24] Hanneke Bol, Jakob de Haan, Bert Scholtens, and Ralph de Haas.
How important are foreign banks in european transition countries? a
comparative analysis. EconWPA International Finance Series, No.
0209005, 2002.
[25] J.P. Bonin, I. Hasan, and P. Wachtel. Bank performance, e?ciency and
ownership in transition countries. Journal of Banking and Finance,
29:31–53, 2005.
[26] R.A. Brealey and E.C. Kaplanis. The determination of foreign banking
location. Journal of International Money and Finance, Vol. 15 (4):577–
597, 1996.
[27] Claudia Buch. Why do banks go abroad? - evidence from german data.
Kiel Working Papers, No.948, 1999.
[28] Claudia Buch and Gayle de Long. Cross-border bank mergers: What
lures the rare animal? Journal of Banking and Finance, 28:2077–2102,
2004.
[29] Claudia Buch and Alexander Lipponer. FDI versus cross-border ?nan-
cial services: The globalization of german banks. Deutsche Bundesbank
Discussion Paper, No 05/2004, 2004.
[30] G.A. Calvo, L. Leiderman, and C.M. Reinhart. Capital in?ows and
real exchange rates appreciation in latin america - the role of external
factors. IMF Sta? Papers, 40 (1):108–151, 1993.
[31] Mark Casson. Evolution of multinational banks: A theoretical perspec-
tive. In Geo?rey Jones, editor, Banks as Multinationals, pages 14–29,
London, 1990. Routledge.
164
[32] Vittoria Cerasi, Barbara Chizzolini, and Marc Ivaldi. Sunk costs and
competitiveness of european banks after deregulation. LSE Financial
Market Group Discussion Paper 290, 1998.
[33] Judith Chevalier and Christopher Avery. Why do managers undertake
aquisitions? an analysis of internal and external rewards for aquisive-
ness. Journal of Law, Economics and Organization, 14:24–43, 1998.
[34] S. Claessens, A. Demirguec-Kunt, and H. Huizinga. How does foreign
entry a?ect the domestic banking market. Journal of Banking and
Finance, 25:891–911, 2001.
[35] Sophie Claeys and Christa Hainz. Acquisition versus green?eld: The
impact of the mode of foreign bank entry on information and bank
lending rates. ECB Working Paper, No. 653, 2006.
[36] Je?rey Clark. Economic cost, scale e?ciency and competitive viability
in banking. Journal of Money, Credit and Banking, 28:342–364, 1996.
[37] George Clarke, Robert Cull, and Maria Martinez-Peria. Does foreign
bank penetration reduce access to credit in developing countries: Evi-
dence from asking borrowers. World Bank Group Working Paper 2716,
2001.
[38] Rebel A. Cole, Lawrence G. Goldberg, and Lawrence J. Wright.
Cookie-cutter versus character: The micro structure of small business
lending by large and small banks. In Richard W. Lang, editor, Business
Access to Capital and Credit. Federal Reserve Bank of Dallas, 1999.
[39] Bank Austria Creditanstalt. Annual report 2006.
[40] Bank Austria Creditanstalt. 2007.
[41] J.S. Crystal, B.G. Dages, and L. Goldberg. Has foreign bank entry led
to sounder banks in latin america? Current Issues in Economics and
Finance, 8 (1):1–6, 2002.
[42] Elisa Curry, Justin Fung, and Ian Harper. Multinational banking:
Historical, empirical and case perspectives. In A.W. Mullineux and
V. Murinde, editors, Handbook of International Banking, pages 27–59,
UK and USA, 2003. Edward Elgar.
165
[43] B.G. Dages, L. Goldberg, , and D. Kinney. Foreign and domestic
bank participation in emerging markets: Lesson from mexico and ar-
gentina. Federal Reserve Bank of New York Economic Policy Review,
September:17–36, 2000.
[44] Ralph de Haas and Iman van Lelyveld. Foreign banks and credit sta-
bility in central and eastern europe. a panel data analysis. Journal of
Banking and Finance, 30 Issue 7:1927–1952, 2006.
[45] Ralph de Haas and Iman van Lelyveld. Internal capital markets and
lending by multinational bank subsidiaries. De Nederlandsche Bank
Working Paper, 101/June, 2006.
[46] R.T.A. deHaas and I.P.P. vanLelyveld. Foreign bank penetration and
private sector credit in central and eastern europe. Technical report,
De Nederlandsche Bank Sta? Reports No.91, 2002.
[47] Giovanni Dell’Ariccia, Ezra Friedman, and Robert Marquez. Adverse
selection as a barrier to entry in the banking industry. RAND Journal
of Economics, 30 (3):515–534, 1999.
[48] Giovanni Dell’Ariccia and Robert Marquez. Information and bank
credit allocation. Journal of Financial Economics, 72 (1):185–214,
2004.
[49] E. Detragiache, P. Garella, and L. Guiso. Multiple versus single bank-
ing relationships: Theory and evidence. Journal of Finance, 55:1133–
1161, 2000.
[50] Robert DeYoung. Bank mergers, x-e?ciency, and the market for cor-
porate control. Managerial Finance, 23:32–47, 1997.
[51] Robert deYoung and Daniel Nolle. Foreign-owned banks in the us:
Earning market share or buying it? Journal of Money, Credit and
Banking, 28:622–636, 1996.
[52] Emilia Bonaccorsi di Patti and Giorgio Gobbi. Winners or losers? the
e?ects of banking consolidation on corporate borrowers. Journal of
Finance, LXII (2):669–695, 2007.
[53] Douglas W. Diamond. Financial intermediation and delegated moni-
toring. Review of Economic Studies, Vol. 51, No. 3:393 – 414, 1984.
166
[54] Douglas W. Diamond. Monitoring and reputation: The choice between
bank loans and directly placed debt. Journal of Political Economy, 99
(4):689–721, 1991.
[55] Valeriya Dinger and Juergen von Hagen. Risk e?ects of bank special-
ization in central and eastern europe. European Review of Finance and
Economics, forthcoming, 2007.
[56] Marc Melitz Elhanan Helpman and Stephen Yeaple. Export versus
FDI with heterogeneous ?rms. American Economic Review, 94:300–
316, 2004.
[57] Liz Feely and Dave Compare. When worlds collide: culture clash.
Journal of Business Strategy, 24 No.4:9–13, 2003.
[58] Norman Fieleke. The growth of u.s. banking abroad: An analytical
survey. Federal Reserve Bank of Boston Conference Series, October
18:9–40, 1977.
[59] A. Fisher and P. Molyneux. A note on the determinants of foreign
bank activity in london between 1980 and 1989. Applied Financial
Economics, 6:271–277, 1996.
[60] D. Focarelli and A.F. Pozzolo. The determinants of cross-border bank
share holdings: an analysis with bank-level data from oecd countries.
Banca D’Italia Temi di Discussione, 381 - October 2000, 2000.
[61] D. Focarelli and A.F. Pozzolo. The patterns of cross-border bank merg-
ers and shareholding in oecd countries. Journal of Banking and Fi-
nance, 25:2305–2337, 2001.
[62] D. Focarelli and A.F. Pozzolo. Where do banks expand abroad? an
empirical analysis. Universita degli Studi del Molise Economics and
Statistics Discussion Paper, No. 9/03, 2003.
[63] Bank for International Settlements Committee on the Global Finan-
cial System. Foreign direct investment in the ?nancial sector of emerg-
ing market economies. Technical report, BIS, March 2004.
[64] Kenneth French and James Poterba. Investor diversi?cation and in-
ternational equity markets. American Economic Review, 81:222–226,
1991.
167
[65] Robert H. Gertner, David S. Scharfstein, and Jeremy C. Stein. Inter-
nal versus external capital markets. Quarterly Journal of Economics,
109:1211–1230, 1994.
[66] Mariassunta Giannetti and Steven Ongena. Financial integration and
entrepeneurial activity: Evidence from foreign bank entry in emerging
markets. Stockholm School of Economics, mimeo, 2005.
[67] R. Gilbert and X. Vives. Entry deterrence and the free rider problem.
Review of Economic Studies, 53:71–83, 1986.
[68] Lawrence Goldberg and Robert Grosse. Location choice of foreign
banks in the united states. Journal of Economics and Business, 46:367–
379, 1994.
[69] Linda Goldberg. When is u.s. bank lending to emerging markets
volatile? NBER Working paper, 8209, 2001.
[70] Holger G¨org. Analysing foreign market entry: The choice between
green?eld investment and acquisitions. Journal of Economics Studies,
27(3):165–181, 2000.
[71] Christopher Green, Victor Murinde, and Ivaylo Nikolov. The e?ciency
of foreign banks in central and eastern europe: Evidence on economies
of scale and scope. Journal of Emerging Markets Finance, 3 No. 2:175–
205, 2004.
[72] Stuart Greenbaum, George Kanatas, and Itzhak Venezia. Equilibrium
loan pricing under the bank client relationship. Journal of Banking
and Finance, 13:221–235, 1989.
[73] Herbert Grubel. A theory of multinational banking. Banca Nazionale
del Lavoro Quarterly Review, 123:349–363, 1977.
[74] Mauro Guillen and Adrian Tschoegl. At last the internationalization of
retail banking? the case of the spanish banks in latin america. Wharton
School Center for Financial Institutions, University of Pennsylvania,
Working paper 99-41, 1999.
[75] Diana Hancock and James A. Wilcox. The ”credit crunch” and the
availability of credit to small business. Journal of Banking and Finance,
22 (6):983–1014, 1998.
168
[76] Olena Havrylchyk and Emilia Jurzyk. Pro?tability of foreign banks in
central and eastern europe: Does the mode of entry matter? LICOS
Discussion Paper, 16606, 2006.
[77] Jean-Francois Hennart and Young-Ryeol Park. Green?eld vs. acquisi-
tion: The strategy of japanese investors in the united states. Manage-
ment Science, 39 No. 9:1054–1070, 1993.
[78] Bengt Holmstroem and Jean Tirole. Financial intermediation, loanable
funds, and the real sector. Quarterly Journal of Economics, CXII:663–
691, 1997.
[79] Joel Houston, Christopher James, and David Marcus. Do bank internal
capital markets promote lending? Journal of Banking and Finance, 46
(2):899–918, 1997.
[80] Thomas F. Huertas. US multinational banking: History and prospects.
In Geo?rey Jones, editor, Banks as Multinationals, pages 248–267, Lon-
don, 1990. Routledge.
[81] Joseph Hughes and Loretta Mester. Bank capitalization and cost: Ev-
idence of scale economies in risk management and signalling. Review
of Economics and Statistics, 80 (2):314–325, 1998.
[82] W. Hultman and Randolph McGee. Factors a?ecting the foreign bank-
ing presence in the u.s. Journal of Banking and Finance, 13:383–396,
1989.
[83] Wiliam Jackson. Market structure and the speed of price adjustments:
Evidence of non-monotonicity. Review of Industrial Organization, 12
(1):37–57, 1997.
[84] Michael Jensen and Kevin Murphy. Performance pay and top-
management incentives. Journal of Political Economy, 98:225–264,
1990.
[85] Michael C. Jensen. Agency costs of free cash ?ow, corporate ?nance,
and takeovers. American Economic Review, 76(2):323–329, 1986.
[86] Fatih Kasirye Joe Peek, Eric S. Rosengren. The poor performance of
foreign bank subsidies: Were the problems aquired or created? Journal
of Banking and Finance, 23 (2-4):579–604, 1999.
[87] Geo?rey Jones. Banks as multinationals. In Geo?rey Jones, editor,
Banks as Multinationals, pages 1–13, London, 1990. Routledge.
169
[88] Michael Klein, Joe Peek, and Eric Rosengren. Troubled banks, im-
paired foreign direct investment: The role of relative access to credit.
American Economic Review, 92 (3):664–682, 2002.
[89] Paul Krugman. The narrow moving band, the dutch disease and the
competitive consequences of mrs. thatcher: Notes on trade in the pres-
ence of dynamic scale economies. Journal of Development Economics,
27:41–55, 1987.
[90] Myron Kwast, Martha Starr-McCluer, and John Wolken. Market def-
inition and the analysis of antitrust in banking. Antitrust Bulletin, 42
(4):973–995, 1997.
[91] Lawrence J. White Lawrence G. Goldberg. De novo banks and lending
to small businesses. Journal of Banking and Finance, 22 (6-8):851–867,
1998.
[92] Maria Lehner. Entry mode choice of multinational banks. Munich
Economics Discussion Paper, 2007 forthcoming, 2007.
[93] Maria Lehner and Monika Schnitzer. Entry of foreign banks and their
impact on host countries. Munich Economics Discussion Paper, No.
152, 2006.
[94] Ross Levine. Denying foreign bank entry: Implications for bank interest
margins. Working Papers Central Bank of Chile, 222, 2003.
[95] Andzelika Lorentowicz, Dalia Marin, and Alexander Raubold. Owner-
ship, capital or outsourcing: What drive german investment to eastern
europe. Munich Economics Discussion Paper, 2002-03, 2002.
[96] Giovanni Majnoni, Rashmi Shankar, and Eva Varhegyi. The dynamics
of foreign bank ownership - evidence from hungary. The World Bank
Policy Research Working Paper, 3114, 2003.
[97] Dalia Marin and Monika Schnitzer. When is FDI a capital ?ow. Munich
Economics Discussion Paper, 2006-22, 2006.
[98] James Markusen and Anthony Venables. The theory of endowment,
intra-industry and multinational trade. Journal of International Eco-
nomics, 52 (2):209–234, 2000.
[99] Marc Melitz. The impact of trade on intra-industry allocations and
aggregate industry productivity. Econometrica, 71:1695–1725, 2003.
170
[100] Atif Mian. Distance constraints: The limits of foreign lending in poor
economies. Journal of Finance, 61 (3):1465–1505, 2006.
[101] Philip Molyneux, D.M. Lloyd-Williams, and John Thornton. Compet-
itive conditions in european banking. Journal of Banking and Finance,
18 (3):445–459, 1994.
[102] Randal Morck, Andrei Shleifer, and Robert Vishny. Do managerial
objectives drive bad aquisitions? Journal of Finance, 45:31–49, 1990.
[103] D. Morgan, B. Rime, and P.E. Strahan. Bank integration and state
business volatility. Quarterly Journal of Economics, 119 (4):1555–1585,
2004.
[104] F. Moshirian. International investment in ?nancial services. Journal
of Banking and Finance, 25:317–337, 2001.
[105] Thomas M¨ uller. Analyzing modes of foreign entry: Green?eld invest-
ment versus acquisition. Munich Economics Discussion Paper 2001-01,
January 2001, 2001.
[106] Douglas Nigh, Kang Rae Choh, and Suresh Krishnan. The role of
location-related factors in u.s. banking involvement abroad: An empir-
ical examination. Journal of International Business Studies, Fall:59–72,
1986.
[107] Volker Nocke and Stephen Yeaple. Cross-border mergers and acquisi-
tions versus green?eld foreign direct investment: The role of ?rm het-
erogeneity. Journal of International Economics, 72 (2):336–365, 2007.
[108] Daniel Nolle and Rama Seth. Do banks follow their customers abroad?
Federal Reserve Bank of New York, Research Paper 9620, 1996.
[109] Gerhard Smoley Head of Investor Relations Bank Austria Group. Tele-
phone interview 22.01.2007.
[110] U.S. General Accounting O?ce. Foreign banks: Assessing their role
in the u.s. banking system. Report to the Ranking Minority Mem-
ber; Committee on Banking, Housing and Urban A?airs; U.S. Senate,
GAO/GGD-96-26, 1996.
[111] Oliver Pastre. International bank-industry relations: An empirical as-
sessment. Journal of Banking and Finance, 5:65–76, 1981.
171
[112] J. Peek and E.S. Rosengren. Implications of the globalization of the
banking sector: the latin american experience. New England Economic
Review, September/October:45–63, 2000.
[113] Horst Ra?, Michael Ryan, and Frank Staehler. Asset ownership and
foreign-market entry. CESifo Working Paper No. 1676, February 2006,
2006.
[114] Raghuram Rajan. The entry of commercial banks into the securities
business: A selective survey of theories and evidence. In Anthony
Saunders and Ingo Walter, editors, Universal Banking: Financial Sys-
tem Design Reconsidered, pages 282–302. Irwin, 1996.
[115] Lawrence J. White Robert DeYoung, Lawrence G. Goldberg. Youth,
adolescence and maturity at banks: Credit availability to small busi-
nesses in an era of banking consolidation. Journal of Banking and
Finance, 23 (2-4):463–492, 1999.
[116] M. Sabi. An application of the theory of foreign direct investment
to multinational banking in LDCs. Journal of International Business
Studies, 19 (3):433–447, 1987.
[117] Paolo Sapienza. The e?ects of banking mergers on loan contracts.
Journal of Finance, 57 (1):329–367, 2002.
[118] D.T. Savage. Mergers, branch closings and cost savings. Board of
Governors of the Federal Reserve System, 1991.
[119] Jonathan Scott and William Dunkelberg. Bank consolidation and small
business lending: A small ?rm perspective. In Richard W. Lang, editor,
Business Access to Capital and Credit, pages 328–361. Federal Reserve
Bank of Dallas, 1999.
[120] Sherrill Sha?er. Can megamergers improve bank e?ciency? Journal
of Banking and Finance, 17:423–436, 1993.
[121] Farouk Soussa. A note on banking FDI in emerging markets: Literature
review and evidence from m&a data. BIS CGFS Publications, 22, 2004.
[122] Heinrich Stackelberg. Marktform und gleichgewicht. Julius Springer,
Vienna, 1934.
[123] Jeremy Stein. Information production and capital allocation: Decen-
tralized versus hierarchical ?rms. Journal of Finance, Vol. LVIII, No.
5:1891 – 1921, 2002.
172
[124] Joseph Stiglitz and Andrew Weiss. Credit rationing in markets with
imperfect information. American Economic Review, 71 (3):393–410,
1981.
[125] Jean Tirole. The theory of industrial organization. MIT Press, Cam-
bridge, MA, 1988.
[126] Adrian Tschoegl. Who owns the major us subsidiaries of foreign banks?
a note. Wharton Financial Institutions Center Discussion Paper, 03-
11, 2003.
[127] Various. Thinking big (survey: International banking). The Economist,
May 20th, 2006, 2006.
[128] Rudi Vander Vennet. The e?ect of mergers and aquisitions on the
e?ciency and pro?tability of ec credit institutions. Journal of Banking
and Finance, 20 (9):1531–1558, 1996.
[129] Rudi Vander Vennet. Causes and consequences of eu bank takeovers.
In Richard Portes Kees Koedijk, Marco Pagano, editor, The Changing
European Landscape, pages 45–61, Brussel, Belgium, 1998. Centre for
European Policy Research.
[130] Barry Williams. Positive theories of multinational banking: Eclectic
theory versus internalisation theory. Journal of Economic Surveys, 11
(1):71–100, 1997.
[131] Oliver Williamson. Economics of discretionary behaviour: Managerial
objectives in a theory of the ?rm. Kershaw Publishing, London, 1974.
[132] Andrew Winton. Don’t put all your eggs in one basket? diversi?cation
and specialisation in lending. University of Minnesota, 1999.
173
Curriculum Vitae Peter Beermann
Personal Data
Name Peter Beermann
Date of birth January 20th, 1977
Place of birth Munich, Germany
Working Career
Since Nov. 2007 L.E.K. Consulting GmbH, Munich
Associate
Dec. 2002 – Sept. 2007 LMU University of Munich
Chair for International Economics,, Prof. Dr. Dalia Marin
Research and Teaching Assistant
Apr. 2001 – Oct. 2002 ifo Institute for Economic Research, Munich
Directorate of social policy and labour markets
Student worker
May 2001 – Aug. 2001 Logistics Gateway Inc , Toronto (CAN)
Intern Business Development
Apr. 1998 – Apr. 2001 The Boston Consulting Group, Munich
Student worker
June 1999 – Aug. 2000 tecis FDL AG, Munich
Trainee / Junior Consultant
Military Service
09/1996 – 06/1997 Military service in Lenggries and White Sands, Texas (USA)
Academic Career
Apr. 2003 – Jan. 2008 LMU University of Munich
Doctoral Student in Economics
11/1997 – 11/2002 LMU University of Munich
Diploma-Student of Economics
09/1987 – 06/1996 Maria Theresia Gymnasium, Munich
High School Student
doc_465717889.pdf
The ongoing integration of the world economy has not stopped at the real sectors, but has also taken place in the financial sector, where both the crossborder provision of financial services as well as the globalization of banks are on the rise.
Topics in
Multinational Banking
and
International
Industrial Organization
Inaugural-Dissertation
zur Erlangung des Grades
Doctor oeconomiae publicae (Dr. oec. publ.)
an der Ludwig-Maximilians-Universit¨at
M¨ unchen
2007
vorgelegt von
Peter Beermann
Referent: Professor Dr. Dalia Marin
Korreferent: Professor Dr. Monika Schnitzer
Promotionsabschlussberatung: 06. Februar 2008
Contents
1 Introduction 1
2 Multinational Banking - A Literature survey and the case of
Bank Austria in Central and Eastern Europe 6
2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Multinational banking - what do we know? . . . . . . . . . . . 7
2.2.1 Entry motives of multinational banks . . . . . . . . . . 7
2.2.2 Which banks become multinational? Firm and home
country characteristics . . . . . . . . . . . . . . . . . . 10
2.2.3 Which markets attract multinational banks? . . . . . . 12
2.2.4 How do the entry and market strategies of multina-
tional banks look like? . . . . . . . . . . . . . . . . . . 15
2.2.5 How successful are multinational banks operating in
foreign markets? . . . . . . . . . . . . . . . . . . . . . 19
2.2.6 What are the e?ects of multinational bank entry on
the host country? . . . . . . . . . . . . . . . . . . . . . 22
2.3 A case study of successful multinationalization: Bank Austria
Creditanstalt in Central and Eastern Europe . . . . . . . . . . 29
2.3.1 Overview of objects and sources of the case study . . . 29
2.3.2 Bank Austria Creditanstalt: An Overview . . . . . . . 30
2.3.3 Bank Austria Creditanstalt in Eastern Europe . . . . . 31
2.3.4 Entry motives, entry modes and market strategy of
Bank Austria Creditanstalt in CEE . . . . . . . . . . . 32
2.3.5 The future strategy of Bank Austria Creditanstalt in
CEE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.3.6 Same region, di?erent strategy: The case of Erste Bank
AG in Central and Eastern Europe . . . . . . . . . . . 41
2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3 M&A versus Green?eld - Optimal Entry Modes into Markets
with Sequential Entry 47
I
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 The basic model . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.2.1 Analyzing entry (mode) decisions of the sequential en-
trant in period T=2 . . . . . . . . . . . . . . . . . . . 56
3.2.2 The optimal entry mode of the early mover . . . . . . . 61
3.2.3 Benchmark Case (No sequential entry) . . . . . . . . . 64
3.2.4 How does the threat of sequential entry change the
entry mode decision of early movers? . . . . . . . . . . 65
3.2.5 A note on completely endogenous market structure . . 68
3.3 Welfare analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.4 Markets with restricted takeover possibilities . . . . . . . . . . 71
3.5 Country-speci?c learning-by-doing e?ects . . . . . . . . . . . . 73
3.5.1 Sales volumes in period T=1 . . . . . . . . . . . . . . . 74
3.5.2 Sequential entry probabilities in period T=2 . . . . . . 74
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4 When do Banks Follow their Customers Abroad? 84
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.2 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.2.1 Bank pro?t maximization under given liquidation value 91
4.2.2 Loan provision mode choice and endogenous liquida-
tion value . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2.3 When do banks follow their customer abroad? . . . . . 104
4.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5 The E?ect of Bank Sector Consolidation through M&A on
Credit Supply to Small and Medium-Sized Enterprises 115
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.2 Bank Sector Consolidation: An Overview . . . . . . . . . . . . 119
5.2.1 Revenue enhancement as a motive for bank sector M&A120
5.2.2 Cost saving as a motive for bank sector M&A . . . . . 121
5.2.3 E?cient Risk diversi?cation as a motive for bank sector
M&A . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.2.4 Managerial motives for bank sector M&A . . . . . . . . 124
5.3 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.3.1 Inside the small bank . . . . . . . . . . . . . . . . . . . 128
5.3.2 Inside the large bank . . . . . . . . . . . . . . . . . . . 130
5.3.3 Small business ?nancing in small and large banks . . . 138
5.4 The Role of Consolidation . . . . . . . . . . . . . . . . . . . . 139
5.4.1 Changing credit supply within the merged bank . . . . 140
5.4.2 SME credit supply at the market level . . . . . . . . . 141
II
5.4.3 Small ?rms with multiple bank relationships – The odd
?rm out . . . . . . . . . . . . . . . . . . . . . . . . . . 142
5.5 National versus Multinational Consolidation through M&A
and heterogeneous countries . . . . . . . . . . . . . . . . . . . 144
5.6 The Empirics of Small Business Lending and Consequences of
Bank M&A . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
5.6.1 Characteristics of Small Business Lending . . . . . . . 146
5.6.2 Consequences of (Cross-border) Consolidation on Small
Business Lending: The Existing Literature . . . . . . . 146
5.6.3 Bank sector consolidation and relationship manage-
ment strategies of small businesses . . . . . . . . . . . 149
5.7 Findings and Shortcomings . . . . . . . . . . . . . . . . . . . . 149
6 Conclusion and Outlook 157
Bibliography 162
III
List of Figures and Tables
1.1 Foreign Bank Control (% of assets)…………………………………………... …….2
2.1 BA-CA Operations in Central and Eastern Europe…………………………… …...32
2.2 Erste Bank AG Operations in Central and Eastern Europe…………………… …...42
3.1 Time Structure of the Model………………………………………………….. …...52
3.2 The General Structure of Entry Mode Decisions……………………………... …...55
3.3 The Effect of Early Entry Mode on Sequential Entry………………………… …...61
3.4 Comparison of Benchmark and Sequential Entry Case………………………. …..67
3.5 Contingent Sequential Entry Probabilities and Transport Costs……………… …..76
4.1 Time Structure of the Model………………………………………………….. …...91
4.2 Equilibrium Bank Strategies - The Case of Severely Cash-Strapped or
Moderately Cash-Strapped/Small Effort Problem firms………………………
….108
4.3 Equilibrium Bank Strategies – The Case of Moderately Cash-Strapped Firms
with a Large Effort Problem…………………………………………………...
….108
5.1 Volume of M&A Activity in the Banking Sector…………………………….. ….116
5.2 Correlation Analysis of Bank ROE among Nations…………………………... ….124
5.3 Relationship Length and Specialisation Decision…………………………….. ….130
5.4 Bank-Firm Relationships Affected by M&A…………………………………. ….140
IV
Acknowledgements
During the last years working on this thesis I have experienced immense
support, both academically and personally, by a large number of people.
First of all, I am deeply indebted to my supervisor Prof. Dalia Marin for
her wonderful support, giving me the opportunity to pursue my research as
a member of an exceptional faculty. She continuously encouraged me and
o?ered plenty of ideas for this thesis. Discussions with her were extremely
rewarding and instructive. Moreover, her continued rigorous demand for clar-
ity in internal seminars really pushed my understanding and argumentation
to another level.
I am also thankful to my colleagues Alexander Raubold, Thorsten Hansen,
Maria Lehner, Christa Hainz and Georg Gebhardt for countless fruitful scien-
ti?c discussions. Numerous other members of the faculty made this research
experience rewarding.
Special thanks go to Mr Gerhard Smoley, Ms Jana Schwarze and Mr
Christoph Schropp, whose exhaustive practitioner’s insights helped my un-
derstanding of the complex banking industry considerably.
I am deeply grateful to Ms Sabine Schmidberger for encouraging me to
keep working through numerous setbacks one experiences when developing
such a body of research. Last but not least I would like to thank my whole
family for supporting me throughout my academic career.
V
Chapter 1
Introduction
The ongoing integration of the world economy has not stopped at the real
sectors, but has also taken place in the ?nancial sector, where both the cross-
border provision of ?nancial services as well as the globalization of banks are
on the rise.
Though no way a new phenomenon
1
, the internationalisation and multi-
nationalisation of the banking industry has picked up steam (again) in recent
years, with the real new feature being the unprecedented scale of foreign di-
rect investment (FDI) in the banking sector.
This trend was aided by world-wide ?nancial liberalization and deregula-
tion for the banking sector, an increased globalisation of the real sector and
technological advances reducing information costs in international ?nancial
services activities (e.g. Soussa (2004)[121].
In some countries, primarily in Latin America and the former transition
economies in Central and Eastern Europe, foreign direct investment in the
banking sector has gone so far as foreign bank subsidiaries now being the
dominant market players in these countries, cumulatively holding market
shares of close to or more than 50% in many countries in these regions (see
the following table).
1
Jones (1990)[87] traces multinational banking back to the Middle Ages, when Italian
bankers established branches in foreign countries to assist in international banking service.
1
Table 1.1: Foreign Bank Control (% of assets)
0 10 20 30 40 50 60 70 80 90 100
Czech Republic
Hungary
Poland
Turkey
Mexico
Argentina
Peru
Chile
Venezuela
Brazil
Colombia
Malaysia
Korea
Thailand
1994
2001
Source: Inter-American Development Bank
In general ?nancial sector FDI to emerging market economies increased
sharply from 6 $billion in the period 1990-1996 to nearly 50 $billion from
1997 to 2000 (numbers as in BIS report[63]). At the same time an increasing
number of large cross-border merger and acquisition deals has taken place in
the banking industry in OECD countries. In Europe, for example, the top 10
Cross-border M&A deals in banking alone amounted to an investment volume
of 101.4 $billion from 1995 to 2006 (source: The Economist (2006)[127]), with
the largest intra-European bank merger to date having been the acquisition
of HypoVereinsbank AG group (GER) by Uni Credit Group (IT) at a value
of 22.3 $billion.
Amid this surge both new political concerns and scienti?c questions have
arisen. From a normative perspective, governments are in need of answers
concerning the e?ects of foreign bank entry on the economy, while banks are
in need of best practices to shape their international expansion strategies.
Taking a positive perspective, economic science has to come up with expla-
nations concerning politics towards multinational banking (e.g. legal entry
barriers) and the observed patterns of international expansion by multina-
tional banks.
The banking sector is widely perceived to be one of the main important
”‘strategic”’ sectors in an economy, due to its role in providing capital to
2
all other sectors in the economy. Unsurprisingly then, the increasing glob-
alisation of banking has invoked both fears and hopes among governments
concerning the e?ects of foreign bank entry into their local markets. Whereas
there is general hope in emerging market economies, that such entry might
help in accessing international capital sources and aid in the e?ciency de-
velopment of the domestic banking sector, fears have arisen about potential
negative e?ects of foreign bank entry on the survival probability of domestic
banking institutions, potentially harmful ”‘overconsolidation”’ of the sector
and entry e?ects on loan provisioning for informationally opaque market seg-
ments, especially to small and medium-sized enterprises.
Besides scienti?c interest in the latter direct economic policy-relevant
issues, questions on why, when and how banks go abroad are of major interest,
especially because entry motives and strategies might shape foreign bank
entry e?ects on the host economy. For example, as will be discussed in the
literature overview in this thesis, it is observed, that foreign bank entry via
the set-up of a new physical structure (Green?eld Investment) has di?erent
e?ects than entry via the acquisition of a local incumbent bank (M&A). Also,
e?ects of such entry might depend on which banking segment the entrant
foreign bank focuses on in the host country market.
In this thesis, besides trying to give both an overview as well as a prac-
titioners view on the general issues of multinational banking, I focus on
theoretically exploring three questions, concerning banks’ reasoning for in-
ternational expansion due to follow your customer-considerations, the opti-
mal entry mode into foreign markets and how consolidation via M&A in the
banking sector might a?ect credit supply for the most likely victims of such
a development, small and medium-sized enterprises.
Chapter 2 of this thesis starts out with ?rst giving a short literature
overview about six basic questions that arise in the context of multinational
banking, namely
1. What are the motives of banks for expanding internationally?
2. Which banks become multinational players?
3. Which markets attract multinational banks?
4. How do entry and market strategies of multinational banks look like?
5. How successful are multinational banks operating in foreign markets?
6. What are the e?ects of multinational bank entry on the host country?
3
Additionally a case study of a bank highly successful in its international
expansion, Bank Austria Creditanstalt, studying its penetration into the
Central and Eastern European region, is conducted. Based on ?rst-hand in-
formation from a Bank Austria practitioner, I derive a number of insights for
a general discussion about the international expansion strategy of multina-
tional banks. Generally, the study shows how strongly various dimensions of
international expansion strategy are intertwined and how such strategies are
shaped by ex ante characteristics of the bank and its background. Also, the
?ndings shed some further light on the importance of the follow your cus-
tomer motive in banks’ decisions to go abroad, as well as e?cient strategies
in identifying host market characteristics underlying the location decision of
multinational banks.
Chapter 3 tries to answer the question of whether ?rms in general should
enter a host market via the establishment of a completely new structure
(Green?eld Investment) or via the acquisition of a domestic incumbent. I
theoretically discuss this question in a setting of sequential entry and ?nd,
that one reason for the dominance of Green?eld Investment concerning the
number of occurrences might be its strategic advantageous e?ect of deterring
potential further entry into the market. While the theory adds to the general
literature on entry modes in foreign direct investment, the results are valid for
the banking industry, too. Extensions of the model are proposed to discuss
entry modes in industries with limited local takeover possibilities and strong
country-speci?c learning-by-doing e?ects, of which the latter might further
help in understanding entry mode choice in speci?c banking segments such
as retail banking.
Chapter 4 discusses the decision of bank entry into foreign markets under
the follow your customer-motive. This motive is among the most discussed
reasons for banks to go abroad, however no formal theory had been brought
forward to my knowledge. In a setup of a double moral hazard problem
within a real sector ?rm I discuss its home bank’s incentive to follow the ?rm
abroad to provide ?nancing for a host country subsidiary to be set up. I ?nd
that the decision of a bank whether to provided its client with ?nancing from
its home base, via actually establishing a physical presence abroad or not at
all, depends on client, bank and host country characteristics, namely the
relative magnitude of the two moral hazard problems the parent ?rm faces
in interaction with its foreign subsidiary manager, the general e?ciency of the
bank in the liquidation of project assets and the factor endowment or stage
of development of the ?rm’s FDI host country. These single characteristics
are found to not in?uence the banks behaviour uniquely, but rather interact
with each other to shape the respective banks service provisioning strategy.
In chapter 5 I propose a general theory on how active consolidation (via
4
M&A) in the banking sector might a?ect the credit supply for small and
medium sized enterprises (SMEs). Though set up as a general theory for the
banking sector, the model also applies to changes in the industrial organi-
zation in banking markets induced by the entry of foreign banks. I study
this topic by modelling capital allocation decisions within banks, formally
introducing relationship lending considerations. Restricting the analysis to
banks’ incentives to lend to SMEs, the model yields the result that consol-
idation in the form of at least one existing bank becoming part of a larger
organisational structure leads to a potential reduction of credit supply to a
fraction of SMEs in the market. Additionally, I give a smaller literature re-
view on why banks might want to grow larger and how consolidation a?ects
credit supply to SMEs empirically.
I ?nally conclude this thesis by summarizing my ?ndings and discussing
potential future trends that will shape the further evolution of multinational
banking.
5
Chapter 2
Multinational Banking - A
Literature survey and the case
of Bank Austria in Central and
Eastern Europe
2.1 Overview
The purpose of the following chapter is to discuss the main ?ndings of em-
pirical literature on multinational banks as well as to discuss entry motives
and entry strategies by the example of the most important foreign players in
Central and Eastern Europe, namely Austrian Bank Austria Creditanstalt.
For the literature review I will focus on the following questions surround-
ing multinational banking.
1. What are the motives of banks for becoming multinational or generally
acting in foreign markets?
2. Which banks become international/multinational players?
3. Which markets attract multinational banks?
4. How do the entry and market strategies of multinational banks look
like?
6
5. How successful are multinational banks operating in foreign markets?
6. What are the e?ects of multinational bank entry on the host country?
Of course, at least part of these questions are intertwined with each other.
For example, the motives of banks should critically hinge on what kinds of
banks they are, which again also determines whether they become interna-
tional players. Of course, especially taking into account the literature on
vertical versus horizontal FDI, the geographic pattern should also di?er in
the entry motives. Entry motives also will play a role in how the respective
multinational bank will setup their entry and market strategy.
One striking point, is that the e?ect of multinational bank entry should
very much depend on the entry motives. For example, if a foreign bank
enters a host country market to provide services that are neither available
from foreign banks nor close substitutes to the service provided by the latter,
the e?ect of foreign bank entry on the domestic banking sector might be
negligible.
I proceed with a case study of Austrian banks’ entry in Central and East-
ern Europe to shed some further light on these issues. Among the ?ndings
are that home country push factors play a strong role in banks going abroad,
that the availability of follow your customer-strategies signi?cantly supports
international expansion and that the choice of entry mode is strongly con-
nected with the respective market strategy of banks. The latter point also
is a reminder that banks are ?nancial conglomerates, potentially o?ering a
wide variety of services, therefore treating banks as a homogenous group in
a general discussion of multinationalization strategies might be misleading.
2.2 Multinational banking - what do we know?
In the following I want to give a small, by no means complete, review on the
literature about multinational banking.
2.2.1 Entry motives of multinational banks
Broadly speaking, there are two main strands of motives on why banks enter
foreign markets. The one I will not discuss in detail here are managerial
motives. Obviously, entering a foreign market is usually connected with ?rm
growth. Empire-building tendencies of bank managers might therefore be
a simple reasoning why banks want to enter foreign markets, especially if
growth possibilities are restricted in the home country e.g. due to anti-trust
7
considerations. A large literature deals with the empire-building tendencies
of managers and the underlying motives such as status, power, compensa-
tion and prestige of managers of large ?rms (see e.g. Baumol (1959)[10],
Williamson (1974)[131] and Jensen (1986)[85]). An additional managerial
motive for expanding into foreign markets is (potentially ine?cient) risk di-
versi?cation (e.g. Berger et al.(2000)[12]) to prevent cases of bank liquidation
leading to job loss.
The other strand of motives, which I want to discuss in more detail, can
be subsumed under ”‘pro?t-maximization motives”’. One can further di?er
between the motive to expand business in general and the motive to expand
in a speci?c business segment/geographic market.
Concerning motives for general expansion, economies of scale and scope in
various dimensions might play a role. Besides the generally discussed revenue
and cost economies of scale and scope, another focus in the analysis of the
banking industry is on risk diversi?cation economies of scale and scope.
The empirical evidence on the former is mixed as is discussed in more
detail later in of this thesis. However, practitioners strongly support the
view that such economies of scale and scope exist in international expansion.
For example, Spanish bank managers active in the expansion into the Latin
American market perceive a wide variety of such economies e.g. due to the
possibility to develop relatively homogeneous ?nancial products or centralize
back o?ce and transaction processes (Guillen and Tschoegl (1999)[74]. The
majority of studies on this topic also might su?er from old data. Due to tech-
nological advances and changing market possibilities, such economies might
be available to a higher degree today. The evolution of internet banking, the
arrival of Automated Teller Machines (ATMs) and the beginning specializa-
tion of banks along the value chain have most probably increased available
economies of scale (The Economist (2006)[127]). At the same time bank size
and diversi?cation might be an important requirement to be able to place
”‘strategic bets on future markets such as China without putting the whole
bank at risk”’(The Economist (2006), page 4[127]).
The other two well-discussed potential pro?t-enhancement motives for
entering a foreign market are to win new customers in this country (market-
seeking foreign direct investment) or to keep existing domestic customers and
enhance business volume with them (follow your customer-strategy).
The former is the most obvious and generally acknowledged motive for
entry into a foreign market. However, special to the multinational banking
literature, there had been an ongoing debate about whether a foreign bank is
actually capable of successfully entering local retail and commercial banking
markets (e.g. Nolle and Seth (1996)[108]). More recent literature, as well
8
as the overwhelming experience, suggest that at least some foreign banks in
some host markets are able to penetrate local markets on a su?cient scale
(e.g. Berger et al.(2000)[12]).
Compared to the discussion of FDI in other industries, the follow your
customer motive plays a much larger role in the multinational banking lit-
erature. From a ?rm perspective this follow your customer behaviour seems
to have been of large importance for a long time now, as, e.g. in the survey
by Pastre (1981)[111] 52% of U.S. multinational ?rms reported to use one of
their domestic banks for operations in foreign jurisdictions
1
.
The follow your customer motive includes both o?ensive and defensive
strategic traits. For one, a domestic bank might engage in stand-alone non-
pro?table follow your customer-FDI, to secure the respective client’s home
market business with the bank and keep them from switching to another bank
providing global service network capabilities. This so-called defensive expan-
sion approach has been mentioned by e.g. Grubel (1977)[73] or Williams
(1997)[130]. However, following its customer also might be pro?table for
the respective bank per se, as the latter is able to broaden the volume of
business conducted with the respective ?rm, taking over additional trade
?nancing and local cash management services for this ?rm. Also, the geo-
graphical expansion of the bank’s network might attract additional customer
from its home country looking for such ”‘global capabilities”’
2
.
Another pro?t motive for international expansion of banks seems to have
become less of an issue, but has been a big reason for the big wave of inter-
national expansion of U.S. banks in the 1960s and 70s, namely the search for
cheap sources of re?nancing for home market ?nancing activities. This need
arose through ”‘Regulation Q”’ enacted by the Glass-Steagall act in 1933
in the United States, which did put a limit on the interest rates that banks
could pay on deposits in the United States, therefore leading to a shortage
in capital supply for the banks’ loan business. The response was to enter
especially European markets on a large scale to use European deposits to
re?nance U.S. bank ?nancing activities (e.g. Huertas (1990)[80]).
1
Further empirical evidence is discussed at the beginning of chapter 3 of this thesis.
2
Huertas (1990)[80] e.g. cites Frank Vanderlip, former CEO of Citibank on the banks
expansion into Latin America: ”‘ .. I hope to get a very considerable return by o?ering
facilities that other banks cannot o?er to exporters, and thus attract their accounts to
Citibank”’.
9
2.2.2 Which banks become multinational? Firm and
home country characteristics
Just like in the general literature on Foreign Direct Investment (FDI), the
issue of which institution-speci?c characteristics in?uence a banks multina-
tionalization decision has started to garner interest just recently.
This newly arising question is strongly linked with the recent advances
in trade and FDI theory incorporating the fact that ?rms are heterogeneous,
e.g. in the papers by Melitz (2003)[99] and Helpman, Melitz and Yeaple
(2004)[56].
The main proposition of these papers is, that only su?ciently productive
(e?cient) ?rms supply foreign markets. This group can the be further split
into relatively less productive ?rms which will serve foreign markets via ex-
porting, whereas the most e?cient/productive ?rms will establish a physical
presence in foreign markets conducting Foreign Direct Investment.
A similar, and maybe even stronger, view is existent in the multinational
banking literature. As local banks should have inherent advantages over for-
eign banks in the market, due to intimate knowledge about e.g. borrowers’
risks or retail customers’ speci?c preferences, a successful entry by a foreign
bank should only be feasible, if the latter has su?cient advantages in other
bank characteristics to o?set the incumbents’ informational advantages (e.g.
Grubel (1977)[73] or Berger et al.(2000)[12]). Examples for such multina-
tional bank advantages that could be leveraged on a foreign market could
be managerial skill, enhanced risk management and IT systems (e.g. Berger
et al.(2000)[12]) or the ability to re?nance in the capital market or home
deposit market at lower costs.
The empirical literature on e?ciency characteristics is surprisingly scarce
to date. Focarelli and Pozzolo in two studies (Focarelli and Pozzolo (2001)[61]
and Focarelli and Pozzolo (2003)[62]) indeed ?nd, that the more e?cient
3
a
bank, the more likely this institution will run branches and/or subsidiaries
in foreign countries. Also, Buch and Lipponer (2004)[29], studying a sample
of German banks, discover that more pro?table banks
4
are more active
3
Measured as return on assets in these studies.
4
Indeed pro?tability in their study setup is a good proxy for general bank e?ciency, as
the authors control for di?erent business portfolios of the banks as well as for bank size.
Whereas the former therefore controls for pro?t di?erences arising from concentrating on
di?erent segments of the banking industry, the latter ensures that pro?tability is not only
a measure of economies of scale, but rather for underlying X-e?ciency of the respective
bank.
10
internationally, both in the sense that they undertake more foreign direct
investment as well as generate more revenue from international business in
general. However, the authors do not analyze how di?ering pro?tability
a?ects the choice of serving foreign markets predominantly via FDI or cross-
border provision of services.
Bank size per se might be an advantage for banks in their pursuit of inter-
national expansion as well, as large size and scope might enable institutions
to bear larger risks as well as get cheaper re?nancing rates at the capital
market, due to potential economies of risk diversi?cation
5
. A number of
studies indeed ?nd size to have a positive in?uence on the degree of multina-
tional activity of a bank (e.g. Focarelli and Pozzolo (2001)[61],Focarelli and
Pozzolo (2003)[62], Buch and Lipponer (2004)[29] and Tschoegl (2003)[126]).
The latter study even ?nds that, for the US market, it is predominantly the
largest bank from the respective home country, that is most active in the
host market. One reasoning might be, that these respective banks are most
limited in their further domestic growth due to anti-trust regulations.
However, Curry, Fung and Harper (2003)[42] mention that the causality
between the degree of multinationalization and the size of a bank remains a
bit unclear, as bank size might be in?uenced by the fact that the respective
institution is involved internationally and not vice versa.
Furthermore the two studies by Focarrelli and Pozzolo[61],[62] show evi-
dence, that the product focus of banks is a further determinant of the global
scope of its operations. To be precise investment banks, or banks that gen-
erate a high share of their total revenue from non-interest income in general,
are found to be more globalized than traditional loan-processing banks.
Finally, Buch and Lipponer (2004)[29] also show, that previous inter-
national experience increases the probability of a bank to enter a foreign
market
6
.
Incentives and capabilities of banks to become multinational are most
probably endogenous, depending on the characteristics of markets the bank
has been operating in, speci?cally its home market.
The ?nancial development of the home country, the degree of bank sector
competition in the bank’s main market, as well as regulatory conditions in
this home market are deemed by the literature to have an in?uence in both
developing the capabilities of banks to become multinational as well as their
incentive to do so (e.g. Berger et al.(2000)[12], Aliber (1984)[3] and Curry,
5
From this perspective, size is an e?ciency factor, determining available economies of
scale and maybe scope.
6
This ?nding is also backed by information found in conducting the below case study.
11
Fung and Harper (2003)[42]). Theoretically a large market should bread
tendentially larger banks or enable access to larger deposit volumes. Indeed
Brealey and Kaplanis (1996)[26] and Fisher and Molyneux (1996)[59] ?nd a
positive in?uence of home country size on the multinational activity of banks.
Also strong bank sector competition in the host country should lead to
pressure to become more e?cient, while also restricting the possibility to
make abnormal pro?ts, giving incentives and enhancing capabilities to ex-
pand into other (less competitive) markets.
Home market saturation is seen among the main reasons banks start to
look abroad. Guillen and Tschoegl (1999)[74], conducting interviews with
Spanish bank managers, discovered, that the Latin American expansion by
Spanish banks was mainly pursued due to a very saturated Spanish home
banking market featuring strong margin pressure and restricted growth op-
portunities.
Additionally a well-developed capital market might give home banks an
opportunity for cheaper or more ?tting re?nancing options. The role of
home market ?nancial market conditions for general foreign direct investment
has been discussed by Klein, Peek and Rosengren (2002)[88], who ?nd that
the reduction in Japanese Foreign Direct Investment in the 1990s could be
explained by the Japanese banking crisis, which constrained Japanese ?rms’
ability to ?nd su?cient ?nancing for FDI activity.
The role of regulation on incentives to go abroad has already been partly
discussed in the motives for US banks to go abroad (”‘Regulation Q”’). Spe-
ci?c to the U.S., another regulatory restriction for domestic growth might
have played a large role in these banks’ decision to grow abroad. Precisely,
until the Riegle-Neal act in 1994, US banks faced severe restrictions on entry
into multiple US states, making large-scale interstate banking infeasible[12].
So geographic expansion for U.S. banks was mostly only feasible outside of
the U.S..
As mentioned before, another reason for international expansion might be
anti-trust considerations for large domestic banks, which might make further
inorganic (via acquisition) growth in the home market legally infeasible (e.g.
Tschoegl (2003)[126]).
2.2.3 Which markets attract multinational banks?
One area of research on multinational banking, that has already been ex-
plored to a relatively large degree by economic literature, is the location
choice of multinational banks.
The characteristics of an attractive host country can be sorted into two
12
main categories, stand-alone and bilateral host-home country characteristics.
Multiple dimensions of stand-alone characteristics of host country markets
can be distinguished.
Macroeconomic conditions
The macroeconomic conditions attracting banking foreign direct invest-
ment are very much alike these attracting FDI in general.
Numerous studies ?nd a positive in?uence on banking sector FDI of Gross
Domestic Product (GDP) and GDP per capita, e.g. Focarelli and Pozzolo
(2003)[62], Brealey and Kaplanis (1996)[26], Sabi (1987)[116], Buch and Lip-
poner (2004)[29] and Buch and de Long (2004)[28], which is to be expected
from the theoretical literature on horizontal foreign direct investment, e.g.
Markusen and Venables (2000)[98]. Generally, all else equal, a higher GDP
should equal a larger demand for any kind of product. Besides similar rea-
soning, higher GDP per capita might be associated with a higher demand
for sophisticated, high-margin banking services, such as asset management.
In?ation is found to reduce the attractiveness of the market by Focarelli
and Pozzolo (2003)[62], however Buch and Lipponer 2004)[29] do not ?nd a
signi?cant impact of in?ation
7
.
Country risk, as measured by an Index from Euromoney, is found to have
a negative in?uence on international bank activity in a respective country,
be it cross-border lending or bank foreign direct investment, in the study by
Buch and Lipponer[29].
Bank sector and regulation characteristics
One unanimous result across the empirical literature is that countries har-
bouring ?nancial centers (e.g. New York, London, Tokyo) attract a larger
volume of bank sector FDI (see e.g. Focarelli and Pozzolo (2003)[62] and
Buch and de Long (2004)[28]. This easily can be rationalized by a type
forward and backward linkages in the banking industry, as banks buy in-
vestment banking products for their portfolios from other banks as well as
sell such products to other banks. Proximity of respective banks’ investment
banking divisions supports these transactions with heavy information and
trust requirements.
Concerning bank sector regulation, studies ?nd, that the harsher activity
restrictions for multinational banks and the stricter banking regulation in
general, the less bank FDI the respective country will attract (e.g. Focarelli
7
However, Buch and Lipponer argue that positive (higher nominal returns) and negative
(higher instability) e?ects might cancel each other out.
13
and Pozzolo (2003)[62], Berger et al.(2000)[12], Buch and de Long (2004)[28]
and Curry, Fung and Harper (2003)[42]). However, this result does not seem
to hold for less developed countries (e.g. Sabi (1987)[116]).
Other pro?t-in?uencing factors determining whether banks enter the mar-
ket are the size of the banking market (+) (Sabi (1987)[116] the level of con-
centration in the host banking market (-) (Focarelli and Pozzolo (2003)[62]),
and the cost e?ciency of incumbent banks (-) (Berger et al.(2000)[12]). These
in?uence factors obviously shape the expected pro?tability of market entry
through determining the market volume and the degree of competition in the
respective market.
Bilateral home-host country characteristics
Screening the empirical literature, bilateral country characteristics seem
to play a large role in multinational banks’ location decisions, just as the
general theory on Foreign Direct Investment predicts (e.g. Markusen and
Venables (2000)[98]).
Looking at the broad picture one clearly sees strong bilateral patterns
in multinational banking, as found by Soussa (2004)[121]. For example, as
discussed in the case study below, Austrian banks abroad are almost exclu-
sively active in the former transition economies, with 93% of Austrian banks’
FDI stock concentrated in ?ve countries (Czech Republic, Slovak Republic,
Croatia, Poland and Bulgaria). Also, Spanish banks’ multinational activity
is predominantly restricted to Latin America, with a share of 94% of total
outward bank FDI stock in Brazil, Mexico, Argentina, Chile and Colombia.
Similar strict bilateral geographic patterns are found for banks from Belgium
and Italy (strong focus on CEE countries).
From a host country perspective, for example, Spanish banks account for
65% and 58% of total bank foreign assets in Argentina and Brazil, respec-
tively, while e.g. Austrian banks’ foreign asset share in the Czech Republic
is 39%.
A closer, regression analysis-based look on bilateral home-host country
characteristics shows, that the amount of bilateral trade and real sector for-
eign direct investment positively in?uences the volume of bank sector FDI
into the host country (see e.g. Focarelli and Pozzolo (2003)[62], Buch and Lip-
poner (2004)[29], Brealey and Kaplanis (1996)[26] and Fisher and Molyneux
(1996)[59] and Sabi (1987)[116]). This result is consistent with the idea, that
follow your customer-motives play an important role in multinational banks’
location choice.
Additionally common language, low distance and a common legal system
are found to positively in?uence bilateral bank sector FDI in some studies(e.g.
14
Buch and de Long (2004)[28]), however the impact is partly statistically
insigni?cant in other studies (e.g. Focarelli and Pozzolo (2003)[62] and Buch
(1999)[27])
8
.
Looking at how the Latin American and Eastern European markets shape
up concerning the source of inward bank FDI, especially in the Latin Amer-
ican case, language and cultural factors do seem to play a role, as the dom-
inance of Spanish banks among multinational banks in these countries is
blatant. Such a cultural factor is also often mentioned as the reason for the
market leadership of Austrian banks in Eastern Europe, as countries such as
the Czech Republic, Hungary and Slovenia used to be part of the Austrian
Habsburg empire for a long period of time in history.
2.2.4 How do the entry and market strategies of multi-
national banks look like?
Having discussed empirical results on why, which bank, stemming from which
country, enters where, the focus of discussion shifts to the existing literature
on how a bank enters a foreign market.
When deciding upon how to become actively involved in a foreign market,
a bank has two intertwined organizational dimensions to decide on. For one,
the bank has to choose the degree of market involvement (Curry, Fung and
Harper (2003)[42]) it wants to reach. A low level of involvement is achieved
by establishing correspondent banking (cross-border services with the help
of a correspondent incumbent bank in the host country) or by opening a rep-
resentative o?ce or agency. Common to these forms are restricted activities,
with the bank not enabled to engage in deposit taking or direct lending
9
in
the host country. In contrast, via establishing branches or subsidiaries, for-
eign banks are able to ”‘conduct the full range of banking activities”’(Curry,
Fung and Harper (2003)[42]). The di?erence between branch and subsidiary
organizations is, that the former type is not a legally independent organiza-
tional structure, whereas the latter is.
Ball and Tschoegl (1982)[6] ?nd that the main determinant of organiza-
8
In Focarelli and Pozzolo (2003) distance, as expected, in?uences foreign bank activ-
ity negatively and signi?cantly. However, a common language enters insigni?cantly and
ambiguously in the probability of a foreign bank operating either a branch or even a
subsidiary in the respective host country.
9
Agencies might be allowed to engage in commercial lending, but not in other loans or
deposit taking.
15
tional choice in the above dimension is a bank’s experience in the respective
host market as well as its general experience in foreign banking markets.
The former suggests, that markets are entered in a step-by-step approach
starting with low degrees of involvement (e.g. representative o?ce) with the
bank in the following growing its local structure up to possibly establishing
a subsidiary in the market.
Concerning branches and lower-level physical engagement forms, these
will usually be set up via Green?eld Investment.
The second dimension of organizational mode of entry concerns the setup
of a subsidiary in a host market. Entering via the establishment of a sub-
sidiary can be achieved either by setting up a completely new organizational
structure in the market (Green?eld Investment) or via the acquisition of a
local incumbent bank.
Very few empirical studies to date deal with the determinants of entry
modes of multinational banks.
According to de Haas and van Lelyveld (2006)[44], establishing Green-
?eld subsidiaries might be the entry mode of choice, if the parent bank wants
to exercise a high degree of control over the foreign structure. The authors
show evidence for this claim, ?nding that Green?eld subsidiaries are more
closely integrated within the parent organization operationally, having access
to the parent banks’ internal markets for capital and management resources.
In contrast, subsidiaries stemming from the acquisition of local incumbent
institutions enter the multinational bank group with an existing personnel
and capital (deposits and loans) portfolio, and are found to be less integrated
into the parent organizations’ internal markets, and also might need restruc-
turing such as to ?t into the group’s organization and product portfolio. On
a positive side, the promptly available deposits of an acquired bank may en-
able the foreign bank to grow its local loan business faster due to available
local re?nancing
10
. The need for access to the multinational bank group’s
internal capital market might therefore also be less pronounced.
A disadvantage of entry via M&A proposed in the literature especially ap-
plies to entry into non-OECD countries. Whereas acquired banks come with
a portfolio of potentially non-performing loans, a Green?eld subsidiary can
start o? local business without such baggage[35]. This might be an impor-
tant factor, especially when entering countries, that have experienced a loan
crisis lately or harbour a majority of ine?cient banks, also lacking the trans-
parency for multinational entrants to determine their level of engagement in
bad loans
11
.
10
see Curry, Fung and Harper (2003)[42]
11
Indeed, in the case study about Bank Austria Creditanstalt in this chapter, the prac-
16
An important determinant of the entry mode for multinational banks
should also be the respective bank’s segmental strategic focus in the respec-
tive host market. Due to the problems of acquiring soft information about
loan risks, ”‘green?eld banks have an incentive to focus on the most trans-
parent clientele”’(Havrylchyk and Jurzyk (2006)[76]). The two authors, for
banks entering Central and Eastern European banking markets, show indi-
cation, that Green?eld banks on average charge lower interest rates on their
loans than acquired subsidiary banks, which indicates that the former con-
centrate on low risk, informationally non-opaque clients, as this customer seg-
ment should experience more bank competition, and therefore lower interest
margins available, according to theory (Dell’Ariccia and Marquez (2004)[48]).
As the informational advantage of entry via acquisition versus via Green-
?eld investment hinges on the fact of potential customers being informa-
tionally opaque, one might be able to argue, that we should ceteris paribus
observe acquisition to be the dominant mode of entry in countries with in-
su?cient information-providing institutions and a relatively large volume of
informationally opaque potential customers. However these markets might
also be those featuring incumbent banks with a high share of non-performing
loans, which (partially) o?sets the advantage of entering via M&A.
Havrylchyk and Jurzyk (2006)[76] also decompose foreign subsidiaries’
and domestic banks’ pro?ts. A look at their descriptive statistics yields, that
tendentially foreign banks, that enter the foreign markets via acquisition,
show higher net interest margins and higher return on assets than foreign
banks entering via Green?eld Investment . This suggests that banks enter via
acquisition, if they are large, pro?t e?cient and generate a high proportion
of their pro?ts from the traditional banking activity of lending and deposit-
taking.
Some special political developments in countries also play a large role in
the entry mode decision of foreign banks. The main example are the tran-
sition countries in Central and Eastern Europe, where entry via acquisition
of incumbent domestic banks was supported by the large bank privatization
wave in the 1990s, providing available targets for sometimes relatively low,
political prices (ECB (2004)[8]. The above study also claims, that one of
the reasons for entering via M&A in these countries was, that the foreign
entrants main line of business in this region was retail and commercial bank-
ing, operations that require local market knowledge. The authors however
especially caption the importance of the privatization programme ‘All in all,
titioner mentioned uncertainty about the degree of exposure to bad loans of potential local
target banks as a signi?cant barrier to entry via acquisition in Central and Eastern Europe
in the 1990s.
17
the most relevant consideration in the investment strategy of foreign banks
in the accession countries has been to take advantage of the opportunities
provided by privatisation programmes in order to develop a wide and visible
presence in the host markets within a short period of time
12
.
The latter part stresses another crucial advantage of acquisition over
Green?eld Investment, namely entry by acquisition enabling foreign banks
to serve the market on a large scale fast, reducing the time needed to ramp
up business volume to signi?cant levels. For example, a lack of brand name
recognition for an obscure foreign bank trying to enter the retail banking
market via a completely new structure, should lead to slow business growth,
as reputation has to be slowly built up. The argument might however not
hold true for multinational bank icons such as Citibank entering markets har-
bouring incumbent banks with a collective negative reputation, for example
stemming from a recent national banking crisis.
A general empirical literature on entry modes into foreign markets yields
some additional insights, that might also apply to the banking sector.
Hennart and Park (1993)[77], looking at the entry mode decision of
Japanese multinational ?rms into the United States, match ?rm and industry
characteristics with entry mode choice. Results that may also apply to bank-
ing are, that the level of concentration in the respective market positively
in?uences the propensity to enter via Green?eld Investment. Concerning
product strategy, the more similar products of parent and subsidiary, the
more attractive the entry mode of Green?eld Investment compared to ac-
quisition seems to be. However, this might be due to di?ering motives for
acquisition and Green?eld Investment. In di?erence to the latter, the former
might be motivated by trying to add technological knowledge or diversify a
?rm’s business portfolio.
More to the point of a general organizational discussion, the above authors
also ?nd that Green?eld Investment is chosen if operations in the host market
are of small volume. This points back to the discussion about agencies and
branches versus subsidiaries, where the former operations are feasible for low
levels of activity in the host market and can most easily and cheaply be
achieved by setting up this small structure from scratch.
Andersson and Svensson (1994)[4], for a sample of Swedish multinational
?rms, come up with evidence, that older ?rms with strong organizational
skill
13
tend to enter foreign markets via acquisition of local incumbents,
whereas the role of ?rm size in the decision whether to enter via Green?eld
investment or acquisition is unclear. Concerning host country characteristics,
12
(ECB (2004), page 2)[8]
13
The authors use the number of existing a?liates as a proxy for organizational skill.
18
the authors ?nd that acquisition is the preferred mode of entry for developed
countries, as the probability of entering that way positively and signi?cantly
depends on GDP per capita.
Another dimension of entry strategy is concerned with which segment a
bank wants to serve in the respective host market. Tschoegl (2003)[126] dis-
cusses an interesting example how such a decision is, besides obvious market
characteristics, shaped by institutional/regulatory market conditions. Swiss
banks and Deutsche Bank were reluctant to get into retail business in the
United States because of fears they might run into problems with U.S. bank-
ing regulation agencies as an universal bank, as they already were heavily
involved in the securities market in the United States
14
.
2.2.5 How successful are multinational banks operat-
ing in foreign markets?
Having discussed what shapes the structure of multinational banking, I now
take a look at what literature has to say about the success of international
expansion of multinational banks. The following literature gives insights into
how the relative e?ciency of banks in foreign markets look like.
Results by Claessens et al.(2001)[34] suggest, that the relative perfor-
mance of multinational banks compared to local banks in foreign markets is
to some degree host country speci?c. The authors, using a large sample of
7,900 bank observations from 80 countries, ?nd evidence that foreign banks
are more pro?table
15
than domestic banks in developing countries
16
. How-
ever, this ?nding is turned upside-down for developed countries. The latter
result has been seen as puzzling, as from a Bertrand competition-point of
view there should then be no scope for pro?table multinational bank entry
into these developed countries
17
.
14
The Glass-Steagall Act prohibited universal banking structures and was only repealed
in 1993.
15
The authors discuss interest margins, tax payments and general pro?tability as vari-
ables for pro?t-e?ciency of banks.
16
This result is replicated by a number of other empirical studies e.g. Bonin, Hasan and
Wachtel (2005)[25] and Majnoni, Shankar, Varheggyi (2003)[96].
17
However, this need not be a real puzzle. As the banking industry can be divided into
multiple heterogeneous service segments, one could argue that multinational bank entry
into developed countries is mostly restricted to segments of the industry, where pro?tability
19
Berger et al.(2000)[12] use a more advanced estimation strategy deriv-
ing banks’ X-e?ciency from a standard banking cost function
18
and further
break down the analysis to country levels within developed countries. They
focus on ?ve countries, France, Germany, Spain, UK and the United States.
The authors generally con?rm the result by Claessens et al.(2001)[34], that
domestic banks are more e?cient than foreign banks in developed countries,
with the notable exception being Spain.
Decomposition into pro?t and cost e?ciency gives a hint on where this
pro?t di?erences between domestic and foreign banks stem from. In all of
the analyzed countries the average domestic bank shows signi?cantly higher
pro?t e?ciency than the average foreign bank. However, in Spain, this is
more than o?set by a lower cost e?ciency of domestic banks. Cost e?ciency
in the U.S. is also lower for domestic banks (2,9% higher costs), however
this is vastly lower than the pro?t e?ciency advantage for these domestic
institutions (25,1% higher pro?ts). Therefore the authors argue, that the
cost disadvantage might not be due to ine?ciency but rather ”‘these high
expenses more likely re?ect e?orts to produce a quality or variety of ?nancial
services that generate substantially greater revenues”’
19
.
One interesting result concerns home country e?ects in the discussion of
relative e?ciency. Though, as laid out, foreign institutions are on average
less e?cient than domestic banks in most developed countries, this result
does not hold for multinational banks from the U.S.. Indeed, Berger et
al.(2000)[12] observe, that U.S. banks are more e?cient than domestic banks
in France, Germany and Spain, the only exception to this pattern being the
UK. It therefore seems that U.S. banks are so overwhelmingly more e?cient
in general, that they are able to even outperform domestic banks in their
own backyards in most countries.
is lower. Domestic banks might voluntarily leave these segments to foreign players, instead
growing their business in other ?elds of banking. Also, the empirical literature partly
su?ers from the problem, that the majority of multinational banks entered foreign markets
rather recently. Therefore, these banks might predominantly be in a phase of pursuing
aggressive growth strategies in respective foreign markets, asking for low interest spreads
and providing services at low prices to attract away customers from incumbents.
18
The speci?ed cost function uses four variable outputs (consumer loans, business loans,
real estate loans and securities), one ?xed output (o?-balance-sheet activity), two ?xed
inputs (physical capital, ?nancial equity capital) and three variable inputs (purchased
funds, core deposits and labor). This setup therefore at least partially controls for di?erent
business portfolios/strategies of the respective banks.
19
Berger et al.(2000), page 57
20
Concerning developing countries, Green, Murinde and Nikolov (2004)[71],
exploring economies of scale and scope achieved by domestic and foreign
banks in selected Central and Eastern European countries from 1995-1999,
?nd selective counter-evidence on whether foreign banks are generally more
e?cient than domestic banks in lesser developed countries. The authors do
not ?nd signi?cant di?erences in scale and scope e?ciency between domestic
and foreign banks in the analyzed region. However, their results should be
handled with care, as due to the nature of the transition markets, foreign
banks had just recently entered the market prior to the observation period,
probably still in the process of reaching their e?cient scale and scope in
respective markets.
Havrylchyk and Jurzyk (2006)[76] study the relative e?ciency of for-
eign banks in Central and Eastern Europe in the period 1993-2004. Using
BankScope data they measure pro?tability via return on assets (ROA). They
?nd that the mode of entry plays an important role in the relative e?ciency
of foreign banks. Whereas foreign bank subsidiaries resulting from the acqui-
sition of domestic incumbents do not signi?cantly di?er from other domestic
banks concerning pro?tability, foreign subsidiaries established via Green?eld
Investment are signi?cantly more pro?table than domestic banks
20
. How-
ever, decomposition of the pro?tability variable, shows that this di?erence
in pro?ts might not be due to ine?ciency but rather due to di?erent strate-
gies and business segments. Green?eld banks seem to focus on low-cost, low
risk business segments, which do yield lower risk and service premia, but
these premia seem to be too low in transition economies for the risk born via
non-performing loans and high overhead costs
21
.
The two authors also discover, that Green?eld foreign banks show sig-
ni?cantly lower deposit-to-asset rates than either ”‘takeover foreign banks”’
or domestic banks (60% versus 79/76%). Whether this is by choice, as the
former banks want to re?nance via other sources, or by their ineptitude to
raise su?cient volumes of deposits, e.g. due to a missing large scale branch
network, is at ?rst sight debatable. A lack of access to local deposits in gen-
eral might be one of the reasons for the relatively poor performance of foreign
banks in developed markets. DeYoung and Nolle (1996)[51] ?nd that foreign
banks in the US work under an ine?cient input mix, predominantly having
to rely on re?nancing via purchased funds, whereas local banks use domes-
tically raised deposits to a much higher degree. For the US, this even seems
20
”‘Green?eld foreign banks”’ show a mean return on asset ratio of 1.45, whereas do-
mestic and ”‘Takeover foreign banks”’ show mean ROA of 0.86 and 0.87 respectively.
21
High overhead costs might stem from focusing on loan projects that require a personal
gathering of information about the respective client.
21
to hold true for ”‘takeover foreign banks”’. Peek et al.(1999)[86] ?nd that
foreign banks in the US tend to acquire targets that show an above-average
reliance on purchased funds, resulting in a low deposit-to-asset ratio.
2.2.6 What are the e?ects of multinational bank entry
on the host country?
From a political point of view, the main question concerning multinational
banking is, which e?ects the entry of foreign banks into respective markets
has on economic conditions. With banks still being one of the main providers
of ?nance to the real sector, questions about the entry e?ects on domestic
banks’ performance and loan supply are heavily discussed.
The e?ects of multinational bank entry depend on the respective entrants’
(relative) e?ciency, the product/segment strategy they implement and what
kind of entry mode they choose.
One commonly accepted e?ect of entry into any oligopolistic market is,
that market power of incumbents diminishes and market volumes increase, if
the entrant is su?ciently e?cient to put competitive pressure on the incum-
bents. Due to competitive pressure, incumbents additionally might or might
not be incentivised to operate more e?ciently, either by bank managers being
pushed to operate the structure more e?ciently, or by increasing/decreasing
the need/incentive for incumbents to invest in e?ciency-enhancing new tech-
nologies and practices.
These e?ects should theoretically be costly for the domestic banking sec-
tor up front, but should at ?rst sight improve loan market conditions for
borrowers.
Systematically, the potential e?ects of bank entry can be decomposed
into the direct market e?ect via the introduction of an additional (Green?eld
Investment) or at least the e?cient restructuring of an existing local bank (ac-
quisition), an indirect e?ect induced by domestic incumbents adapting their
strategies in this case, and some direct spillovers of banks technology and
institutional requirements on domestic banks’ and domestic regulatory/legal
institutions.
An increasing theoretical literature has provided both a basis for further
(and ex post explanation for existing) empirical analysis.
Concerning the e?ect of entry on domestic banking sector e?ciency,
Lehner and Schnitzer (2006)[93] in a setup of competition in horizontally
di?erentiated products, discuss the direct e?ects of increased competition
and direct spillovers in screening technology on the e?ciency of the domestic
22
banking sector, where the two direct e?ects have the negative side-e?ect of
reducing incumbent banks’ incentives to invest in screening themselves. Dif-
fering between e?ects from entry via Green?eld Investment and acquisition,
the authors ?nd, that for the case of weakly competitive market (high product
di?erentiation) tendentially entry via acquisition is relatively harmful in less
developed countries, whereas it is less harmful in developed countries. For
the case of low product di?erentiation Green?eld Investment is the welfare
maximizing mode of entry in all kinds of countries.
The e?ect of foreign bank entry on incumbent banks’ lending practices
is theoretically analyzed by Dell’Ariccia and Marquez (2004)[48], who de-
velop a model capturing information asymmetries in loan markets. One of
their main results is that entry by an ”‘uninformed”’ outsider (multinational
bank) should lead to incumbent banks shifting their loan portfolio towards
more informationally-opaque borrowers, due to increased competition in the
segment of non-opaque potential clients. From a general perspective this
could be seen as a bene?cial indirect e?ect of entry, if the opaque segment
had previously been ine?ciently loan-constrained by incumbent banks.
Claeys and Hainz (2006)[35], building on the above setup, develop a model
to discuss di?ering e?ects of foreign bank entry via Green?eld and acquisition
on bank lending rates in the respective host market. They ?nd that foreign
bank entry reduces required lending rates by incumbent domestic banks,
more so if the majority of entry is via Green?eld Investment.
Proceeding to test their hypothesis empirically, Claeys and Hainz (2006)[35]
indeed ?nd, that a higher foreign bank share in loans reduces average lending
rates in the respective market. Also they show evidence, that ”‘Green?eld
foreign banks”’ charge higher lending rates than ”‘takeover foreign banks”’.
Concerning di?ering competition e?ects of entry via Green?eld or acquisi-
tion, the former is observed to reduce average domestic bank lending rates
signi?cantly more than the latter.
Havrylchyk and Jurzyk (2006)[76] also discuss the di?ering e?ect of dif-
ferent entry modes on domestic banks’ performance. They discover that, no
matter the entry mode, a higher market share of foreign banking increases
costs for domestic banks. The authors note that this might seem to be coun-
terintuitive at ?rst sight, as one would expect higher competition to lead to
higher cost e?ciency, however they argue that this cost increase might be
short-term, due to domestic banks’ arising need for investing into competitive
risk, IT and management systems. In contrast to a high share of ”‘takeover
foreign banks”’, a large market share of ”‘Green?eld foreign banks”’ addi-
tionally decreases domestic banks’ non-interest income but also their loan
loss provision volume. In total, however, the e?ects seem to balance out each
23
other, as domestic banks’ pro?ts seem to be una?ected by the foreign banks’
market share in the respective market.
Other empirical studies ?nd that entry of foreign banks reduces the prof-
itability of domestic banks. Additionally to the result from Claeys and Hainz
(2006)[35] e.g. Claessens et al.(2001) [34] ?nd that this reduction in prof-
itability is mainly due to reduced net interest margins, a hint that the e?ect
on pro?tability indeed works through a reduction in market power of incum-
bent banks in the provision of loans and/or deposit taking.
The latter authors ?nding can also be perfectly related to the miss-
ing ”‘foreign bank market share-e?ect”’ found by Havrylchyk and Jurzyk
(2006)[76]. Indeed Claessens et al.(2001)[34] ?nd that the market share of
foreign banks is not an important determinant of domestic banks pro?tabil-
ity, but rather only the number of foreign banks has the expected negative
signi?cant e?ect, suggesting that the threat of foreign banks taking market
shares away already induces domestic banks to reduce net interest spreads,
allowing them to keep domestic bank market shares at a high level.
Another accommodating result is found by Levine (2003)[94], who uses
a unique data set on cases of regulatory institutions denying foreign bank
entry in 47 developed and less-developed countries to analyze the relation-
ship between political entry restrictions for foreign banks and bank interest
margins in a respective country. The author shows evidence, that the more
restrictive entry regulation for foreign banks in a country, the higher the net
interest margins for banks in the market. Foreign bank entry also is found to
be special, as restricting domestic bank entry does not alter operating banks
net interest margins. The results further con?rm the hint from Claessens
et al.(2001)[34] that the contestability of a respective host market primarily
determines the competitive behaviour of operating banks, not the actually
incurring amount of entry into the market. It seems that the existence of
multinational banks threatening to enter a market already disciplines incum-
bent banks.
One important point of note concerning the e?ects of foreign bank entry
is, that such entry from a global view went hand in hand with a consolidation
of the international banking industry. The chapter on bank sector consolida-
tion in this thesis discusses the e?ects of general consolidation in the banking
industry in detail. In that chapter the potential e?ect of multinational bank
entry on the availability of loans to a speci?c segment of the market, infor-
mationally opaque small and medium-sized enterprises (SMEs), is discussed
in detail.
Generally, foreign bank presence is found to increase access to loans, at
least for larger and transparent ?rms in emerging market economies (see e.g.
Mian (2006)[100], Giannetti and Ongena (2005)[66], Clarke et al.(2001)[37]).
24
Concerning informationally opaque smaller ?rms, e.g. Mian (2006)[100] how-
ever ?nds that
‘greater cultural and geographical distance between a foreign bank’s head-
quarters and local branches leads it to further avoid lending to ”information-
ally di?cult” yet fundamentally sound ?rms requiring relational contract-
ing”’
suggesting that the bene?cial e?ect of foreign bank entry might not extend
to this segment of the market. This results is independent of bank size, bank
risk preferences or legal institutions in the home country. This biased lending
strategy e?ect of distance also is identi?ed to be large enough to completely
exclude some types of borrowers in the economy from foreign bank ?nance.
Concerning the increased availability of bank ?nancing for at least a
fraction of borrowers in the host market through MNB entry, it at ?rst
sight seems as this is simply due to an underlying capital transfer into the
host country via the foreign banks’ internal capital market. However, as
Havrylchyk and Jurzyk (2006)[76] discovered foreign banks to predominantly
re?nance their local lending activity with local deposits, the international
capital transfer through entering banks seems less pronounced than origi-
nally thought. Especially for banks entering via acquisition the deposit-to-
asset ratio of 79% suggests minimal cross-border capital transfers within the
banks’ activity. However, if entering foreign banks are more e?cient screen-
ing potential borrowers than existing incumbents the availability of deposits
for commercial loan ?nancing might rise in total, as owners of liquid assets
might be more willing to extend deposits to the banking system (instead of
e.g. transferring assets abroad or invest in government bonds).
Finally, one topic that has also received a substantial amount of interest
in the economic literature is how the presence of foreign banks a?ect the
general stability of the host country banking sector and ?nancial market.
The empirical studies to date focus on the two main recipient less-developed
regions, Latin America and Central and Eastern Europe (CEE).
Theoretically, two speci?c advantages of multinational banks compared to
local domestic banks might determine whether credit supply is stabilized by
foreign banks or not. On the one hand, a multinational bank might have ad-
ditional/cheaper sources of re?nancing, such as better access to international
capital markets or to deposits in the home country or other countries of op-
erations. Therefore these banks might be able to support their subsidiaries
in cases of market-speci?c liquidity/bank crisis, therefore dampening local
shocks. On the other hand, due to operating in multiple markets, the bank
might transfer liquidity/assets from badly-performing/crisis-plagued coun-
25
tries to other regions, therefore strengthening local shocks.
Supporting a subsidiary should generally be a question of a parent bank’s
capability to do so, whereas substituting e?ects should be seen as an incentive-
based decision within in the bank structure, depending on banks’ opportunity
costs of keeping capital in a respective host country subsidiary.
Due to multinational banks operating in multiple regions, they might
therefore transfer shocks from one region to the other via reallocation of
assets within their international internal capital market.
Recent theoretical work by Morgan et al.(2004)[103], who develop a model
of a multinational bank (along the general banking model of Holmstr¨om and
Tirole (1997)[78]), that rebalances its international bank capital reacting to
shocks in bank and real sector capital in respective regions, show what kind of
country-speci?c shocks might be dampened or strengthened by the presence
of multinational banks. The model predicts that multinational banks dampen
local bank-capital shocks by supporting their local subsidiaries but increase
the volatility of the business sector by reallocating bank capital away from
regions experiencing real sector capital shocks.
Empirically, for Latin American countries, Dages et al.(2000)[43], Peek
and Rosengren (2000)[112], Goldberg (2001)[69], Crystal et al.(2002)[41] and
other studies, ?nd strong evidence that foreign bank presence increases the
stability of domestic banking sectors. Foreign banks seem to show stronger
and less volatile credit growth, and positive growth even in periods of ?-
nancial market crisis
22
. However, this line of reasoning does not seem to be
speci?c to the ownership structure of banks, but rather to the underlying rel-
ative ?nancial health of foreign banks in this region. Dages et al.(2000)[43]
observe that domestic banks show the same involatile behaviour as foreign
banks, if they are characterized by similar health e.g. similar levels of shares
of non-performing loans in their portfolio.
For the Central and Eastern European region, de Haas and van Lelyveld
(2006)[44] reinforce the notion of foreign banks as a stabilizing force as they
are found to keep up lending volume during times of ?nancial distress whereas
domestic banks strongly contract lending. However, this seems to hold true
for ”‘Green?eld foreign banks”’
23
only.
22
Indeed, it seems that foreign banks see these occasions as opportunities to expand
their market shares at the expense of ?nancially stricken local banks.
23
Again this is in line with the results of Havrylchyk and Jurzyk (2006)[76], which show
that Green?eld banks are much more embedded in a multinational bank group’s internal
capital market, whereas acquired banks within the group seem to be organized as rather
independent capital centers.
26
Home market and parent bank e?ects, a?ecting the whole bank group via
internal capital markets, also signi?cantly in?uence the lending activity of
multinational banks abroad.
De Haas and van Lelyveld (2006)[44] ?nd evidence for both a substitution
and a support e?ect
24
, depending on parent bank and home market condi-
tions. Concerning the former, they ?nd that foreign banks reduce credit
supply in foreign markets if GDP growth in the home country accelerates,
leading to more potential value-adding business in the home country market.
Again, this substitution e?ect only applies to ”‘Green?eld foreign banks”’.
The capability to support a subsidiary should critically hinge on the parent
bank’s ?nancial status. Indeed, the authors ?nd that parent banks showing
strong ?nancial health
25
have subsidiaries in the CEE region growing credit
volume faster than subsidiaries of weak parent banks. The latter result does
hold for all kinds of foreign bank subsidiaries, however the e?ect is more
pronounced for Green?eld operations.
The majority of empirical studies on home market e?ects con?rm the
results concerning the substitution hypothesis, as worsening home country
conditions seem to have led banks to enlarge their lending activity in foreign
markets (e.g. Moshirian (2001)[104] and Calvo et al.(1993)[30]).
Summing up, concerning the e?ect of multinational bank presence on the
stability of local banking markets, such presence seems to increase stability in
the analyzed less developed regions during times of ?nancial distress. How-
ever, there is strong evidence that, while foreign bank presence might dampen
?nancial crisis in the respective host countries, it could also strengthen or
even import these when home or third country market developments lead the
bank to substitute business and assets from the foreign to its home market
or third countries. The advantages and disadvantages of multinational bank
presence strongly depend on the level of integration of the respective local
subsidiary in the bank group’s international internal capital market. Both
disadvantages and advantages seem to be less pronounced for foreign sub-
sidiaries established via the acquisition of local incumbent banks, as these
structures seem to be more or less ?nancially segregated from their parent
banks.
After discussing what existing literature has to say about multinational
24
The following results are also con?rmed in a newer study by the same authors (de
Haas and van Lelyveld (2006b)[45].
25
The authors use the ratio of loan loss provision over net interest revenue as a proxy
for ?nancial health.
27
banking, the chapter now turns to a practitioner-oriented view, discussing
the behaviour of multinational banks with the help of a case study of one of
the success stories of multinational banking, Bank Austria Creditanstalt and
its successful expansion into Central and Eastern European markets.
28
2.3 A case study of successful multination-
alization: Bank Austria Creditanstalt in
Central and Eastern Europe
2.3.1 Overview of objects and sources of the case study
One of the hot spots of the evolution of multinational banks in recent years
have been the former Communist countries in Central and Eastern Europe.
Nowadays, concerning the scope of internationalization of their banking sys-
tem, these countries are special in that their banking system is dominated by
foreign bank subsidiaries. For example Bol, de Haan, Scholtens and de Haas
(2002)[24],for the year 2000, ?nd foreign bank asset share in total banking
assets to be 54% in Central Eastern Europe and 87% and 77% in South East-
ern Europe and the Baltic States respectively. This dominance holds true in
large countries like Poland (69%) and the Czech Republic (66%).
Within this environment one other fact of note is that the leading home
country of multinational banks operating in these transition economies is
Austria, with the regional market leader in this region being Bank Austria
Creditanstalt (BA).
These two facts rationalize choosing the respective bank in the respec-
tive region for a case study on Bank Austria’s operations in these (former)
transition countries.
The following insights have been won by screening annual reports and
presentations of Bank Austria and its Austrian competitor Erste Bank, but
the major insights were won by interaction with Mr Gerhard Smoley, Head
of Investor Relations of the Bank Austria Group at this time
26
.
I proceed as follows. First a general overview about Bank Austria Cred-
itanstalt (BA) and its history is given. In the following the focus is on the
bank’s operations in the Central and Eastern European (CEE) region, giving
an overview of the development of the bank’s market position. A discussion
of BA’s entry motives, modes and market strategy follow, yielding insights
into how, from a practitioner’s view, home country, host country and bank-
speci?c characteristics determined the internationalization strategy of Bank
26
The following insights strongly base on telephone interviews[109] as well as additional
email communication.
29
Austria. After discussing the future strategic focus of Bank Austria Credi-
tanstalt in the region, an overview of operations of Erste Bank, BA’s main
competitor in the CEE region concerning size, is given, allowing some in-
teresting comparisons between two very di?erent, yet both highly successful
market entry strategies. The case study ?nally is concluded by deriving gen-
eral insights into topics in multinational banking, that can be won from the
proposed case study.
2.3.2 Bank Austria Creditanstalt: An Overview
Bank Austria Creditanstalt today is the leading bank in its original home
country Austria (1.8 million clients at a country population of 8 million) and
what it now deems as its ”‘second home market”’[40] Central and Eastern
Europe
27
. Since November 2005 Bank Austria Creditanstalt, formally a part
of the HypoVereinsbank (DE) group, has become a member of the UniCredit
(IT) banking group via the latter acquiring the former. 95% of Bank Austria
shares are now held by UniCredit with 5% in free ?oat.
Concerning historical roots, Bank Austria was founded in 1991 via the
merger of ”‘Oesterreichische Landesbank”’, ”‘Zentralsparkasse”’ and ”‘Kom-
merzialbank”’. In 1997 Bank Austria took over the Austrian government’s
shares in ”‘Creditanstalt”’ fully integrating into the Bank Austria Credi-
tanstalt group in 1999. As ”‘Creditanstalt”’ was privatized by the Austrian
government only in 1990, all formerly independent parts of the newly arising
bank had formally been state-owned.
In the year 2000 Bank Austria Creditanstalt merged with HypoVereins-
bank and became the competence center for CEE business of the group.
BA fully concentrated on the Austrian as well as 11 selected Eastern Eu-
ropean countries, taking over HVB business in these countries while trans-
ferring other international business to the HVB organization. Bank Austria
embraced this friendly merger, as management saw Bank Austria’s further
growth possibilities limited due to the fact, that Bank Austria stand-alone
had grown too large for its home market but was too small to establish
strong operations in an integrating European banking market
28
. After the
HVB-UniCredit merger in 2005, Bank Austria became part of UniCredit
27
In Bank Austria Group de?nition this also includes CIS countries as well as Turkey.
28
Additionally Bank Austria had become too small for the risk it carried in its portfolio.
Especially its exposure to risks in the North American market was too large for further
growth stand-alone.
30
group structure and now acts as the primary holding and operations center
for the group’s Central and Eastern European business[40].
2.3.3 Bank Austria Creditanstalt in Eastern Europe
Bank Austria Creditanstalt was the ?rst mover in the Central and Eastern
European states, especially with Creditanstalt already having had established
agencies in Hungary (1975), Prague and Moscow (both 1987) even before the
fall of the iron curtain. Creditanstalt also was the ?rst foreign bank to take
over a domestic target in the CEE states, acquiring a majority stake in the
Slovenian ”‘Nova banka”’ as early as 1992[40].
In general, between 1989 and 1991 both Bank Austria and Creditanstalt
started to enter markets in Central and Eastern Europe on a large scale.
Nowadays Bank Austria Creditanstalt is by far the market leader, not
only among foreign but all banks, in Central and Eastern Europe, with to-
tal assets of 41bn and 39.000 employees in 1,800 branches serving roughly
18 million clients in the region. As the CEE competence center of Uni-
Credit group, this includes former subsidiaries of UniCredit and HypoVere-
insbank.Concerning assets, BA-CA-group subsidiaries are the No.1 bank in
Croatia, Bulgaria and Bosnia-Herzegowina, and are among the ?ve largest
banks in eight countries in total[39].
Table 1 gives a detailed look at the group’s positioning in the respective
countries
29
.
29
data as reported 31.12.2006
31
Table 2.1: BA-CA Operations in Central and Eastern Europe
Country Subsidiaries Total
Assets
(€bn)
Branches Market
position
Bosnia • HVB Central Profit Banka
• Nova Banjalucka Banka
• UniCredit Zagrebacka Banka
1.7 180 No. 1
Bulgaria • UniCredit Bulbank 4.2 300+ No. 1
Croatia • Zagrebacka banka 10 127 No. 1
Czech
Republic
• HVB Bank Czech Republic
• Zivnostenska Banka
9.2 79 Top 5
Estonia • HVB Bank Talinn 0.07 NA Top 10
Hungary • UniCredit Bank 5.3 76 Top 10
Lithuania • HVB Bank Vilnius 0.38 2 Top 10
Macedonia Representative Office only
Montenegro Representative Office only
Poland • Bank BPH
• Bank Pekao
33 1,292 No. 1
Romania • UniCredit Tiriac Bank 3.7 130 Top 5
Russia • International Moscow Bank
• Yapi Kredi Moscow
6.6 NA Top 10
Serbia • UniCredit Bank 0.86 46 Top 5
Slovak
Republic
• UniCredit Bank 3.6 93 Top 5
Slovenia • Bank Austria Creditanstalt
Ljubljana
2.2 14 Top 5
(Turkey) • Yapi Credi 29.5 653 Top 5
Ukraine • HVB Bank Ukraine
• UniCredit Bank
0.25 6 13
Source: Bank Austria Creditanstalt Corporate websitehttp://www.bankaustria.at
2.3.4 Entry motives, entry modes and market strategy
of Bank Austria Creditanstalt in CEE
The following section draws most of its information from communication with
Bank Austria group head of investor relations Mr Gerhard Smoley.
Three headline motives for BA’s entry in CEE countries became apparent
in discussions. I describe these hand in hand with what seems like underlying
or complementing characteristics of Bank Austria.
Home country push factors seem to have played a signi?cant role in
Bank Austria’s decision to expand internationally. The Austrian market
32
was severely overbanked in the early 1990s and no further signi?cant growth
in the home market was deemed feasible. In this surrounding, Bank Austria,
a recently privatized group of formerly state-owned bank, had to grow its
pro?ts fast after privatization to reach su?cient levels of shareholder value.
Therefore BA was one of the banks most committed to ?nding pro?table
business opportunities. The political development in Eastern Europe pre-
sented an unexpected window of opportunity for this.
‘By chance the possibility for feasible international expansion arose by
the fall of the iron curtain.”’(Smoley, 2007)(Smoley (2007)[109])
Host country factors in the beginning of the geographical expansion were
more restricted to indirect follow your customer considerations. The ?rst
clients in these markets were Austrian commercial clients of BA entering the
respective markets. According to Mr Smoley this was the main motive or
building stone in the beginning. ”‘If not for so many of BA’s customers from
Austria entering the CEE markets, Bank Austria would most probably have
entered these markets signi?cantly later in the process”’(Smoley (2007)[109].
Additionally the CEE markets turned out to be in large need for stable banks
(Bank Austria back then o?ered an AAA rating).
Also, Bank Austria, almost from the beginning, did not only service Aus-
trian clients in Eastern Europe, but also a variety of multinational ?rms from
other home countries, such as McDonalds and VW-Skoda. The bank might
have had an additional home country advantage in the latter business, as
Vienna functions as a management platform for a number of multinational
?rms operating in the CEE region. For example, McDonalds Eastern Euro-
pean operations center is located in Vienna, therefore close personal contact
between Bank Austria top management and McDonalds’ CEE management
was available quite easy.
A helpful ”‘snowball e?ect”’, stemming from business in the CEE re-
gion with such multinational customers, partly materialized, as these multi-
national customers often used local suppliers in the value chain. Getting
into business with these local ?rms was made signi?cantly easier by this
link[109]
30
.
The general identi?cation strategy of Bank Austria concerning attractive
host country markets, which was applied for location choice in the latter
process of transition, was based on basic macroeconomic indicators, such
as GDP, GDP growth and population. Additionally BA analyzed potential
cross-market synergies, trying to identify markets, where entry would also
30
This might be another avenue for research as one could discuss how the possibility of
follow your customer-motivated entry into foreign markets supports local market-seeking
operations later on.
33
leverage opportunities in BA’s global business network[109].
Compared to banks from other countries, Bank Austria might have had
an advantage concerning ”‘appetite for entry”’ by an indirect home market
e?ect. Bank Austria stems from a relatively small home market with a
population of 8 million. Therefore the bank was less reluctant to enter, what
banks from other home countries like Germany might have perceived as rather
too small, markets in CEE, as from the perspective of a small country bank
the market seemed to be su?ciently large[109]. Additionally the bank had
experience how to pro?tably operate in a small market. These two points
might also explain the fact, that Austrian banks in general have been heavily
involved in the CEE region already early in the transition process.
Another reason why banks from the small host country Austria were
?rst-movers in the CEE region, outpacing German banks who, according
to other theoretical considerations, such as the availability of large deposit
volumes and the e?ect of strong competition in large markets on e?ciency
of market participants, should have bene?ted from their larger home country
market while also being close geographically, might have been the fact, that
German banks, in the early period of transition, very much concentrated
on growing their business in the former GDR. This focus possibly restricted
organizational capabilities left to explore the CEE markets.
The early geographic pattern of Bank Austria’s expansion into CEE mar-
kets also was closely linked to the follow your customer-nature of its early
FDI projects. Bank Austria ?rst expanded into markets close to Vienna,
namely Prague in the Czech Republic, Bratislava in the Slovak Republic and
Budapest in Hungary. Whereas in the economic literature (e.g. Buch and
de Long (2004)[28]) the in?uence of distance on bank FDI is predominantly
perceived to base on an intuition of information costs, Mr Smoley stated,
that, in the case of Bank Austria, distance was simply negatively correlated
with the amount of activity of potential real sector customers to be followed
abroad[109].
Additionally though, the transition economies closest to the EU-15 also
pro?ted most from the fall of the iron curtain, experiencing faster economic
growth than more distant countries. According to the practitioner, and prob-
ably due to considerations of eastern enlargement of the EU, a more rapid
increase in institutional quality also was evident in these countries. There-
fore they were also the most interesting markets from a pure foreign market-
seeking perspective in the region.
Digging deeper into the expansion strategy of Bank Austria, to under-
stand the entry modes preferred by the bank one has to take the general
34
strategic setup of Bank Austria in the 1990s as well as the changing land-
scape in the ?nancial sector in CEE into consideration.
Both Bank Austria and Creditanstalt started to enter the CEE markets
in an extensive way, ?rst setting up small agencies in the countries’ capi-
tals. Bank Austria’s general strategy was to enter the market via Green?eld
investment. Usually its ?rst step was to enter via its investment banking
division, which acted as a path?nder to screen respective market character-
istics and the legal environment. As laid out before, primary early clients
for BA in the region were Austrian corporate customers and other multina-
tional ?rms. In general Bank Austria, in Austria as well as in CEE, heavily
focused on the commercial wholesale segment, at least in the 1990s. Due to
this strategic positioning, a large branch network to attract retail and SME
customers and soft information about these informationally opaque poten-
tial clients were not mandatory for successful operations in these markets.
Indeed, for conducting wholesale business, it is often su?cient to establish a
handful of regional headquarters in the big cities[109]. This latter argument
might make clear why Bank Austria, compared to other multinational banks
in CEE, was less active in the takeover market for local banks.
However, as margin pressure in commercial banking increased in Eastern
Europe, due to the development of capital markets and an increasing number
of foreign, e?cient banks operating in the region in this segment, Bank Aus-
tria added retail banking as an additional focus for growth in the region[109].
This rededication to retail banking in the region was further motivated by
the rapidly growing demand for services such as asset management and tra-
ditional retail banking arising through the economic development of these
countries.
One of the most attractive, due to being the largest, markets for retail
banking in Central and Eastern Europe is Poland. Bank Austria took a share
in the Polish bank PBK in 1997, becoming majority shareholder by 2000
31
.
However, Bank Austria was not able to grow its business via acquisi-
tions at this time in many CEE countries strongly. The reasoning shows how
strategic choice in one market is in?uenced by an institutions’ business devel-
opment in other markets. Bank Austria was one of the very few international
banks heavily involved in Russia when the Ruble crisis materialized in 1998.
The incurred losses of Bank Austria in its Russian business restricted the
possibility for growing its business in other parts of Eastern Europe for the
following years, due to a lack of ?nancial power. BA was not able to partic-
31
Polish regulation at that time was such that only 10% stakes in a domestic bank could
be bought per round. Therefore reaching a majority share in a Polish bank could not have
been attained immediately.
35
ipate full scale in the bank privatization rounds in Eastern Europe in 1999
and 2000[109]. Its main Austrian competitor in Eastern Europe size-wise,
Erste Bank AG, took advantage of this strategic weakness very actively tak-
ing part in takeover/privatization markets in Eastern Europe. Erste Bank
AG also was less reluctant to pay high strategic prices for banks especially
o?ering su?cient retail and SME client contacts, as these business segments
also happen to be the strategic focus of Erste Bank in its Austrian home
market.
So in general, Bank Austria was relatively weakly involved in signi?cant
acquisitions in Eastern Europe. BA tried to grow its retail business via
acquisitions in the Czech Republic, Slovak Republic and Hungary, but failed
to do so, due to a regional takeover market ever more characterized by bidding
wars on attractive targets. From an ex post point of view the bank’s forced
low acquisition activity in the early years however seems to have been a
blessing in disguise, as Bank Austria was one of the foreign banks in the
region not negatively a?ected in their growth possibilities by having to take
on a large share of non-performing loans from an acquired local bank[109]
32
.
An additional explanation for the bank’s relatively low M&A activity,
according to Mr Smoley[109], was that Bank Austria had no punctual geo-
graphic expansion strategy, therefore being less dependent on speci?c acqui-
sition, rather focusing on entering the Eastern European market on a broad
regional basis. With this implemented strategy Bank Austria has also been
able to gain signi?cant economies of scale on the regional basis via synergies
both on the pro?t as on the cost side. On the cost side, Bank Austria Cred-
itanstalt was able to build up centralized transaction centers for the whole
region and also introduced a, though partially adapted to local needs, com-
mon IT infrastructure. Additional economies of scale and scope according to
the practitioner are available on the revenue side[109]. As the CEE region
itself strongly integrates trade-wise, with intra-region trade becoming ever
more important, as well as an ongoing legal integration due to the adoption
of common EU law, two potentially pro?t-enhancing opportunities arise for
a bank with a complete network across the whole region. For one, the trade
32
Indeed with this ”‘Green?eld”’ strategy Bank Austria itself was surprised with the
low ex post risks in their loan portfolio. Ex post risks for some time now have been lower
in BA Eastern European operations than in Austrian operations, which was a completely
unexpected development for the bank. One advantage according to the practitioner is
social culture in the CEE states, in the respect that e.g. in Poland not paying back a
loan is still seen as a personal shame, leading to a perceived lack of moral hazard in loan
provision.
36
integration increases the need of local ?rms for banking services in multiple
countries in the region, favouring a bank with a complete network. Concern-
ing the legal integration, the possibility for selling homogeneous products in
the whole region increases, yielding cost-saving and reputation-spillover pos-
sibilities for Bank Austria. These available economies suggest, that banks,
analyzing single country markets stand-alone (and deeming them too small),
might miss an important point. As Mr Smoley stated
‘Bank Austria is the No.5 bank in Slovenia, a country with a population
of 2 million. Therefore looking at the market stand-alone one could ques-
tion whether signi?cant value-added can be achieved by entering this market.
However being present there creates network e?ects and value-added for com-
mercial customers in Poland, Czech Republic, etc.. One has to take this into
account when thinking about entering a market.”’(Smoley, 2007[109])
Bank Austria also indirectly grew by acquisition in Eastern Europe, tak-
ing over Creditanstalt in 1997, who itself already had a signi?cant physical
presence in numerous market in this region, such as for example the Slovenian
”‘Nova banka”’ Creditanstalt had acquired back in 1992. One of the success
factors in this integration was that brands stayed independent for ?ve years
to keep goodwill immanent in both brands in these countries, while at the
same time exploring economies of scale centralizing back o?ce operations for
both brands.
Additionally Bank Austria’s CEE operations grew via two other M&A
deals Bank Austria was part of. As discussed before Bank Austria today
bundles all former CEE operations of HypoVereinsbank and UniCredit
33
.
For Bank Austria Creditanstalt the merger with HypoVereinsbank brought
a number of signi?cant advantages for CEE operations with it. BA was able
to signi?cantly strengthen its network in Eastern Europe. The HVB opera-
tions taken over were further successfully leveraged, as Bank Austria already
was a more established brand name in Eastern Europe. BA was also able to
manage existing HVB operations more e?ciently due to its greater experi-
ence and vaster market knowledge in CEE markets
34
. One clear-cut synergy
was that HVB had already also taken over a Polish domestic bank. As the
respective Polish subsidiaries were merged the newly created subsidiary now
33
Except one polish subsidiary bank which became part of the UniCredit organization
due to political restrictions by the Polish government.
34
The former claim can be backed by the fact that for a while HVB changed the name
of subsidiaries of the group in the region to HVB (country). The brand name was not
as well known as Bank Austria and stunted business growth for some time in the CEE
region.
37
became the third-largest bank in the Polish market and even more impor-
tant, the lone universal bank with a strong retail business in Bank Austria’s
Eastern European portfolio[109]. As stated before, this was of signi?cant
interest especially in Poland. The increased size of the group’s Eastern Eu-
ropean operations also allowed the bank to take on ?xed costs risks of further
expansion into the retail sector.
As Bank Austria tried to grow its retail business, it grew its branch net-
work in Poland, Romania, Croatia and Hungary via Green?eld Investment.
As Hungarian target banks became too expensive for takeover, BA tried to
grow organically on a large scale, establishing 100 new Green?eld branches
in short time.
Additionally, in recent years, Bank Austria became more active in the
takeover market, acquiring banks ”‘Splitska banka”’ in Croatia and ”‘Biochim”’
in Bulgaria in 2002 as well as ”‘Central Pro?t Banka”’ in Bosnia in 2003.
In the retail sector ”‘Bank Austria started out with Green?eld Investment
and then tried to speed up the growth process via acquisitions, even though
acquisition prices grew higher, too.”’(Smoley, 2007[109]. Concerning target
strategy, Bank Austria, contrary to other banks in the region, focused on
acquiring healthy local banks and paying the high prices for these, whereas
other entrants sought a low acquisition price ?rst and foremost, banking on
being able to restructure badly-performing banks burdened with a high share
of non-performing loans[109].
Based on this discussion, one may be able to derive that in general full-
scale retail banking services can probably only be provided in a market en-
tered via M&A. In almost all countries BA was not involved in takeovers it
still focuses on the commercial clients segment[109].
The underlying reasons can be identi?ed by evaluating the banking mar-
ket in the Czech Republic. Bank Austria there has limited scope winning
retail clients away from other banks due to a lack of network size. Whereas,
for example, Erste Bank AG subsidiary Ceska sporitelna has 630 branches
in this country, Bank Austria only operates 40. So due to missing local
structures, there is no real possibility to attract retail customers as well as
SME customers in some regions of the Czech Republic. Still, Bank Austria
in terms of value of assets is the fourth largest bank in CZ concerning assets,
due to its leading position in the ?eld of industry and trade (commercial)
clients[109].
Strong existing bank-client relationships and therefore a general lack of
customer mobility in the retail segment are reasons why Bank Austria is re-
luctant to build up large Green?eld branch structures in most CEE countries,
even more so, as BA tries to position itself as a quality-, not price-leader,
where the former strength can hardly be marketed to potential customers
38
locked in an existing relationship with another bank.
Besides engaging in retail and commercial banking, Bank Austria has a
very well-positioned investment banking division in the CEE region. Due
to having been one of the ?rst-movers in the markets, BA has excellent
long-lasting relationships especially with government agencies in Central and
Eastern Europe, also evidenced by the bank being awarded a price for ”‘Best
Investment Bank in CEE”’ by ”‘Financing New Europe”’. In this ?eld BA
also still pro?ts from the fact, that most large investment banks early on
deemed this regional market to be too small to put a strategic focus on it.
According to Mr Smoley these relationships to local customers that are in
place now are more important for successful investment banking than strong
relationships to the capital market where products are placed
35
. Now being
part of UniCredit group also strengthens BA’s position in the latter respect,
such that BA expects to keep its market leadership in CEE in this area of
business[109].
One ?nal interesting point concerning Bank Austria Creditanstalt’s strat-
egy is the mode of re?nancing of operations in Central and Eastern Europe.
Start-up investments and acquisition prices were ?nanced by the Austrian
parent bank. However concerning the re?nancing of operating business such
as loan provision, the pecking order is local ?rst, global only if needed. Sub-
sidiaries should ?rst and foremost re?nance their operations via raising local
deposits. BA tries to re?nance all of its local loans in the respective local
market. However the degree of local re?nancing di?ers among countries (be-
tween 100% and 80%). The larger the subsidiary concerning the number of
branches, the higher tendentially the percentage of local re?nancing
36
. Early
on in the transition process. as loan demand exceeded wealth, the picture
percentage-wise looked di?erent, with the bulk of loans by CEE subsidiaries
re?nanced by the parent in Austria. One important part of local deposits for
BA, as the group’s main strength is still with commercial clients, are large
deposits from its commercial clients, who hold liquid assets in local currency
in a substantial amount[109].
If a subsidiary comes short of complete self-re?nancing, Bank Austria
operates an internal capital market structure trying to optimize capital across
the group. Usually the subsidiaries should not re?nance themselves via the
capital market themselves, as Bank Austria has better re?nancing conditions
35
However placement power has to be su?ciently large to keep the trust of customers
in the bank being actually able to secure ?nancing.
36
This again points at the need for a large branch network to attract a large volume of
retail customers.
39
there due to a better rating[109].
The following reasoning can be put forward for such an implemented
pecking order of re?nancing. First, local deposits should be the main and
?rst source of re?nancing, due to a complete lack of currency risks
37
when
re?nancing that way, as well as an increasing distaste of stock analysts for
intra-group cross-subsidiaziation[109].
If local deposits are however not available to a su?cient degree, the in-
terest rate advantage of Bank Austria Creditanstalt over its subsidiaries in
capital markets should be made use of.
2.3.5 The future strategy of Bank Austria Creditanstalt
in CEE
With Bank Austria Creditanstalt now a part of UniCredit group, the geo-
graphic focus of BA’s business is shifting further to the east. Bank Austria
has recently entered markets in Russia and Turkey and is in the process of
doing so with a large commitment in Kazahstan and the Ukraine. In Rus-
sia, Bank Austria group is present via the International Moscow Bank. In
Turkey, BA holds a 50% share in the Yapi Credi Bank. In Kazahstan and
the Ukraine Bank Austria is in the process of taking over ATF Bank and
Ukrsotsbank respectively[40].
This further eastern expansion is driven by market dynamics in the CEE
regions already serviced as well as by a perceived comparative advantage of
Bank Austria over other competitors in the new markets.
Large CEE markets have already become quite consolidated and further
signi?cant growth is hard to achieve there. One prime example is the Czech
market. The ?rst three market positions are ?rmly established there, espe-
cially in the retail banking sector characterized by a low customer switching
rate. The market is more or less divided in stable market shares, and BA is
restricted to keep operating within its niche strategy there. While organic
growth can hardly be achieved, pro?table growth via acquisitions is also not
feasible as attractively priced and available targets are missing[109].
In contrast, vast pro?table growth potential in Russia and Turkey exists,
especially due to the fact that a lot of foreign banks are still not present in
37
Note that additional to the direct costs of currency risks, Basel 2 guidelines force
banks to either completely costly hedge these risks or build up legal reserves for them,
which is also very costly as the respective amount of capital can then not be used in more
pro?table opportunities.
40
these markets due to perceived political risk. There is high pro?t potential
in the retail segment as well, as especially in Turkey demographic trends lead
to an arising interesting market for ?nancial products[109].
According to Bank Austria, two comparative advantages of the bank over
competitors exist concerning these markets. For one, Bank Austria is one of
the very few banks that have build up vast relevant experience from being
present in the very early years in transition economies. With target countries
for further expansion being similar to CEE countries in the early years of
transition (Smoley, 2007[109]), Bank Austria might have best practices to
deal with such surroundings including political risk, a lack of institutional
quality and a just developing modern real sector.
Additionally the integration into the large UniCredit group enables Bank
Austria to expose itself to such higher risks. Bank Austria Creditanstalt, with
strong ?nancial backing of the complete group, is able to pursue a long-run
oriented strategy. Turkey and Russia might even be candidates for ?nancial
crisis in the next year in the eyes of BA, but the is be able to sustain 2-3 loss-
making years in the market, being able to focus on the long-term potential,
especially for Turkey.
2.3.6 Same region, di?erent strategy: The case of Er-
ste Bank AG in Central and Eastern Europe
Erste Bank AG also started out as a purely Austrian bank, rooting from
the mutual savings bank structure in Austria. Like Bank Austria the pub-
licly stated geographic focus of further business development is Central and
Eastern Europe[1].
In contrast to Bank Austria Erste Bank AG started to enter the CEE
markets relatively late, starting with the acquisition of Mezbank in Hungary
in 1998. This ?rst step was followed by further acquisitions of Cakovecka
banka, Bjelovarska banka and Trgovacka banka in Croatia in 1999
38
, Ceska
sporitelna in the Czech Republic in 2000
39
, Rijecka banka in Croatia in 2002,
Postabank in Hungary in 2003 and Novosadska banka in Serbia as well as
Banca Comericala Romana S.A.
40
in Romania in 2005. In July 2007 Erste
Bank, like Bank Austria Creditanstalt, expanded further eastwards, acquir-
ing Bank Prestige in the Ukraine[1].
38
These three banks were later merged into Erste&Steiermaerkischen banka.
39
To be precise Erste Bank bought a 52% majority share in this bank in 2000 gradually
increasing its share to 100% by 2005.
40
BCR is the largest bank in Romania with a market share of 32%.
41
Erste Bank AG operations in the respective countries today are subsumed
in Table 2.2.
Table 2.2: Erste Bank AG Operations in Central and Eastern Europe
Country Subsidiary Clients
(million)
Branches Market
position
(number
of clients)
Croatia • Erste Bank Croatia 0.6 114 No. 3
Czech
Republic
• Ceska Sporitelna 5.3 637 No. 1
Hungary • Erste Bank Hungary 0.9 186 No. 2
Romania • Banca Comerciala Romana 3.5 485 No. 1
Serbia • Erste Bank Serbia 0.3 NA No. 9
Slovak
Republic
• Slovenska Sporitelna 2.5 279 No. 1
Ukraine • Erste Bank Ukraine Just founded in December 2005
Source: Erste Bank AG Corporate websitehttp://www.erstebank.com
The main di?erence concerning general strategy between Bank Austria
Creditanstalt and Erste Bank AG is, that the former predominantly focuses
on the wholesale banking segment, including loans to larger commercials
as well as investment banking services, whereas Erste Bank AG’s primary
focus is the retail segment and, to a lesser degree, the provision of ?nancial
services to small and medium-sized (SME) ?rms[1]. This general di?erence
also seems to at least partly explain both di?erent timing and mode of entry
between these banks. Whereas the strategy of Erste Bank AG requires a large
branch network as well as soft information about informationally-opaque
potential loan clients, Bank Austria’s need for these due to its focus on
the wholesale business, was less pronounced (at least until 2000 when BA
started to increase its interest in the retail segment in the CEE region).
Therefore Erste Bank was much more active acquiring these needed assets or
capabilities via the acquisition of local banks. Erste Bank did not enter any
market on a large scale via Green?eld Investment but predominantly choose
an acquisition strategy from the beginning[1].
Also, the fact that Erste Bank AG entered the region way later than Bank
Austria Creditanstalt could also be seen in light of di?erent core business
segments. Whereas historically BA had a large portfolio of large Austrian
?rms equipped to enter the Eastern European markets, and therefore scope
42
for applying follow your customer-strategy, Erste Bank AG commercial client
portfolio was more skewed towards SMEs, who were less ready to engage in
the CEE region from the beginning. Additionally the retail segment in these
markets itself might have been to risky and small to enter in the early phase
of transition.
Erste Bank AG also focuses on less markets in the region, being active
in only 7 countries[1]. It has a very strong position in the retail segment
especially in the Czech and Slovak Republic and Romania, as well as to
a lesser degree in Hungary. Interestingly Erste Bank AG is not active in
Poland, deemed to be the most interesting market for retail banking in CEE
due to its large population.
Concerning future growth strategies Erste Bank AG tries to make use of
the banking sector development cycle in the heterogeneous markets in the
region. Whereas the Ukraine, Serbia and Romania are now seen as ”‘emerg-
ing markets”’ characterized by a low market penetration rate of banking
services, where growth short-term is driven by demand for simple banking
services such as savings, payment transfers and debit cards, these countries
will become what Erste Bank AG calls ”‘developing markets”’ like the Czech
and Slovak Republic, Hungary and Croatia, where bank penetration rates
are somewhat higher therefore experiencing tougher competition, but at the
same time enabling growth in high-margin products such as mortgage loans,
consumer loans, credit cards and wealth management products[1].
2.4 Conclusion
Quite a lot can be learned for multinational banking in general from the
above case study.
Concerning home country e?ects in multinational banking, it is con-
cluded, that banks from saturated markets are more likely to expand abroad,
due to missing signi?cant growth opportunities in the home market. This
result has already been discussed in the empirical literature (e.g. de Haas,
van Lelyveld (2006)[44]). Interestingly home country characteristics in the
majority of studies are most often seen as shaping ?rms’/banks’ capabilities
to become multinational (e.g. Buch and de Long (2004)[28], Focarelli and
Pozzolo (2000)[60], not so much as a push factor in incentives to become
multinational.
New to this discussion are soft home market e?ects mentioned by Bank
Austria Creditanstalt, namely that the relative size of the home country
43
has an e?ect on whether a potential target country is perceived by a multi-
national bank as su?ciently pro?table (large). Whereas this result seems
somewhat irrational, an additional home market size e?ect concerning ge-
ographical characteristics of expansion might be, that banks from a small
market might simply be better equipped to be pro?table in foreign small
markets, due to experience in this type of surroundings.
This knowledge in similar markets might also explain why some banks are
able to enter risky accending markets at an earlier stage than others. Bank
Austria seems to be more equipped to enter markets such as the former CIS
states due to experience in similar markets, namely the Central and Eastern
European countries at an early stage of transition.
One very clear result of the case study is the importance of follow your
customer strategies in multinational banking. Especially having customers
to follow seems to allow early entry in accending markets, which stand-alone
might not be attractive enough at this stage of development.
Additionally the geographic pattern of bank internationalization, at least
in non-OECD countries might to a large degree be explained by FYC strate-
gies. This result adds an additional explanation on the in?uence of distance
on multinational bank location choice, as distance might play a minor role
in this decision directly through information costs considerations (e.g. Buch
and de Long (2004)[28]) but a larger role indirectly as distance is an im-
portant variable in the location decision of real sector ?rms to be followed
who actually face a decision concerning the trade of physical goods over this
distance
41
.
Another mentioned bene?t of follow your customer strategies is that local
?rms might be easier to attract for a bank if the former has business ties to
multinational customers of the respective bank.
The most interesting point of the case study might be the ?ndings on how
a bank’s product strategy shapes entry modes and entry timing into foreign
markets, as well as the geographic pattern of such entry.
One conclusion derived is that banks focusing on the wholesale business,
characterized by informationally non-opaque clients with low switching costs
41
Along the FDI theories of e.g. Markusen and Venables (2000)[98] we would expect
market-seeking horizontal foreign direct investment between distant countries and vertical
production-cost minimizing foreign direct investment between neighbouring countries. For
less-developed countries we would rather expect in?owing vertical FDI. So in the early
CEE case we would expect countries close to Austria to be main recipients of vertical
FDI, therefore close countries to be main recipients of Austrian bank FDI.
44
and low requirements concerning the scope of the branch network, are more
likely to enter via Green?eld investment, whereas banks focusing on the retail
and SME
42
business are more likely to enter a foreign market via the acquisi-
tion of a local bank due to the client group’s informational opacity (especially
SMEs), high switching costs from existing bank relationships (especially re-
tail customers) and large requirements concerning the branch network scope
(especially retail customers). These assets might most easily be acquired via
the acquisition of an existing bank, whereas wholesale-focused entrants might
probably not be willing to pay for these assets, as they are of second-order
importance for their business focus.
Additionally, timing of entry in a respective market
43
might be a function
of product strategy of a bank. As discussed above, a retail-oriented bank not
having large real sector clients in its portfolio ex ante, might want to wait
to enter an accending market until the retail market in this country has
grown su?ciently large. In contrast, a wholesale-oriented bank might be
able to enter a market earlier on, banking on early pro?t from follow your
customer-business, business with third-country multinational ?rms as well
as investment banking services for the host country government and large
corporations.
For the economic literature these results propose a research strategy both
treating banks as conglomerates as well as on a market level taking care of
the fact that banking is not a single homogeneous industry. Therefore multi-
national bank behaviour might require di?erentiated theories for respective
banking segments. Entry modes, further strategic decisions and potentially
associated di?erent impacts of bank entry into foreign host markets might
have to be discussed separately for banks with a retail and SME ?nancial
services focus and wholesale/investment banking-oriented institutions.
The empirical literature, when discussing the location choice of multina-
tional bank subsidiaries, primarily focuses on host country e?ects. Taking
a result of the above case study into consideration, this picture might be
partially misleading. Indeed location strategies might take complete regions
into account, not single countries stand-alone. An isolated view on a re-
spective country market might be insu?cient, as due to potential network
e?ects within a region the market’s attractiveness also depends on potential
cost or revenue synergies with other markets in the region. The notion of
network economies of scope across regions also seems rather new to economic
42
Small and Medium-sized enterprises
43
This could also be seen as locational choice, as a market A in period t
i
might mimic
a market B in period t
j
.
45
theory concerning banking, but has been a standard feature of business lit-
erature discussing international one-stop-shopping possibilities for clients as
well as follow your customer strategies, which constitute a special case of
such network economies of scope.
The ?nal point I want to stress is the result obtained on the re?nancing
of multinational bank activity in host countries. Indeed it seems that in
su?ciently developed markets foreign bank activity does not lead to a large
volume of capital in?ows into this market, as the subsidiaries’ re?nancing
might primarily take place in the host country (via deposits). Therefore
internal capital markets in multinational bank organizations might only play
a residual role in the ?nancing of these subsidiaries.
46
Chapter 3
M&A versus Green?eld -
Optimal Entry Modes into
Markets with Sequential Entry
3.1 Introduction
One of the topics concerning foreign direct investment, that has just recently
become a core focus of economic literature, is the choice of the exact entry
mode of ?rms into foreign markets.
Foreign direct investment can take place via the acquisition of a local ?rm
in the target market (M&A), via the set-up of a completely new structure
(Green?eld Investment) or some impure organizational designs in between
these two modes (e.g. Joint Ventures, Brown?eld Investment).
The most general advantage for a ?rm entering via M&A instead of Green-
?eld lies in the di?ering e?ect on the ex post host market structure and
therefore on the degree of competition in the market ex post. A ?rm enter-
ing a market with d incumbents via Green?eld Investment will ex post face
d competitors in the market, whereas, when entering via M&A, it will only
face d-1 competitors (abstracting from the possibility of additional entry).
In markets with imperfect competition, e.g. due to horizontal product dif-
ferentiation, all else equal, pro?ts for the entrant should therefore be higher
if the ?rm enters via M&A. This positive characteristic of entry via M&A is
for example modelled by M¨ uller (2001)[105] and G¨org (2000)[70].
47
However, this bene?cial ”competition” e?ect might be a short-sighted
motivation for entry via M&A. If one would consider foreign ?rms to enter
markets sequentially, such that some ?rms due to various reasons
1
move into a
speci?c host market faster than others, the entry mode of early movers might
a?ect the entry decisions of sequential entrants in the future. Obviously, if
the early mover entered via Green?eld Investment, a sequential entrant would
face the decision whether and how to enter a market with d+1 incumbents,
whereas with early mover entry via M&A he would face a decision on a market
with d incumbents. Therefore sequential entry might be accommodated by
the early moving ?rm entering via M&A instead of Green?eld Investment.
So the static positive ”competition” e?ect of entry via M&A might to some
degree be o?set up by an increase in the likelihood of further sequential entry.
Two types of questions arise in such a ”dynamic” setting, the ?rst one
being ”ex post”, how the entry mode of the early mover a?ects potential
sequential entrants’ decisions, and therefore ”ex ante” how the optimal entry
mode choice for the early mover looks like given its e?ect on sequential entry.
To analyze entry mode decisions in a less static setting a simple two-period
model, featuring a potential early and sequential entrant, is proposed.
To focus on pure market structure e?ects of entry modes I, at least in
the basic model, abstract from other incentives for entry via M&A such
as asset complementarity between acquirer and target (see e.g. Nocke and
Yeaple (2007)[107]), but from the possibility, that an entrant might be bet-
ter equipped to operate a company in general compared to its host country
target. In this simple setup ?rms are assumed to be symmetric in their mar-
ginal costs, only di?ering in what can be called their ?xed costs of operating
in a market. In such a model I show, that entry via Green?eld Investment
indeed has a valuable strategic entry deterring e?ect for early entrants, such
that sequential entry in general is less likely than if the early entrant enters
via M&A. It is also shown, that the chosen entry mode of later entrants dif-
fers in expectations for di?ering early mover entry modes, as do acquisition
prices for domestic targets. Which early movers should enter via Green?eld
Investment or M&A, and how modelling potential sequential entry reduces
the relative probability of early movers choosing M&A over Green?eld In-
vestment as entry mode, is then derived.
To sum up, the model will have the following intuition. The general trade-
o? faced by ?rms, when choosing their entry mode, is between increasing net
variable pro?ts (variable pro?t minus acquisition price) and not having to
1
For example practitioners mentioned that German banks did not enter the market
in Eastern Europe early on because they focused on building up their business in East
Germany at this time.
48
sustain the ?xed costs of Green?eld entry when entering via M&A, versus
working in the market at lower ?xed operating costs per period when entering
via Green?eld Investment. The heterogeneity in possible entrants’ general
?xed operating cost levels determines their respective ?xed cost disadvantage
when entering via M&A.
Modelling potential sequential entry reduces the expected variable pro?t
advantage of entry via M&A, as this kind of entry is shown to increase the
probability of pro?t-reducing sequential Green?eld entry. Additionally, the
pro?t reducing e?ect for the early entrant of sequential entry via Green?eld
Investment is more pronounced if the early entrant entered via M&A, than if
it had entered via Green?eld Investment. In sum these two e?ects are shown
to reduce the attractiveness of M&A entry for all types (concerning ?xed
operating costs) of early entrants, even though the threat of sequential entry
also works to reduce the acquisition price to be paid for takeover targets. As
the ?xed cost disadvantage of M&A is una?ected by such an entry threat,
a smaller relative fraction of potential early entrants (namely only the very
?xed-cost e?cient ?rms) chooses entry via M&A compared to a static setting
not taking the threat of sequential entry into account.
The main result of the basic model is, that taking into account the threat
of sequential entry reduces the incentive for early movers to enter via M&A
compared to a static world. This result holds true for all markets where both
kinds of entry modes are principally probable
2
. These market characteris-
tics should be present in the majority of real-world markets, as we generally
observe both kinds of foreign direct investment modes in the respective coun-
tries.
The possibility of market incumbents to deter or accommodate further
entry is well-established in economic literature, starting with the work of
Stackelberg (1934)[122] and continuing with contributions by e.g.Bernheim
(1984)[20] and Gilbert and Vives (1986)[67]. However, according to my
knowledge, no one has linked entry deterrence in models with sequential
entry to the entry mode choice of multinational corporations. Also, the ma-
jority of theories have been restricted discussing entry deterrence in markets
with Cournot competition.
The proposed basic model can therefore contribute to the understand-
ing why Green?eld Investment still makes up a substantial share of foreign
2
This means, that structures are such that ?rms entry choice will be heterogeneous in
equilibrium, with some ?rms entering via Green?eld Investment, some via the takeover of
an incumbent ?rm and some will not enter the market at all.
49
direct investment. For example, Ra?, Ryan and Staehler (2006)[113] ?nd,
that for Japanese ?rms Green?eld Investment is the dominant mode of entry,
with cases of Green?eld entry outnumbering M&A entry by a factor of 2.5 to
1. Even concerning value of investments, Lorentowicz, Marin and Raubold
(2002)[95] show evidence that for German Direct Investment in Eastern Eu-
rope Green?eld Investment makes up 56% of total FDI ?ows. Concerning
number of projects, this dominance of Green?eld Investment in their study
is even more striking when one considers that the mean size of M&A invest-
ments is roughly 2.4 times the mean size of Green?eld Investments in the
authors’ dataset.
Recent theoretical contributions on optimal entry modes have however
stressed additional reasons of ?rms to enter via M&A, especially the impor-
tance of asset complementarity (e.g. Nocke and Yeaple (2007)[107]). Con-
cerning the found dominance of Green?eld Investment in the number of oc-
currences, we think a counterbalancing theoretical contribution stressing the
advantage of Green?eld Investment is needed.
Also, when controlling for ?rm size e?ects, Ra?, Ryan and Staehler
(2006)[113] ?nd, that the widespread perception that the most e?cient ?rms
enter via Green?eld Investment, does not hold true. Indeed, when controlling
for ?rm size, they ?nd, that tendentially the more e?cient ?rms enter via
M&A, though the e?ect is not signi?cant. Andersson and Svensson (1994)[4]
?nd, that ?rms with strong organizational skill tend to enter foreign markets
via the acquisition of local incumbents rather than via Green?eld Invest-
ment. Theoretically the latter point is theoretically rationalized in the paper
by Nocke and Yeaple (2007)[107], as an internationally mobile asset such as
organizational skill can be leveraged by combining it with non-mobile capa-
bilities of acquired local market incumbents. I incorporate these ?ndings,
that have else been neglected in the literature, to date, with the general per-
ception being that the most e?cient ?rms enter via Green?eld Investment
(e.g. Mueller (2001)[105]), in my model setup.
Additional to the above main result it is shown, that the advantageous
e?ect of early entry via Green?eld Investment is more pronounced when
takeover possibilities in a market are limited, such that early entry via M&A
deletes the option of M&A entry for sequential entrants. Interestingly de-
creasing potential competitors options then has a negative e?ect for early
movers. I therefore come up with a new, strategic indirect, ”‘perverse”’ ef-
fect of missing takeover targets on the choice of entry mode of ?rms into
foreign markets.
Welfare analysis within the proposed basic model yields, that in a world
50
with sequential entry, the negative e?ect welfare e?ect of entry via M&A is
even more pronounced than in a static setting. This is due to the fact, that
additional sequential entry, which is accommodated by early entry via M&A,
in the proposed setup reduces pro?ts of foreign ?rms more than it increases
consumer rent.
A ?nal proposed extension of the basic model is to incorporate what can
be called country- or market-speci?c learning by doing-e?ects, which changes
the results from the basic model for some types of markets. Indeed it is
shown, that when the degree of product di?erentiation in the market is low
and learning-by-doing e?ects are su?ciently strong, an ”e?ciency e?ect”
dominates the ”competition e?ect” of entry modes, such that in this case
early entry via M&A deters sequential entry. In this setting then taking into
account the threat of sequential entry increases the incentive for early movers
to enter via M&A.
One could argue that this latter extension might yield an especially valu-
able insight into the retail and commercial banking sector, where ex ante
product di?erentiation between banks is generally perceived to be low and
learning-by-doing e?ects perceived to be both strong as well as predominantly
country-speci?c, due to the heterogeneity of banking regulation across coun-
tries. Indeed the result, that in this industry M&A entry should be very
dominant due to its additional sequential entry deterring nature, is getting
support from the ?ndings in the case study in this thesis.
From a general theoretical point of view, this chapter contributes to the
existing literature by being, to my knowledge, the ?rst to leave a static world
of entry mode decision analysis to implement a more forward-looking behav-
iour of potential foreign direct investors. Additionally, within the model, I am
able to discover yet undiscussed potential e?ects of limited takeover possibil-
ities on entry mode choice, as well as the e?ect of country- or market-speci?c
learning by doing e?ects on the choice of entry modes, in at least a stylized
way.
The rest of the chapter proceeds as follows. Section 2 lays out the basic
model and assumptions, analyzes the contingent sequential entry structure
and the derives the optimal entry mode decision of early movers, ?nally
comparing the results to a benchmark world without sequential entry. Section
3 discusses welfare implications Two extensions/setup changes are introduced
in the following. In Section 4 the e?ect of limited takeover possibilities is
analyzed. Then the model is extended by including country-speci?c learning-
by-doing e?ects in Section 5. The ?nal section concludes and discusses open
questions.
51
3.2 The basic model
The building stone of the model is a host country market characterized by
Bertrand competition in horizontally di?erentiated goods. The market is set
up as a Salop circle of size Y = 1 and transport costs or degree of product
di?erentiation t > 0.
Before any foreign entry into the market, there are two incumbent do-
mestic ?rms A and B operating in the market at marginal costs c
A
and c
B
and ?xed costs of operation per period of O
A
and O
B
, respectively.
The simple timing structure of the model is illustrated by the following
graphic.
Figure 3.1: Time Structure of the Model
Foreign
firm 1 entry
(mode)
decision
T=1: First
period of
market
competition
Foreign
firm 2 entry
(mode)
decision
T=2: Second
period of market
competition
T
Foreign entry happens sequentially, such that an early-moving ?rm enters
the market one period before the next potential entrant.
At the beginning of period T = 1 there is one potential foreign entrant
F
1
with marginal costs c
1
and ?xed costs of operation per period of O
1
.
The potential entrant ?rm has three options concerning entry into the
host country market. It can either enter the market via the acquisition of
an incumbent ?rm (M&A), by establishing a completely new ?rm structure
in the host country (Green?eld Investment) or not enter the market at all.
After the entry decision the ?rst market game in the host country market
takes place.
At the beginning of period T = 2 a second potential entrant F
2
with
marginal costs c
2
and ?xed costs of operations per period of O
2
, with O
2
52
assumed to be uniformly distributed between 0 and 1
3
, decides on whether
and how to enter the market. After the sequential entry decision the second
market game takes place.
For simplicity symmetric marginal costs of ?rms are assumed, such that
c
1
= c
2
= c
A
= c
B
= c. Therefore it is straightforward that ?rms will locate
equidistantly to each other on the Salop Circle
4
.
Additionally, incumbent ?rms are assumed to be perfectly symmetric,
such that O
A
= O
B
= O
D
> 0, where O
D
is then the level of ?xed costs of
operation per period for all domestic ?rms.
Concerning entry modes the setup is as follows.
If a ?rm enters via Green?eld Investment it bears ?xed costs of entry of
F > 0. The newly setup structure will then work at marginal costs c and
?xed costs O
i
.
If it enters via M&A, the ?rm has to pay an endogenous acquisition price
A for the respective target. Bargaining power is assumed to reside with
the acquirer, such that the acquisition price will equal the outside option of
the target ?rm, which is the respective targets (expected) foregone pro?ts
when staying independent. The acquired structure will work at marginal
costs c and ?xed costs ?O
i
with ? = 2 for simplicity. The results would be
unchanged as long as we assume ? > 1.
Therefore ?xed costs of operation are assumed to be higher under M&A
then under Green?eld Investment. ? determines how relatively large this
di?erence is. The usual explanation for higher (operational) costs under
operating with an acquired organization is, that company cultures between
target and acquirer may clash (see e.g. Feely and Kompare (2003)[57])or that
there might be some costs of restructuring the target to ?t into the acquirer’s
organizational structure (see e.g. M¨ uller (2001)[105]). In the proposed setup
the ?xed costs of operation are twice as large under M&A compared to stand-
alone Green?eld operation of entrants.
Special to the proposed setup is then, that the absolute negative e?ect of
operating with an acquired organizational structure instead of a completely
new setup structure (Green?eld Investment) on operating ?xed costs depends
on the general ?xed costs e?ciency O
i
of the entering bank. The intuitive
idea is, that a generally well managed ?rm should also be better equipped to
handle post-merger integration problems than a badly managed ?rm.
3
The actual realisation of O
2
of the potential entrant is then drawn by nature.
4
Obviously this would be the pro?t maximizing location choice. As in this model
incumbents have a ”‘location history”’, it is implicitly assumed that location switching
costs are zero.
53
Note that for reasons of tractability the model abstracts from directly
arising di?erences in marginal costs between Green?eld and M&A operation.
Therefore I do not claim to fully discuss the e?ect of a ?rms general e?ciency
on its choice of entry mode. I’ll be content to discuss e?ects of non-marginal
cost e?ciency on entry mode choice.
The ?nal restriction made to reduce cases to be analyzed is what can
be called a ”no passive consolidation”-clause, which means that the focus of
the analysis will be on cases of transport costs t and incumbent ?rms’ ?xed
costs O
D
such that incumbent ?rms are not driven out of the market by
competition
5
. Such passive consolidation would occur if pro?ts of incumbents
would be negative in the case of four players in the market, so i? t < 16O
D
6
.
To sum up, general pro?t functions of ?rms then look as follows:
For ?rms entering via Green?eld Investment:
?
i,j
(GF) = (p
i,j
(GF) ?c
i
) ×x
i,j
(GF) ?O
i
?F (3.1)
For ?rms entering via Acquisition:
?
i,j
(MA) = (p
i,j
(MA) ?c
i
) ×x
i,j
(MA) ?2O
i
?A
j
(MA) (3.2)
with i ? {1, 2} denoting early and sequential entrant and
j ? {MA, GF, NE} the entry mode of the other entrant.
For incumbent ?rms, pro?ts of operating in the market are simply respec-
tive variable pro?ts depending on the entry mode(s) of entrant(s) minus the
?rms’ respective ?xed operating costs.
?
A,j
= ?B, j = (p
A,j
?c
A
) ×x
A,j
?O
D
(3.3)
In appendix 1 the respective variable pro?t for ?rms for di?erent cases
of entry are derived. In this basic model the only di?erence between entry
modes concerning variable pro?ts lies in the respective number of ?rms that
compete in the market.
5
Besides reducing cases to be analyzed let us abstract from passive consolidation among
incumbents to keep the analysis non-trivial. If e.g. any kind of entry of ?rm 1 would lead
incumbent ?rms to leave the market in period 1 then there would never be an incentive
for 1 to enter the market via M&A and no possibility for 2 to enter the market via M&A
in period , as there would be no targets to be acquired.
6
The respective pro?ts under a given number of market participants are derived in
appendix 1.
54
Before getting onto the core questions one intermediate result is derived
to help determine equilibrium acquisition prices.
Lemma 1
The structure of entry decisions is such, that ?rms will enter via M&A if
they have low ?xed costs of operation, via Green?eld Investment for medium-
level ?xed costs and will not enter the market at all with high ?xed costs of
operation.
The Lemma is proved in appendix 2 and the following graph illustrates
this result.
Figure 3.2: The General Structure of Entry Mode Decisions
O
i
?
i
?>1
1
Greenfield entry M&A entry No entry
?
GF
?
MA
?
i
Õ
i
Therefore one can conclude that ?rms making an acquisition o?er to
an incumbent ?rm will be those that would enter via Green?eld if the ac-
quisition fails to materialize. Therefore the outside option of the target is
its (expected) pro?ts under the case that the potential acquirer enters via
Green?eld Investment.
First now the ex post question will be analyzed to determine how entry
patterns di?er for the sequential entrant in period 2 depending on the early-
mover’s choice of entry mode.
55
3.2.1 Analyzing entry (mode) decisions of the sequen-
tial entrant in period T=2
In general the entry decision of ?rm 2 can be subsumed as follows
a) Firm F
2
enters via M&A i?
?
2,j
(MA) > ?
2,j
(GF) and ?
2,j
(MA) > 0
7
b) Firm F
2
enters via Green?eld i?
?
2,j
(GF) > ?
2,j
(MA) and ?
2,j
(GF) > 0
c) Firm F
2
does not enter the market i?
?
2,j
(GF) < 0 and ?
2,j
(MA) < 0
8
From this general decision structure sequential entrants’ decisions contin-
gent on early entrant entry mode decisions can be derived.
Early mover entry via Green?eld Investment
In this case ?rm 2 decides upon entry and entry mode into a market with three
incumbents, the two domestic ?rms and the early entrant ?rm 1. Therefore
if ?rm 2 enters via Green?eld four ?rms will divide market pro?ts between
them. Firm F
2
pro?t then is
?
2
(GF) = ?
V ar
(4) ?F ?O
2
=
t
16
?F ?O
2
Firm F
2
will not enter the market at all if Green?eld pro?ts are negative,
so i?
t
16
?F ?O
2
< 0 or ?xed costs of operation above
O
GF
2
=
t
16
?F (3.4)
Firm F
2
pro?t when entering via M&A in this case is
?
2
(MA) = ?
V ar
(3) ??
V ar
(4) +O
D
?2O
2
9
7
Where we do know that the former is binding due to the structure of the entry mode
decision.
8
Where the former is binding again due to the structure of the entry mode decision.
9
As the acquisition price to be paid in this case equals ?
V ar
(4) ?O
D
.
56
Firm F
2
will enter via M&A if M&A pro?ts are larger than Green?eld
pro?ts, so i?
?
V ar
(3) ??
V ar
(4) ??
V ar
(4) +F +O
D
?2O
i
=
t
9
?
t
16
?
t
16
+O
D
+F ?O
2
> 0
or ?xed costs of operation below
O
2
GF
= ?
t
72
+O
D
+F (3.5)
Firm F
2
will enter via Green?eld Investment if such entry yields both
positive pro?ts and higher pro?ts than entry via M&A, so i?
t
16
?F ?O
2
> 0
and ?
t
72
+O
D
+F ?O
2
< 0, or ?xed costs of operation O
2
in the range
O
2
GF
< O
2
< O
GF
2
(3.6)
Early mover entry via M&A
In this case ?rm F
2
decides about entry and entry mode into a market with
two incumbents, one independent domestic ?rm and the early entrant orga-
nization. Firm F
2
pro?t under Green?eld Investment then is
?
2
(GF) = ?
V ar
(3) ?F =
t
9
?F ?O
2
In this case F
2
will not enter the market i?
t
9
?F ?O
2
< 0 or ?xed costs
of operation above
O
MA
2
=
t
9
?F (3.7)
Firm F
2
pro?t under M&A then is
?
2
(MA) = ?
V ar
(2) ??
V ar
(3) +O
D
?2O
2
=
5
36
t +O
D
?2O
2
The condition for F
2
to enter via M&A is then
?
V ar
(2) ??
V ar
(3) ??
V ar
(3) +F +O
D
?2O
i
=
t
4
?
t
9
?
t
9
+O
D
+F ?O
2
> 0
or ?xed costs of operation below
O
2
MA
=
t
36
+O
D
+F (3.8)
F
2
will therefore enter via Green?eld if its ?xed operational costs are such
that
57
O
2
MA
< O
2
< O
MA
2
(3.9)
As a side result, from the above cases it is obvious, that M&A entry in this
model in general is preferred to entry via Green?eld Investment if ?xed costs
of Green?eld entry as well as the operating ?xed costs of domestic banks are
large, as the latter leads to a reduction in the acquisition price to be paid.
Comparing contingent entry and entry mode probabilities
One can now compare the respective entry and entry mode probabilities for
sequential entrants for the cases of ?rm F
1
entering via Green?eld Investment
or via M&A. With O
2
uniformly distributed between 0 and 1 the probabilities
are easily matched with the respective ?xed costs threshold levels for the
respective entry mode
10
.
The following analysis is restricted to cases where parameters t, O
D
, F are
such that all probabilities are larger than zero and smaller than one. Intu-
itively that means we look at markets where all entry modes are possible in
general, so depending on operating cost levels of a respective potential entrant
market structure leads to heterogeneous pro?t-maximizing entry strategies,
with some ?rms preferring entry via M&A, others preferring Green?eld entry
and some maximizing pro?t by not entering at all.
Taking into account the threshold levels and the characteristics of the
uniform distribution, the probability of any sequential entry given that ?rm
F
1
entered via Green?eld is
P
1
= O
GF
2
=
t
16
?F (3.10)
The probability of any sequential entry given that ?rm F
1
entered via
M&A is
P
2
= O
MA
2
=
t
9
?F (3.11)
10
The probability of e.g. O
i
< O
i
for any distribution of O
i
simply equals the cumulative
distribution function from the lower bound of the distribution up to O
i
. For a uniform
distribution between bounds a = 0 and b = 1 the cumulative distribution then is
F(O
i
) =
O
i
?a
b?a
therefore F(O
i
) = O
i
58
Intuitively, a potential sequential entrant with overhead costs below these
threshold level will enter the market in some mode, while if the potential
entrant has higher costs he will not enter, as entry would result in negative
pro?ts then. The following proposition directly follows from comparing the
above probabilities.
Proposition 1
The probability of sequential entry is lower if ?rm 1 entered via Green?eld
Investment instead of M&A. Potential sequential entrants deterred from entry
in the former case are ?rms with ?xed operating costs of O
MA
2
< O
2
< O
GF
2
.
The probability of sequential entry is reduced by P
E
=
7
144
t.
The proposition is derived in appendix 3. Note that the reduction of
the probability of sequential entry is larger the larger, the degree of prod-
uct di?erentiation t in the market. The reasoning is, that the larger t, the
stronger the negative e?ect on pro?ts of an increase in the number of market
participants is for market participants
11
.
The absolute probability of sequential entry via M&A equals the oper-
ating cost threshold level for M&A entry of the potential sequential entrant
F
2
, therefore
P
3
=
O
2
GF
= ?
t
72
+O
D
+F (3.12)
if F
1
entered via Green?eld Investment, and
P
4
=
O
2
MA
=
t
36
+O
D
+F (3.13)
if F
1
entered via M&A.
Comparing these probabilities gives further insight into the e?ect of early
entrant’s decisions on successive entry modes, as stated in Lemma 2.
Lemma 2
The absolute probability of sequential entry via M&A is lower if ?rm 1
enters via Green?eld Investment instead of M&A . Potential sequential en-
trants with
O
2
MA
< O
2
<
O
2
GF
would enter via Green?eld Investment in the
former case and via M&A in the latter. The probability of sequential entry
via M&A is reduced by
t
24
.
11
This can also be seen technically, as the equilibrium price chargeable in the market is
p =
1
n
t, so
?p
?n?t
= ?n
?2
. This shows that an increase in the number of market participants
reduces prices in the market stronger, the larger the degree of product di?erentiation.
59
The Lemma is proved in appendix 4.
As we will discuss later on, the most important question for the early
entrant is, how its’ entry mode will a?ect the absolute probability of se-
quential entry via Green?eld Investment. This probability simply equals the
probability of any kind of sequential entry minus the probability of sequential
entry via M&A. So the absolute probability of sequential entry via Green?eld
Investment is
P
1
?P
3
=
t
16
?F ?[?
t
72
+F +O
D
] =
11
144
t ?2F ?O
D
for early entry via Green?eld Investment and
P
2
?P
4
=
t
9
?F ?[
t
36
+F +O
D
] =
t
12
?2F ?O
D
for early entry via M&A.
Again comparing these probabilities the following important Lemma can
be stated.
Lemma 3
Early entry via Green?eld Investment compared to entry via M&A reduces
the probability of sequential entry via Green?eld Investment by
t
144
.
The Lemma is derived in Appendix 5.
The e?ects of the early movers’ entry mode on F
2
’s entry decision can be
subsumed by the following graphic.
60
Figure 3.3: The E?ect of Early Entry Mode on Sequential Entry
To sum up, early entry via Green?eld Investment compared to early entry
via M&A has the following e?ects on sequential entry. For one, the absolute
probability of sequential entry is lower in the former case. This is however
not a su?cient result for our following analysis, as it is shown, that part of
the total reduction of entry also occurs via the reduction of the probability of
sequential entry via M&A. Further analysis however shows, that early entry
via Green?eld Investment is shown to de?nitely decrease the probability of
sequential entry via Green?eld Investment, which is the actual harmful type
of sequential entry from the point of view of the early mover in this model.
3.2.2 The optimal entry mode of the early mover
Now knowing how the entry mode of the early mover shapes sequential entry
probabilities, one can analyze the optimal entry mode for the early-moving
?rm F
1
.
To that end, pro?ts of ?rm F
1
under entry via Green?eld and entry via
M&A, given the respective e?ects on sequential entry, are simply compared.
Firm 1 pro?t when entering via Green?eld Investment can be written as
61
?
1
(GF) =
?
V AR
1
(3) which is the variable pro?t of F
1
in period 1
+P
3
×?
V AR
1
(3) which is the probability weighted variable pro?t of F
1
with F
2
entering via M&A in period 2
+(P
1
?P
3
) ×?
V AR
1
(4) which is the probability weighted variable pro?t of F
1
with F
2
entering via Green?eld in period 2
+(1 ?P
1
) ×?
V AR
1
(3) is the probability weighted variable pro?t of F
1
with F
2
not entering in period 2
?2O
1
?F is the ?xed costs of operation in both periods and of Green?eld entry
for F
1
Inserting variable pro?ts and rearranging the terms we can rewrite pro?t
under Green?eld as
?
1
(GF) = 2 ×
_
t
9
?O
1
?
F
2
_
?(P
1
?P
3
) ×
7
144
t (3.14)
where (P
1
?P
3
) ×
7
144
t constitutes the expected negative e?ect of sequen-
tial entry via Green?eld on ?rm 1 pro?ts, as (P
1
?P
3
) constitutes the ab-
solute probability of sequential entry via Green?eld Investment.
In the proposed setting, it is quite obvious that the only sequential en-
try mode a?ecting ?rm 1 pro?ts negatively is Green?eld Investment as this
increases the number of ?rms in the market
12
.
Firm 1 pro?t when entering via M&A can be written in similar style
12
Entry via M&A in this model does not change pro?ts of ?rm 1 as it is assumed that
all foreign entrants and local incumbents have the same marginal costs. Therefore sequen-
tial entry via M&A, from the point of view of ?rm 1, simply constitutes interchanging
identically behaving ?rms in the market.
62
?
1
(MA) =
?
V AR
1
(2) ??
V AR
A
(3) which is the variable pro?t of F
1
minus the ”variable part of
the acquisition price” in period 1
+P
4
×
_
?
V AR
1
(2) ??
V AR
A
(3)
_
is the probability weighted variable pro?t of F
1
minus the ”variable part of the acquisition price” with F
2
entering via M&A in
period 2
+(P
2
?P
4
) ×
_
?
V AR
1
(3) ??
V AR
A
(4)
_
is the probability weighted variable pro?t
of F
1
minus the ”variable part of the acquisition price” with F
2
entering via
Green?eld in period 2
+(1 ?P
2
) ×
_
?
V AR
1
(2) ??
V AR
A
(3)
_
is the probability weighted variable pro?t of
F
1
minus the ”variable part of the acquisition price” with F
2
not entering in
period 2
?4O
1
+ 2O
D
is the ?xed costs of operation in both periods and the e?ect of a
targets ?xed costs on the acquisition price.
Again inserting variable pro?ts and rearranging yields
?
1
(MA) = 2 ×
_
5
36
t ?2O
1
+O
D
_
?(P
2
?P
4
) ×
13
144
t (3.15)
with (P
2
?P
4
) ×
13
144
t again being the expected negative pro?t e?ect of
sequential entry via Green?eld Investment of ?rm 1 pro?ts.
Solving for respective operating ?xed costs pro?t levels like it has been
done for the sequential entrant in T=2 the following result can be derived.
Lemma 4
Firm 1 will enter via M&A, if its ?xed costs of operation are low, such
that
O
1
<
¯
O
1
=
t
36
+O
D
+
F
2
?(P
2
?P
4
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
via Green?eld Investment if ?xed costs of operation are of medium size,
such that
¯
O
1
< O
1
< O
1
=
t
9
?
F
2
?(P
1
?P
3
) ×
7
288
t
and will not enter the market if ?xed costs of operation are high, such
that
O
1
> O
1
63
The Lemma is derived in Appendix 6. The general pecking order of the
entry mode of the early moving ?rm 1 is not in?uenced by the threat of
sequential entry and therefore arising strategic e?ects of entry mode, which
is unsurprising due to the way the model is setup, as this order is determined
by assumptions on the in?uence of entry modes on the ex post operating
e?ciency of the foreign structure in the market.
The more interesting question is whether the threat of sequential entry
reduces the probability or amount of M&A and Green?eld entry in compar-
ison to a world, where there no such threat exists. In order to analyze this,
a two-period benchmark case without a sequential entry threat is derived.
For further use, the relative probability of early entry via M&A given
general entry is simply
¯
O
1
O
1
, which intuitively is simply the proportion of
?rms with operating ?xed costs such that they enter via M&A, divided by
the proportion of ?rms with operating ?xed costs such that they enter the
market at all.
3.2.3 Benchmark Case (No sequential entry)
The following benchmark case without a sequential entry threat is con-
structed. The benchmark would be a two-period market game with perfectly
symmetric periods concerning market structure from the perspective of the
early entrant.
In such a setting, cumulative pro?ts of ?rm 1 under Green?eld Investment
over both periods are then
?
BM
1
(GF) = 2 ×
_
t
9
?O
1
?
F
2
_
(3.16)
which is simply twice the pro?t of ?rm 1 in period 1 under Green?eld
Investment
13
.
Cumulative pro?ts of ?rm 1 under M&A entry are therefore
14
?
BM
1
(MA) = 2 ×
_
5
36
t ?2O
1
+O
D
_
(3.17)
13
With three players in the market in each pro?t each ?rm makes variable pro?ts of
1
9
t
per period. Fixed costs of entry are F, which occur only once, so per period ?xed costs of
Green?eld entry are
F
2
.
14
Entry via M&A leaves two players in the market generating per-period variable pro?ts
of
1
4
t. Foregone target pro?ts per period would have occurred in a market with three
players, so the acquisition price per period would be
1
9
t ?O
D
.
64
Solving for ?xed operation cost levels O
1
one ?nds that in the benchmark
case ?rm 1 will enter via M&A if
O
1
<
¯
O
BM
1
=
t
36
+O
D
+
F
2
(3.18)
will not enter the market if
O
1
> O
BM
1
=
t
9
?
F
2
(3.19)
and for ?xed overhead cost levels in-between ?rm 1 would enter via Green-
?eld Investment.
It can be immediately seen, that
¯
O
BM
1
=
¯
O
1
+ (P
2
?P
4
) ×
13
288
t ?(P
1
?P
3
) ×
7
288
t
and
O
BM
1
= O
1
+ (P
1
?P
3
) ×
7
288
t
The relative probability of entry via M&A given general entry is
¯
O
BM
1
O
BM
1
for
the benchmark case.
3.2.4 How does the threat of sequential entry change
the entry mode decision of early movers?
Comparing the benchmark case with the case of potential sequential entry
yields the following ?rst result.
Lemma 5
The threat of sequential entry leads to a lower probability of general entry
for early movers.
The result is derived in Appendix 7.
This Lemma is very intuitive. With some positive probability another
?rm will enter the market in period 2 via Green?eld Investment in the case of
potential sequential entry. Therefore, with this positive probability, variable
pro?ts of the early mover will be lower in period 2. So some potential early
moving ?rms able to make small positive pro?ts in a market with two
65
other competitors will in expectation make negative pro?ts, as there is the
possibility of having to compete with three other competitors in period 2.
Our main interest however concerns the non-trivial e?ect of introducing
the threat of sequential entry on the relative probability of the early entrant
choosing entry via Green?eld Investment over entry via M&A. Analyzing this
question gives the following main result of the basic model.
Proposition 2
The threat of sequential entry leads to entrants choosing Green?eld In-
vestment over M&A with a higher probability, formally
¯
O
BM
1
O
BM
1
>
¯
O
1
O
1
if the general structure of the market supports all available entry modes
with positive probability.
This proposition is proved in Appendix 8.
The main result is easily explained intuitively. With sequential entry,
Green?eld Investment, compared to early entry via M&A, reduces the prob-
ability of sequential Green?eld entry, which would reduce early mover pro?ts.
Due to expectations, acquisition prices for targets will also be lower for
the early entrants with the threat of sequential entry, as incumbent domestic
?rms expect lower pro?ts in period 2. However the variable pro?t-bene?t of
entry via M&A decreases disproportionately due to the non-linear relation-
ship between pro?ts and number of ?rms in the market.
Altogether the existence of a sequential entry threat makes M&A a less
interesting entry mode option compared to a static setting.
The di?erence between this model of sequential entry and the benchmark
case is illustrated by the following graphical example, where it is quite easy
to see that the relative probability of M&A entry decreases strongly when
modelling sequential entry.
66
Figure 3.4: Comparison of Benchmark and Sequential Entry Case
In the proposed model setup there are markets that would only be entered
via M&A by early movers, no matter sequential entry threat or not. This
would be the case if either ?xed costs of Green?eld entry and ?xed operative
costs of incumbents are su?ciently large
15
. One can easily see that O
1
?
O
1
would become negative in this case as well as P
2
> P
4
, such that entry would
only occur via M&A.
In general, however, neglecting the market dynamics e?ects of entry
modes underestimates the attractiveness of entry via Green?eld Investment.
15
Restricted to F being small enough for potential acquirers entering via Green?eld
Investment if takeover negotiations fail and operating ?xed costs O
D
such that no passive
consolidation takes place.
67
3.2.5 A note on completely endogenous market struc-
ture
For simplicity the model only discusses at 2 potential entrants. In a com-
plete equilibrium we would have sequential entry in period 2 until pro?ts for
potential entrants become zero.
However, if e.g. one would assume n potential symmetric sequential en-
trants with operating ?xed costs of O
2
the qualitative results would not
change at all. In fact entry deterrence considerations would become more
important for ?rm 1 as its pro?ts would reduce much more for n
Entry
> 1
?rms entering the market.
3.3 Welfare analysis
In this section it is shown how welfare e?ects of entry modes di?er between
a model without potential sequential entry and with such a threat. To that
end, the welfare e?ects of ?rm 1 entering via Green?eld and via M&A are
analyzed.
As a simple welfare measure, the sum of pro?ts of domestic ?rms (which
include acquisition prices if a domestic ?rm is acquired) and the consumer
rent in the market, is used.
Welfare in the case of ?rm 1 entering via Green?eld is
W
GF
= ?
GF
Dom
+CR
GF
= ?
GF
A
+?
GF
B
+CR
GF
(3.20)
where ?
GF
Dom
is the sum of pro?ts of domestic incumbent ?rms A and B in
both periods and CR is the consumer rent in both periods, which in general
per period is
_
s ?td ?p
_
×X, where s is consumer willingness to pay, td is
the average transport costs incurred by customers and p is the equilibrium
price in the market and X is the trade volume in the market. As we assumed
market size equal to 1 and s large enough, such that total demand in the
market equals market size, CR reduces to s ?p ?td
16
.
Total domestic pro?ts are then simply
17
16
It is common knowledge, that average transport costs for customers in the Salop setup
equal
t
4n
with n the number of ?rms in the market.
17
to reduce notational clutter in the following it is assumed that c = 0 without loss of
generality.
68
?
GF
Dom
=
4
9
t ?4O
D
?P
1
×
7
72
t +P
3
×
7
144
t
18
where it is straightforward, that
4
9
t ? 4O
D
= ?
GF
Dom
(BM) is the pro?t of
domestic ?rms without potential sequential entry. Compared to the bench-
mark case, one can observe that domestic ?rm pro?ts are lower in markets
that generally support all entry modes
19
, which is obvious as a potential
additional competitor reduces pro?ts for each market participant.
Inserting case dependent consumer rents and case probabilities and rear-
ranging we get the following consumer rent in case of early entry via Green-
?eld Investment
CR
GF
= 2s ?
7
18
t + (P
1
?P
3
) ×
5
72
t
where again it is straightforward, that 2s ?
7
18
t = CR
GF
(BM) is the
consumer rent without potential sequential entry. Again comparing to the
benchmark case we can see that consumer rent is higher in markets that gen-
erally support all entry modes, which is intuitive, as an additional sequential
entrant reduces both prices to be paid as well as average transport costs for
the consumers.
Welfare in the case of entry via Green?eld Investment by ?rm 1 is there-
fore
W
GF
= ?
GF
Dom
+CR
GF
=
t
18
?4O
D
+ 2s ?P
1
t
36
?P
3
t
48
(3.21)
or W
GF
= W
GF
(BM) ?P
1
t
36
?P
3
t
48
.
Deriving welfare under early entry via M&A works the same way.
The sum of pro?ts of domestic ?rms in this case is
?
MA
Dom
=
13
18
t ?4O
D
?P
2
27
144
t +P
4
7
144
t
18
Deriving these pro?ts is simple. If ?rm 1 enters via Green?eld Investment both local
?rms make pro?ts
t
9
? O
D
in period 1 respectively. In period 2, if a sequential entrant
enters via Green?eld both local ?rms make pro?ts
t
16
?O
D
each, if the sequential entrant
enters via M&A one local ?rm gets his outside option as the takeover target of
t
16
? O
D
and the other local ?rm makes pro?ts
t
16
?O
D
. If no sequential entry occurs, local ?rms
make pro?ts
t
9
? O
D
in period 2. Multiplying with respective probabilities of sequential
entry decisions yields the above domestic pro?t level.
19
So P
3
< P
1
69
where again ?
MA
Dom
(BM) =
13
18
t ?4O
D
is straightforward.
Consumer rent in the case of early entry via M&A is
CR
MA
= 2s ?
3
4
t + (P
2
?P
4
) ×
13
72
t
Therefore welfare in this case is
W
MA
= ?
MA
Dom
+CR
MA
= 2s ?
t
36
?4O
D
?P
2
t
144
?P
4
19
144
t (3.22)
where again 2s ?
t
36
?4O
D
= W
MA
(BM).
A comparison between the welfare e?ect of entry modes between the
benchmark case and the case of sequential entry can now be made.
The welfare di?erence between entry via Green?eld Investment and entry
via M&A in the benchmark case is
?W
BM
= W
BM
(GF) ?W
BM
(MA) =
t
36
so in the benchmark case entry via Green?eld Investment is the welfare
maximizing entry mode, yielding
t
36
higher welfare than entry via M&A.
In the case of sequential entry after some rearranging, the welfare di?er-
ence can be shown to be
?W = W(GF) ?W(MA) =
t
36
?
t
144
[?
31
72
t ?19F ?16O
D
]
. ¸¸ .
0
or
?W = W(GF) ?W(MA) = ?W
BM
?
t
144
[?
31
72
t ?19F ?16O
D
]
. ¸¸ .
0
(3.23)
Therefore the following propositions can directly be stated.
Proposition 3
Early entry via Green?eld Investment is the welfare maximizing mode of
entry, with or without potential sequential entry.
70
Proposition 4
With sequential entry, the welfare advantage of entry via Green?eld In-
vestment over entry via M&A is more pronounced, than without the threat of
sequential entry.
Whereas the ?rst proposition does not come as a surprise, the second
result is not so obvious. Intuitively entry via M&A in the case of potential
sequential entry accommodates further entry. However such accommodated
further entry, besides increasing expected consumer rent, also decreases ex-
pected pro?ts (including acquisition prices) for the domestic ?rms. In the
proposed setup, the second e?ect dominates the ?rst, not least because only
sequential entry via Green?eld Investment increases consumer rent, whereas
both types of entry reduce domestic pro?ts for at least one ?rm
20
.
3.4 Markets with restricted takeover possi-
bilities
Another interesting result can be derived from the model if one changes the
setup such, that only 1 out of 2 incumbents can be taken over. Assume that
only ?rm A can be taken over, while B, due to various possible reasons
21
can not be taken over.
Intuitively, one might at ?rst guess that in this setup the choice of entry
mode of the early mover shifts towards M&A, because by taking over ?rm A
it disables the sequential entrant to enter via M&A, as there are no targets
left in this case. So, by taking away this option from sequential entrants, one
might think ?rm 1 should be better o?. Therefore M&A might seem to be a
more appealing entry mode for ?rm 1 now.
However with symmetric ?rms concerning marginal costs this is not the
case as will be shown. The reason is, that, with symmetric ?rms, sequential
20
The reasoning runs through acquisition prices to be paid. If the sequential entrant
enters via M&A he pays the target its outside option, which is pro?ts the latter would
make if the sequential entrant enters via Green?eld Investment. Therefore the threat of
Green?eld Investment reduces the targeted ?rm’s pro?ts even when being taken over, so
the sequential entrant entering via M&A.
21
e.g. the ?rm might be state-owned with the state having a strategic interest in keeping
the ?rm state-owned, or the ?rm might be family-owned, where the family might not be
interested in selling the ?rm to some entrant due to non-monetary reasons.
71
entry via M&A is harmless to the early mover, whereas Green?eld entry re-
duces early mover pro?ts in period 2. Within the setup of restricted takeover
probabilities now ?rm 1, by choosing to enter via M&A, does not change the
entry consideration of sequential entrants that would have entered via Green-
?eld anyway, but it will push sequential entrants that would have entered via
M&A if a target had been available to enter via Green?eld Investment now.
We show this in the following.
If ?rm F
1
enters via Green?eld nothing changes compared to the base
model, as no additional limitation is introduced for the strategic entry deci-
sion of the sequential entrant.
Sequential entry (mode) probabilities still are P
1
= O
GF
2
=
1
16
t ? F and
P
3
=
O
2
GF
= ?
1
72
t +O
D
+F.
If ?rm F
1
enters via M&A the probability of any entry is still P
2
= O
MA
2
=
1
9
t ?F, but now P
4
= 0, as there is no available target for ?rm 2 left in the
market now.
Analyzing ?rm 1 behaviour there is no di?erence concerning the threshold
level of operative costs for general entry of F
1
.
A look at the threshold level for entry via M&A for ?rm 1 points out the
di?erence to the base speci?cation.
In the base model
¯
O
Base
1
=
1
36
t +O
D
+
F
2
?(P
2
?P
4
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
Without a target for sequential entrants this reduces to
¯
O
Extension
1
=
1
36
t +O
D
+
F
2
+ (0 ?P
2
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t (3.24)
It is straightforward that
¯
O
Extension
1
=
1
36
t +O
D
+
F
2
+ (0 ?P
2
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
<
¯
O
Base
1
=
1
36
t +O
D
+
F
2
+ (P
4
?P
2
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
as the inequality reduces to 0 < P
4
×
13
288
t, which is true for t > 0 and
P
4
> 0, which is true for all markets that in general support entry via M&A.
As O
Extension
1
= O
Base
1
it is also straightforward that the relative probabil-
ity of M&A entry of the early mover ?rm 1
¯
O
Extension
1
O
Extension
1
is lower with restricted
takeover possibility than in the basic model
¯
O
Base
1
O
Base
1
.
72
As the benchmark case is una?ected by the proposed change in the model
setup it then must be true that
Proposition 5
The negative e?ect of the threat of sequential entry on the relative proba-
bility of early entry via M&A is more pronounced, when takeover possibilities
in the market are limited, such that early entry via M&A eliminates the
option of sequential entrants to enter via M&A. In absolute terms such lim-
ited takeover possibilities increase the probability of early movers entering via
Green?eld Investment instead of entering via M&A.
This result is quite astonishing. Whereas general literature discusses ob-
vious e?ects of missing targets on entry mode choice, namely that there is
no possibility for a ?rm to enter via M&A if no target is available for this
?rm, we come up with an additional indirect missing target e?ect, due to
strategic considerations about potential sequential entry taken into account
by an early entrant.
3.5 Country-speci?c learning-by-doing e?ects
A ?nal interesting extension of the model proposed is incorporating the
availability of learning-by-doing e?ects in the market. With country-speci?c
learning-by-doing e?ects it is meant, that in the respective industry increas-
ing sales volume in other countries does not change the e?ciency of a ?rm in
the respective host country market, such that e?ciency of a ?rm in a market
only increases with the volume of former (sales) experience in this speci?c
market.
This extension is considered in a further simpli?ed version of the basic
model. We will see that, if learning-by-doing e?ects are strong for some
kind of industry, M&A is actually the sequential entry deterring mode for
early movers and therefore it is obvious that the threat of sequential entry
leads early movers to rather enter the market via M&A compared to the
benchmark case.
Assume for simplicity that in period T=1 all ?rms have marginal costs
of c = 1. Further assume that incumbents have been in the market for
such a long time that they have already used up all learning-by-doing e?ects
available, such that c
A
= c
B
= 1 in both periods. The potential sequential
entrant does by design not participate in the market inT = 1, so his marginal
costs in T=2 are also c
2
= c = 1.
73
For the early entering ?rm 1 assume that his marginal costs in period 2 are
a function of how much 1 has sold in the market in period T=1. Speci?cally
let us assume
c
1
(T = 2) =
1 if x
1
(T = 1) < 0, 5
0 if x
1
(T = 1) ? 0, 5
This is a very simple, special form of learning-by-doing e?ects, but the
results can be generalized
22
. By assuming this speci?c form we can abstract
from any strategic selling behaviour of ?rm 1 in period 1 as well as reduce
notational clutter.
3.5.1 Sales volumes in period T=1
With ?rms 1, A, B being symmetric in marginal costs and market size equal
to 1 it is straightforward that the sales volume of ?rm 1 in period 1 is x
1
(T =
1) =
1
3
if 1 enters via Green?eld Investment and x
1
(T = 1) =
1
2
if 1 enters
via M&A.
Therefore c
1
(T = 2) = 1 if 1 enters via Green?eld and c
1
(T = 2) = 0 if 1
enters via M&A.
3.5.2 Sequential entry probabilities in period T=2
If F
1
enters via Green?eld Investment results do not di?er from the base
model, as marginal costs of market participants are unchanged.
The probability of any entry as well as the absolute probability of entry
via M&A is still P
1
=
1
16
t ?F and P
3
= ?
1
72
t +O
D
+F respectively.
If F
1
now enters via M&A the sequential entry probabilities di?er from
the base model, due to the additional e?ect of M&A on marginal costs of
?rm 1.
The probability of any entry in T=2 is now
23
22
It is always true that ?rst period sales will be higher for ?rm 1, if it enters via M&A
compared to Green?eld Investment. Therefore its marginal costs in the second period will
always be lower if it enters via M&A compared to Green?eld Investment.
23
One has to stay simple in this analysis, therefore I keep assuming that ?rms will
locate equidistantly from each other, which is a strict assumption, but actually even coun-
terbalances our results, so the assumption does not drive our results. To be concrete, it
is additionally assumed that the sequential entrant locates farthest away from the early
entrant at the other side of the Salop circle with the two local incumbents in between.
74
P
2
=
1
t
_
1
3
t ?
1
5
_
2
?F (3.25)
and the absolute probability of entry via M&A is now
P
4
=
1
t
_
_
1
2
t ?
1
3
_
2
?2 ×
_
1
3
t ?
1
5
_
2
_
+O
D
+F (3.26)
Analyzing probabilities we come up with the main interesting result of
analysis of this extension.
Proposition 6
With learning-by-doing e?ects in the market and product di?erentiation
low early-mover entry via M&A compared to entry via Green?eld deters se-
quential entry. If product di?erentiation is high early-mover entry via M&A
still accommodates sequential entry.
The proposition is proved in Appendix 9.
The intuition is straightforward. If product di?erentiation is of low de-
gree, the base model-e?ect of entry deterrence via Green?eld through an
increase in the number of market participants faced by a sequential entrant
is small, as has been shown before. With learning-by-doing-e?ects entry via
M&A leads to ?rm 1 being able to price more aggressively in period 2, due to
lower marginal costs than under Green?eld Investment. This second ”‘e?-
cient competitor”’e?ect then dominates the ?rst ”‘more competitors”’ e?ect
for su?ciently low levels of transport costs, as a more e?cient competitor
hurts potential sequential entrants in expectations more than a larger number
of competitors.
The following ?gure shows how the respective probabilities of entry and
entry modes depend on ”‘transport costs”’ t in this extension.
75
Figure 3.5: Contingent Sequential Entry Probabilities and
Transport Costs
2 2,1 2,2 2,3 2,4 2,5 2,6 2,7 2,8 2,9 3
t
P
r
o
b
a
b
i
l
i
t
i
e
s
P1
P2
P3
P4
For F=0,01; O
D
=0,05
How the threat of sequential entry then in?uences the entry mode choice
of the early mover is obvious. As the threat of sequential entry already
reduces the acquisition price to be paid by the early mover as well as the
strategic entry deterring e?ect also favouring M&A entry it follows that
Proposition 7
In markets with low degrees of product di?erentiation and strong learning-
by-doing e?ects the threat of sequential entry increases the relative probability
of entry via M&A.
3.6 Conclusion
Modelling the threat of sequential entry makes it possible to discuss more
forward-looking entry mode strategies of potential foreign direct investors.
I come to the conclusion, that when the ”competition” e?ect, due to a
di?erence in the number of competitors in the market is the main di?erence
between entry modes, Green?eld Investment of early entrants compared to
entry via M&A has the pro?t-enhancing e?ect of reducing the probability of
harmful sequential entry by other foreign ?rms. Even though the former en-
try mode might additionally have the negative e?ect of pushing the sequential
entrant’s decision towards Green?eld Investment, the absolute entry deter-
ring e?ect is pro?t-enhancing, as the absolute probability of sequential entry
via Green?eld Investment is reduced.
76
And while acquisition prices will also be lower with the threat of sequential
entry, it is show, that early movers will choose Green?eld Investment with
a higher probability than in a static world without sequential entry, as the
entry deterring e?ect on pro?ts outweighs all other e?ects.
The above result holds true for all markets where both kind of entry
modes are chosen by the group of potential entrants with positive probability,
so if the degree of product di?erentiation, ?xed costs of Green?eld entry and
?xed cost structure of incumbent banks give support to both M&A as well
as Green?eld Investment. Obviously there are markets, where modelling
sequential entry does not change entry mode choices, due to incentives for
one speci?c entry mode being too dominant.
Additionally it is shown in the model, that with sequential entry, the
e?ect of early entry via M&A instead of entry via Green?eld Investment is
even more harmful to welfare than in the static case. The reason is, that
early entry via M&A accommodates further entry of foreign ?rms, which
decreases domestic ?rm pro?ts by more than it increases customer rent.
Slightly changing the setup of the model to account for ”scarcity” of
takeover targets, further shifts early entrants’ incentives towards Green?eld
Investment, due to the e?ect that entry via M&A in this setup takes away
the entry mode option for sequential entrants that is less/not harmful to
early mover pro?ts. I therefore ?nd a ”‘perverse”’ missing target e?ect on
entry mode choice, such that missing targets for other entrants pushes an
early entrant towards entering via Green?eld Investment.
All in all the proposed basic model helps explain the still very signi?cant
share of Green?eld Investment in total FDI (see e.g. Ra?, Ryan and Staehler
(2006)[113] via the market structure e?ect of entry modes and the latter e?ect
on entry dynamics.
Extending the model to account for country-speci?c learning by doing
e?ects in a stylized way, it is found that in such a setting the early mover
can become a ”terrifyingly” e?cient competitor in period 2 by entering via
M&A early on. The e?ect of becoming a stronger competitor via M&A
then dominates the ”competition” e?ect of reducing the number of market
participants by M&A concerning sequential entry probability for some kind
of markets. It can be shown then, that in this extended setup M&A is the
entry deterring early entry mode if the respective market shows a low degree
of product di?erentiation and if country-speci?c learning-by-doing e?ects are
su?ciently strong (like they are modelled here).
77
Referring to the general scope of this thesis, I deem this latter extension
to be a very good ?t for international retail and SME loan markets. Indeed
one can probably state, that product di?erentiation concerning ?nancial ser-
vices for these types of customers is rather low
24
. Also, due to heterogeneous
(banking) regulation in countries, banks (and also probably insurance com-
panies, for that matter) compared to ?rms in e.g. the real sector, might
be less able to transfer knowledge won by learning-by-doing in one market
to other markets. So e?ciency in a speci?c market in the former industries
should mainly depend on the cumulative volume of business in exactly this
market, and not so much on general world-wide level of experience. The fact
that heterogeneous regulation in part prohibits banks from making use of
other market experience in a respective host country market has been shown
by various studies of e.g. Berger et al.(2000)[12], who show that, except U.S.
banks in some speci?c markets, foreign banks are almost always less e?cient
than domestic banks in OECD countries. A similar story is discussed in the
case study, stating that general international market experience only helps
in conduction business in a speci?c country, if the other markets mimic this
speci?c market.
24
At least this should be true ex ante. Ex post there might be some product di?erenti-
ation stemming from an existing lending relation with a respective customer.
78
APPENDIX
Appendix 1: Deriving variable pro?ts
For starters I assume that willingness to pay s for the product is such,
that in equilibrium the whole market Y is served by the ?rms in the market.
In general sales volume of ?rm i can be derived by the indi?erence condition
of the marginal customer of the respective ?rm. The condition yields
x
i
=
p
j
+p
k
?2p
i
2t
+
Y
n
where p
j
,p
k
are the prices of the two closest competitors and n is the
number of ?rms in the market.
Inserting into ?rm i pro?t function yields
?
V ar
i
= (p
i
?c
i
) ×
_
p
j
+p
k
?2p
i
2t
+
Y
n
_
Solving for the optimal price and taking into consideration that all ?rms
will charge the same price due to same marginal costs c we get equilibrium
price charged by all ?rms
p =
Y
n
×t
Inserting into the sales volume we get equilibrium sales volume for all
?rms
x =
Y
n
Therefore variable pro?t for each ?rm is
?
V ar
=
_
Y
n
_
2
×t
With market size 1 and potential number of ?rms ranging between 2
(both foreign entrants enter via M&A or no entry at all) and 4 (both for-
eign entrants enter via Green?eld) the relevant variable pro?t levels for the
analysis are
?
V ar
(2) =
1
4
×t
?
V ar
(3) =
1
9
×t
?
V ar
(3) =
1
16
×t
79
Appendix 2: Proof of Lemma 1
The di?erence between M&A and Green?eld pro?ts for a respective bank
and a respective entry mode of another bank is of the form
= ?
V ar
i
(MA) ??
V ar
i
(GF) ?A +F ?2O
i
+O
i
=
?
V ar
i
(MA) ??
V ar
i
(GF) ?A +F ?O
i
So
d
dO
i
= ?1 < 0
which means the larger O
i
so M&A is preferred over Green?eld Investment
for low values (as ?
V ar
i
(MA) ??
V ar
i
(GF) ?A > 0 for the proposed setup) of
O
i
and vice versa for higher values of O
i
.
Green?eld pro?ts are of form
?
V ar
i
(GF) ?O
i
so
d
dO
i
= ?1 < 0 and entry becomes less likely if O
i
is high.
Appendix 3: Proof of Proposition 1
I want to show that
P
1
< P
2
so
1
16
t ?F <
1
9
t ?F
which reduces to
7
144
t > 0
which is true for t > 0,which is ful?lled by assumption.
The di?erence between P
2
and P
1
is
1
9
t ?
1
16
t =
7
144
t
Appendix 4: Proof of Lemma 2
I want to show that
80
P
3
< P
4
which is equal to P
3
?P
4
< 0. Inserting yields
?
1
72
t ?
1
36
t = ?
1
24
t < 0
true for t > 0, which is true by assumption.
The absolute di?erence between probabilities is P
4
?P
3
=
1
24
t.
Appendix 5: Proof of Lemma 3
We want to show that
P
2
?P
4
> P
1
?P3
equal to
1
12
t ?2F ?O
D
>
11
144
t ?2F ?O
D
which reduces to
1
144
t > 0
which is true for t > 0. which is given by assumption.
The di?erence [P
2
?P
4
] ?[P
1
?P3] =
1
144
t.
Appendix 6: Proof of Lemma 4
Solving pro?t inequality
?
1
(MA) = 2 ×
_
5
18
t ?2O
1
+O
D
_
+ (P
4
?P
2
) ×
13
144
t
>
?
1
(GF) = 2 ×
_
1
9
t ?O
1
?
F
2
_
?(P
1
?P
3
) ×
7
144
t
for O
1
yields
O
1
<
1
36
t +O
D
+
F
2
?(P
2
?P
4
) ×
13
288
t + (P
1
?P
3
) ×
7
288
t
Solving inequality ?
1
(GF) > 0 for O
1
yields
O
1
<
1
9
t ?
F
2
?(P
1
?P
3
) ×
7
288
t
81
Appendix 7: Proof of Lemma 5
It should be shown that an early mover with some overhead ?xed costs
O
1
would enter the market without the threat of sequential entry, but will
not enter if the threat of sequential entry exists. This is true i?
O
BM
1
> O
1
Inserting yields
(P
1
?P
3
) ×
7
288
t > 0
By the pecking order of entry modes we know that P
1
> P
3
as long as
the market structure in general supports both kinds of entry. As additionally
t > 0 it must hold true that RHS > 0.
Appendix 8: Proof of Proposition 2
I want to show that some potential entrant ?rm 1 with operating ?xed
costs of O
1
would enter via M&A without the threat of sequential entry and
via Green?eld if the threat of sequential entry exists. Therefore we got to
show that
¯
O
BM
1
O
BM
1
>
¯
O
1
O
1
Inserting in the RHS yields
¯
O
BM
1
O
BM
1
>
¯
O
BM
1
+
[
(P
4
?P
2
)×
13
288
t+(P
1
?P
3
)×
7
288
t
]
O
BM
1
+(P
3
?P
1
)×
7
288
t
which can be simpli?ed to
(P
3
?P
1
) ×7 ×
¯
O
BM
1
> [(P
4
?P
2
) ×13 + (P
1
?P
3
) ×7] ×O
BM
1
for markets where entry is feasible (O
BM
1
+ (P
3
?P
1
) > 0)
Inserting for probabilities and benchmark threshold levels after some ma-
nipulation yields the following inequality
239
18
t
2
?(428F + 290O
D
) t + (2016O
2
D
+ 5904FO
D
+ 3744F
2
) > 0
Analyzing the LHS yields
82
?LHS
?t
< 0 for t <
3852
239
F +
2610
239
O
D
?LHS
?t
= 0 for t =
3852
239
F +
2610
239
O
D
?LHS
?t
> 0 for t =
3852
239
F +
2610
239
O
D
as well as LHS(t = 0) = 2016O
2
D
+ 5904FO
D
+ 3744F
2
> 0
and LHS(t ??) ??.
So, the LHS has a minimum at t =
3852
239
F +
2610
239
O
D
.
Therefore it is su?cient to show that the inequality LHS > 0 holds for
this minimum.
Inserting the minimizing t-level yields
70488
239
F
2
+
294776
239
FO
D
+
103374
239
O
2
D
> 0
which is ful?lled by F > 0 and O
D
> 0.
Appendix 9: Proof of Proposition 6
I show that P
2
< P
1
for su?ciently low levels of t. P
2
< P
1
i?
P
2
=
1
t
_
1
3
t ?
1
5
_
2
?F <
t
16
?F
which reduces to inequality
7
144
t
2
?
2
15
t +
1
25
< 0
which is ful?lled for t < 2, 4. For t > 2, 4 it is still true that P
2
> P
1
.
83
Chapter 4
When do Banks Follow their
Customers Abroad?
4.1 Introduction
Among the often-stated motives for foreign direct investment (FDI) in the
banking sector, besides the classical market-seeking reasoning, is a bank’s
desire to follow its existing customer base abroad. This motive is well-
established both in the economic as well as the business literature (e.g. Aliber
(1984)[3],Casson (1990)[31], Williams(1997)[130] and Bain, Fung and Harper
(1999)[5]). As Casson (1990) states
”‘... US banks capitalize on their goodwill by following their
customers overseas; the multinationalization of manufacturing
?rms creates a derived demand for the multinationalization of
banks as well”’
Nolle and Seth (1996)[108] cite a study conducted by the U.S. General
Accounting O?ce
1
reporting that in the United States ”‘most foreign banks
serve customers of their home countries. An industry representative told us
that only a few banks are large enough to penetrate through home country
loyalties to attract other customer”’.
1
”‘Foreign Banks: Assessing their Role in the U.S. Banking System”’, Report to the
Ranking Minority Member, Committee on Banking, Housing and Urban A?airs, U.S.
Senate, GAO/GGD-96-26 (1996)[110]
84
Though this latter claim might be too strong
2
, the general statement
drives home the point, that follow your customer(FYC)-considerations are
an important reason for banks to establish a physical presence abroad.
In some instances existing clients even actively lobbied for their respec-
tive primary banks to follow ?rm expansion abroad. Well-documented and
often-stated early examples were U.S. multinational companies US Steel and
DuPont urging Citibank to establish a foreign presence in South America to
provide their local operations with banking services (e.g. Huertas (1990)[80]).
A large empirical and business/case study literature deals with this topic
and indeed ?nds strong indications that this motive plays a signi?cant role
in the multinationalization decision of banks.
In one of the earliest studies Fieleke (1977)[58], in his study on the de-
terminants of U.S. banks’ overseas expansion, found that ”‘?nancial need
of U.S. ?rms abroad”’ was the major factor in U.S. banks foreign location
choice. This early result for U.S. banks has been supported by numerous
studies, e.g. Nigh, Cho and Krishnan (1986)[106].
A similar role for follow your customer motives in the foreign direct in-
vestment decision of multinational banks has been found for foreign banks
entering the U.S. market by e.g. Hultmann and McGee (1989)[82] and Gold-
berg and Grosse (1994)[68] and for bank entry into less developed markets
(Sabi (1987)[116]). For example, Goldberg and Grosse (1994) study bank
sector foreign direct investment in respective U.S. states. They come up
with evidence, that states, that attract a large volume of real sector foreign
direct investment, also attract more bank FDI.
Nolle and Seth (1996)[108] also analyze the U.S. banking market. They
?nd indication that follow your customer strategies indeed seem prevalent
in foreign bank strategy in the U.S. banking market, as evidence points to
foreign banks devoting the dominant part of total extended loans to foreign
real sector ?rms. However their approach yields the some indication, that
this strategy might not be the main reason for foreign banks entering the U.S.
market. Then again, in contrast to some recent remarks in the industry, they
discover, that follow your customer strategies are still at least as important
in the entry decision of foreign banks as in previous times
3
.
2
As at least some banks seem able to attract local customers, see e.g. Berger et
al.(2000)[12].
3
The authors ?nd, that the share of loans to foreign ?rms as a percentage of total loans
of foreign banks in the U.S., after decreasing for some time, had reached and partly even
exceeded former levels again in the 1990s.
85
Even though the topic is has drawn a lot of attention in the empiri-
cal literature, there is a lack of formal theory on this subject. This seems
bothersome, especially when one considers the direction of argumentation
of a growing literature on market-seeking bank foreign direct investment.
This literature is strongly concerned with information asymmetries and prob-
lems of foreign banks when trying to serve local markets (e.g. Dell’Arricia,
Friedman and Marquez (1999)[47], Dell’Arricia and Marquez (2004)[48] and
Lehner (2007)[92]). These theories might not yield a su?cient explanation
for bank FDI that is induced by the follow your customer-motive.
Particularly the question arises, why a home bank, with an existing rela-
tionship to a home country multinational ?rm, does not simply provide bank-
ing services cross-border, or indirectly via providing the loan to the parent
company from its home base, to the multinational ?rm’s foreign subsidiary.
Physical transport costs seem negligible for loans and similar ?nancial ser-
vices, and informational requirements to provide a loan might already be
met by the bank’s general intimate knowledge about the client ?rm due to
previous and ongoing interaction.
Whereas there might still be motivation for following clients to assist in
local cash management and other services requiring face-to-face contact
4
, the
question is whether there also is such motivation concerning the provision of
loans.
In the following, a model is proposed to motivate such latter follow your
customer-behaviour by applying and re?tting a well-established theoretical
literature on the choice of type of ?nancing on this speci?c topic.
The following model setup is based on a view ?rst clearly laid out by Gert-
ner, Scharfstein and Stein (1994)[65], who discuss the bene?ts and costs of
debt ?nancing compared to internal ?nancing in a setup of a two-dimensional
moral hazard problem (managerial e?ort and managerial discretion in di-
verting project payo?s to himself) faced by a ?rm conducting a manager-run
investment project.
Marin and Schnitzer (2006)[97] broaden the scope of such analysis by
introducing a geographic dimension of debt ?nancing in the ?nancing decision
of a multinational ?rm setting up a manager-run subsidiary in a foreign
country. Letting liquidation e?ciency of banks di?er exogenously by their
proximity to the respective investment project, they ?nd that, depending
on project/?rm characteristics, multinational ?rms will either use host or
home/third country bank ?nancing (or ?nancing from internal sources) for
4
However the question then is, whether sales volume of such services are su?cient for
banks to have an incentive to enter a foreign country.
86
their FDI projects.
The trade-o? faced by the investing ?rm here is, that high liquidation e?-
ciency of the chosen bank type on the one hand allows the investing multina-
tional ?rm to capture a larger share of ex post project returns if the project is
successful, as well as in some cases being able to claim a larger excess liquida-
tion value (liquidation value minus debt repayment) in case the project fails.
On the other hand, however, managerial incentives to spend e?ort, therefore
increasing the expected size of the pie (larger expected project payo?s) to be
shared, are negatively a?ected by high liquidation e?ciency, as the manager
expects a lower share of project returns to be available to himself due to
a better outside option of the investing ?rm in negotiations about sharing
project continuation value.
I closely follow this main idea in this paper
5
. However, in order to in-
corporate strategic choice of domestic banks into the model, I allow for this
home bank to potentially make non-zero pro?ts as well as having discretion
in locational choice, or to put it simply, for home country banks to establish a
physical presence in the ?rm subsidiary’s host country. That way the respec-
tive bank has discretion in a pro?t-determining choice of the loan provision
mode, as the decision then faced by this bank is to whether establish a phys-
ical presence abroad or to serve the client from its home base. As the focus
will be on this conditional location decision faced by the bank, the model
abstracts from cases discussed in Marin and Schnitzer (2006)[97], where the
parent ?rm can potentially completely ?nance this FDI project from internal
cash ?ow or wealth.
Additionally the location-speci?c liquidation e?ciency of a bank is endo-
genized. This is ?rst done in a very simple fashion, where I lay out the fact
that physical transport costs might after all play a role in the decision of how
to supply a loan, as physical assets to be liquidated of a bank’s unsuccessful
client abroad might have to be transferred back to the bank’s home country
to sell these assets at a high price. With the help of a simple political econ-
omy story transport costs are then further endogenized by introducing the
ability of host country governments to (partially) keep assets to be liquidated
in the home country, restricting cross-border physical asset ?ow. In a sim-
pli?ed setting the respective government’s incentive to do so are discussed,
contingent on the country’s endowment in human capital. I do so by using
the notion of asset-embedded human capital, that therefore is immanent in
5
I’d really like to thank Prof. Marin for pointing out this paper as a possible starting
point for discussing bank location choice.
87
the liquidation value of the respective project
6
.
Introducing bank location choice as well as country-dependent liquida-
tion values, allows an analysis on why and under which circumstances banks
engage in foreign direct investment induced by the follow your customer-
motive.
The reason for engaging in follow your customer-bank FDI is then a po-
tentially di?erent liquidation e?ciency attained by the bank, in comparison
to the bank supplying the loan to its multinational client without a physical
presence in its client host market. The model therefore o?ers the possibility
to discuss project/?rm-speci?c and host country-speci?c optimal provision
modes for the respective domestic bank, showing that these two dimensions
of contingency are intertwined in shaping the location choice of the bank.
Using basic results attained from the model, the argumentation of Marin
and Schnitzer (2006)[97] concerning the link between foreign direct invest-
ment and international capital ?ows can be scrutinized.
The rest of this chapter is organized as follows. Section 2 lays out the
basic model, analyzing ?rm level, bank level and government level decision
making to at the end come up with the optimal loan provision modes from
the bank’s perspective for respective circumstances. Section 3 concludes.
First, I discuss the empirical observations the model can explain. Second, I
analyze how my results compare to those obtained by Marin and Schnitzer
(2006)[97]. Finally I point to the potential obstacles to exposing the theory
to an empirical test.
4.2 The Model
The following setup is considered. There are two countries, Home and For-
eign. A ?rm (investor) from Home considers entering the market in Foreign
via the establishment of a subsidiary in this host country. In order to do so,
the ?rm has to hire a manager to operate the subsidiary. Additionally the
?rm is cash-strapped in so far, as it is not able to ?nance this foreign direct
investment completely via internal funds. Therefore the ?rm has to take on
a loan of size K from a bank.
The foreign direct investment project of the ?rm analyzed is a two-period
project, yielding return X
1
in period 1 with probability p and 0 with proba-
bility 1?p and return X
2
in period 2 with certainty. The subsidiary manager
controls the success probability p in period 1 by his choice of e?ort level which
6
A similar notion is proposed in a chapter of the upcoming PhD. thesis by Yanhui Wu,
University of Munich.
88
equals the probability of success. The costs of e?ort are assumed to be of
the quadratic form C(p) =
1
2
zp
2
with z ?]0; ?].
The problem from the point of view of the parent ?rm is, that neither the
e?ort level of the manager, nor the project returns in the respective periods
are veri?able. Therefore the choice of e?ort level cannot be in?uenced by
the investor ?rm directly via an e?ort-based contract. Additionally the ?rm
needs to give indirect incentives to the manager to at least partially transfer
project payo?s back to the parent company.
Following Marin and Schnitzer (2006)[97] the two incentive problems can
be called the e?ort problem, and the repayment problem, where the former
must be solved by the parent ?rm by to maximize the expected return of
the FDI project, while the latter must be solved to maximize its share in the
expected return.
The combination of the other stakeholders in the subsidiary, bank and
parent ?rm, has two means to in?uence the actions of the manager. For one,
the parent has the ability to monitor the project closely, therefore being able
to make the payo? partially veri?able. That way it is able to capture a share
? of period 1 project payo?. However, monitoring comes at a cost which is
assumed to be of the quadratic type C(?) =
1
2
?
2
.
Additionally the bank has the right to liquidate the project after period
1, if the loan is not paid back. Precisely, the parent ?rm has to pay back the
demanded repayment D
i
to the bank, else the bank will liquidate. Following
(implicitly) Marin and Schnitzer (2006)[97] it must be the case, that only the
parent ?rm, not the local subsidiary manager has the right to pay back the
loan to the bank
7
.
A negotiation stage between manager and investing parent ?rm, after
period 1 payo?s are realized and before the credit repayment is due, is mod-
elled. If the parent ?rm does not repay the loan to the bank the project
is liquidated by the bank. As long as it is e?cient to continue the project
(X
2
> L
i
), as will be assumed in the following analysis, the parent ?rm and
the subsidiary manager can bargain over the continuation value, leaving both
stakeholders better o? ex post by continuing. Furthermore it is assumed, that
the bargaining power in such a renegotiation is exogenous, with the investing
7
When latter discussing the model it becomes obvious that when the parent ?rm could
decide on whether to repay the loan itself or to make the subsidiary a relatively independent
capital center it would always choose the former. As we will implicitly see, the reason is
that if the manager could repay the loan himself the investor would not be able to extract
anything more than net ? ×X
1
from the manager, therefore not participating in period 2
pro?ts at all.
89
parent ?rm able to claim a fraction 0 < ? < 1 of the continuation value and
the manager therefore able to claim (1 ??) of it.
The banking market structure is as follows. There is a bank monopolist
in the home market, which has an existing relationship with the above men-
tioned ?rm
8
. In the foreign market a large number of operating homogeneous
banks is assumed, such that there is perfect competition between these banks
leading to zero pro?ts for them in equilibrium.
The domestic bank can di?er from the foreign banks in terms of the ability
to e?ciently liquidate the ?rm’s subsidiary if the loan is not paid back. Bank
ex post liquidation value for the client ?rm’s subsidiary is denoted L
d
for the
domestic bank and L
f
for all foreign (local) banks
9
.
Within this setting the domestic bank can decide on whether to set up a
subsidiary abroad itself (follow its customer), or try to win the loan contract
and extend the loan without such a physical presence abroad
10
. The dif-
ference between the two modes of loan provision will be potentially di?ering
liquidation values the bank can achieve, which will be discussed in greater de-
tail below. Finally, if the bank decides to follow its customer abroad it incurs
small ?xed costs of setting up an agency or other organizational structure of
size F
11
.
The structure of the model can be subsumed in the following ?gure.
8
Actually one would not need to assume only one bank being active in the home market,
but from its existing relationship with the respective ?rm the bank analyzed might be able
to act as a monopolist concerning this ?rm due to various reasons.
9
As bank e?ciency should be shaped by characteristics of its main market, the assump-
tion, that banks stemming from the same country should, all else equal, be more similar
to each other concerning e?ciency than banks from other markets, seems reasonable.
10
The latter simply means, that the bank ?nances a project abroad without establishing
a local physical presence there, but does so by either actually giving the loan to the
domestic parent ?rm, which then internally transfers the capital to its subsidiary, or by
extending the loan to the subsidiary cross-border, where in the latter case the contract
still must call for repayment by the parent ?rm only as well as excess liquidation value
in case of failure falling back to the parent ?rm directly. Again this follows the implicit
setup by Marin and Schnitzer (2006)[97].
11
In the remaining paper these ?xed costs will not be discussed explicitly, only func-
tioning as a tiebreaker between loan provision modes when the bank is indi?erent else.
90
Figure 4.1: Time Structure of the Model
Firm
announces
interest in
investing
abroad
Domestic
bank
decides
upon going
abroad
Investor
chooses
bank, hires
manager
X
1
realized,
Negotiation
about return
sharing
Investor
chooses ?,
Manager
chooses p
Bank loan
due
Repayment
Liquidation
X
2
realized
4.2.1 Bank pro?t maximization under given liquida-
tion value
In a ?rst step the domestic banks pro?t maximization decision for a given
liquidation value it can achieve is analyzed.
Two possible cases have to be distinguished, the case of a moderately
cash-strapped ?rm and the case of a severely cash-strapped ?rm. From the
perspective of the bank, the ?rst case translates into a small loan (small K)
compared to the total size of the investment project, whereas in the latter case
the loan size would be large compared to the investment volume (large K).
As the liquidation value should depend on the total size of the investment
project, di?erent relative
K
L
i
result, which lead to completely di?erent risk
structures between these two types of ?rms from a bank’s perspective.
The case of a moderately cash-strapped ?rm
Let us ?rst consider the case of a relatively small loan, such that for all banks
L
i
> K, therefore bank i gets back at least the face value of the loan in all
circumstances, even when the project fails. The expected pro?t for bank i is
then
?
i
= ´ pD
i
+ (1 ? ´ p)min[L
i
; D
i
] ?K (4.1)
where ´ p is the expected success probability, equalling the e?ort level cho-
sen by the manager in equilibrium, of the investment project, depending, as
91
we will see, on a bank’s liquidation e?ciency L
i
.
Financing by a foreign (local) bank
As the local banks are symmetric concerning the liquidation value L
f
they
will demand the same repayment D
f
in equilibrium
12
. As the liquidation
value in this case of a small loan always su?ces to repay at least the loan
size, the project is riskless to all banks. Note also, that even in the case the
liquidation value is very large (L
i
> D
i
), the bank will only be allowed to
keep D
i
and give the excess liquidation value (L
i
?D
i
) to the investing parent
?rm. Therefore the local banks will compete themselves down to demand a
repayment of D
f
= K, such that demanded repayment equals the size of the
loan and these banks make zero pro?ts in expectation
13
.
In order to analyze the e?ort and monitoring choices inside the ?rm,
which determine rents to be distributed, the problem is solved by backward
induction.
If the project is not liquidated before, the project yields a payo? of X
2
in period 2. Due to the non-veri?ability of the payo?, the manager can keep
the whole payo? to himself. Also, there is no more control right the other
stakeholders can use to give the manager an incentive to hand over part of
this payo? ex post.
At the end of period 1, at the negotiation stage between parent ?rm and
subsidiary manager, two possible cases are to be distinguished. If the project
failed in period 1, yielding a return of 0, which happens with probability
(1 ? p
i
) no payo?s can be transferred from the manager to the parent ?rm.
Therefore the loan can not be repaid to the bank by the parent ?rm and the
bank will liquidate the foreign subsidiary. In this case the manager does not
get any rent from the project and bears his e?ort costs of C(p
i
). The invest-
ing parent ?rm’s payo? in this case is the excess liquidation value (L
i
?D
i
)
it gets from the bank minus its cost of monitoring C(?
i
) =
1
2
?
2
i
. Obviously,
monitoring goes to waste, if the project is not successful. If however the
project is successful in period 1, which happens with probability p
i
, yielding
a payo? of X
1
, the manager can pay the investor part of period 2 payo?s ac-
cruing to him from his share of the ?rst-period payo?, to prevent liquidation
12
This symmetric equilibrium is a standard result in such a game of Bertrand Compe-
tition with perfectly symmetric ?rms.
13
As L
i
> K it is straightforward that D
i
= K leads to zero pro?ts in expectation. If
any foreign bank would demand repayment K+ another foreign bank could win the loan
contract and make positive pro?ts by e.g. demanding repayment K +
2
.
92
of the project by enabling the investor to repay the loan. If the negotiation
is not successful, leading to liquidation of the subsidiary, the parent ?rm gets
?
i
X
1
+ (L
i
? D
i
) ?
1
2
?
2
i
and the manager’s payo? is (1 ? ?
i
)X
1
?
1
2
zp
2
i
. As
e?ort and monitoring levels have already been chosen at the stage of nego-
tiation, therefore also costs of e?ort and monitoring sunk at this stage, the
negotiation-relevant outside option of the parent ?rm is therefore (L
i
? D
i
)
and 0 for the subsidiary manager, respectively. As continuation is assumed
to be e?cient we will have the ?rm and the manager getting their outside
option plus their share of the continuation value via renegotiation
14
.
Altogether the respective expected payo?s given ?nancing via a foreign
bank are then
E
I
(f) = K+p
f
[?
f
X
1
+(L
f
?D
f
)+?(X
2
?L
f
)]+(1?p
f
)[L
f
?D
f
]?
1
2
?
2
f
(4.2)
for the parent ?rm and
E
M
(f) = p
f
[(1 ??
f
)X
1
+ (1 ??)(X
2
?L
f
)] ?
1
2
zp
2
f
(4.3)
for the subsidiary manager.
Foreign banks will compete themselves down to a required require repay-
ment
D
f
= K. Inserting for D
f
into (3.2) then yields
14
Note however, that two additional constraints have to be ful?lled to prevent ine?cient
liquidation. For one, the parent ?rm has to be able to extract enough repayment from
the manager to be able to repay the loan, formally ?
i
X
1
+ (L
i
?D
i
) + ?(X
2
?L
i
) ? D
i
or ?
i
X
1
+ ?X
2
+ (1 ? ?)L
i
? 2D
i
, which is ful?lled for the investment project being
su?ciently pro?table and parent ?rm bargaining power vis-a-vis its manager su?ciently
high (note that with e?cient continuation X
2
> L
i
, so
??X
2
+(1??)L
i
??
> 0). Additionally,
the manager’s share of period 1 pro?ts must be su?ciently high to be able to transfer
the above payment to the parent ?rm, formally (1 ? ?)X
1
> (L
i
? D
i
) + ?(X
2
? L
i
) or
D
i
> (1??)L
i
+?X
2
?(1??)X
1
. Note that the latter constraint does not present an upper
bound to required repayment levels of banks! I assume both of these constraints to be
ful?lled, because else the project would never be continued (if these constraints would not
at least hold for one type of bank i). I check all upcoming results on whether they interfere
with these constraints but only report when they do so. The implicit constraints will not
play any role in the qualitative results but are just stated for the sake of completeness.
93
E
I
(f) = p
f
[?
f
X
1
+L
f
+?(X
2
?L
f
)] + (1 ?p
f
)L
f
?
1
2
?
2
f
(4.4)
To ?nd the equilibrium payo? for the parent ?rm one has to analyze the
e?ort and monitoring choice of manager and parent ?rm respectively.
In appendix 1 the equilibrium e?ort level chosen by the manager under
?nancing by a foreign (local) bank is derived.
p
f
=
X
1
+(1??)(X
2
?L
f
)
z+X
2
1
which is larger zero if project continuation is e?cient, and
?
f
=
X
1
[X
1
+(1??)(X
2
?L
f
)]
z+X
2
1
is the equilibrium monitoring level chosen by the parent ?rm
15
.
The intuition behind the above e?ects of some right-hand side variables
are quite obvious. (1 ? ?) denotes the bargaining power of the manager
in negotiating sharing the second period payo?. The higher the manager’s
expected share in second period payo?, the higher is his willingness to spend
e?ort, which leads to a higher probability of the project to continue into
the second period. From the perspective of the parent ?rm, this leads to
the choice of a higher monitoring level. The reason is that the investor ?rm
can now take a large share of period 1 payo? for himself without harming
the managers’ incentive to spend e?ort too much, as the latter is su?ciently
incentivised to do so by his high share of second period payo? when the
project is continued
16
. Also, a higher continuation value, due to similar
reasoning, induces the manager to spend more e?ort and the parent ?rm to
15
Before discussing these intermediate results, one should note, that in the following
the analysis for moderately cash-strapped ?rms (due to the structure of the analysis this
restriction does not play a role for the analysis of severely cash-strapped ?rms) has to be
restricted to levels of variables, such that p
i
< 1 and ?
i
< 1,as neither pro?t shares nor
probabilities can be allowed to exceed 100%. Evaluating the inequality at the equilibrium
levels of the former variables, we can state this restriction e.g. via the de?nition of su?cient
levels of e?ort cost parameter z. So the analysis is restricted to e?ort costs and project
payo? characteristics such that
z ? Max[(1 ??)(X
2
?L
i
)X
1
; X
1
(1 ?X
1
) + (1 ??)(X
2
?L
i
)].
16
In absolute terms ?
f
is still restricted, as the manager must still have su?cient liquidity
(1 ??
f
)X
1
to induce the parent ?rm to not let the project get liquidated by the bank.
94
monitor more. A higher ”‘marginal”’ e?ort cost z leads both the manager
to spend less e?ort and the parent to monitor less. Whereas the former
is obvious, the latter stems from the fact that, with higher e?ort costs, the
parent ?rm would destroy e?ort incentives of the manager too much by taking
away too large a share of ?rst-period pro?ts.
Financing by the domestic bank
The domestic bank is allowed to generally di?er from the local foreign
banks by the liquidation value L
d
it is able to generate. As the payo? of
the parent ?rm, who chooses the bank it wants to work with, positively
depends on the equilibrium monitoring and e?ort level ?
i
and p
i
, parent
?rm payo? directly and indirectly depends on the liquidation value of the
bank used in ?nancing. Intuitively a higher liquidation value increases the
expected payo? of the parent ?rm when the project fails in the ?rst period,
due to a higher excess liquidation value falling back to the parent ?rm, and
also enhances the parent ?rm’s position in bargaining over second period
pro?ts. Additionally, a higher liquidation value however reduces the e?ort
the manager will spend for the project, as the outside option of the parent
?rm in bargaining over second period payo?s increases, leaving less payo? of
continuing for the manager in expectation.
As the domestic bank can have a liquidation value di?ering from the local
foreign banks it might be able to demand a higher loan repayment, therefore
generating positive pro?ts for the bank. For this bank it need not be true
that D
d
= K.
The expected payo? of the parent ?rm under ?nancing by a domestic
bank is
E
I
(d) = p
d
[?
d
X
1
+(L
d
?D
d
)+?(X
2
?L
d
)]+(1?p
d
)(L
d
?D
d
)+K?
1
2
?
2
d
(4.5)
which reduces to
E
I
(d) = p
d
[?
d
X
1
+?(X
2
?L
d
)] + (L
d
?D
d
) +K ?
1
2
?
2
d
(4.6)
and the expected payo? of the manager is
E
M
(d) = p
d
[(1 ??
d
)X
1
+ (1 ??)(X
2
?L
d
)] ?
1
2
zp
2
d
(4.7)
The payo? maximizing e?ort and monitoring levels are of identical struc-
ture to the case of ?nancing by a local foreign bank, the only di?erence being
the di?erent liquidation values, formally:
95
p
d
=
X
1
+(1??)(X
2
?L
d
)
z+X
2
1
and
?
d
=
X
1
[X
1
+(1??)(X
2
?L
d
)]
z+X
2
1
Note that ?
d
and p
d
are (negatively) depending on the respective banks’
liquidation e?ciency
17
, but not on the size of repayment demanded by the
domestic bank.
The optimal choice of the repayment size for the domestic bank
One can now derive what the optimal and feasible repayment for the
domestic bank looks like. As the size of repayment D
d
does not in?uence the
equilibrium e?ort and monitoring level, it is true that the optimal demanded
repayment D
d
equals the maximum feasible (contract-winning) repayment
D. When choosing the level of required repayment the domestic bank has to
take into account that, if it requires too large a repayment, the parent ?rm
will instead choose ?nancing by a local bank. So the domestic bank has to
choose D
d
such that it leaves the parent ?rm with its outside option, which is
the expected payo? for the latter when ?nancing the FDI project via a local
bank. Therefore the maximum feasible required repayment D can be derived
from the condition under which the parent ?rm is indi?erent between using
the domestic or a local bank, formally
p
d
[?
d
X
1
+?(X
2
?L
d
)] + (L
d
?D) +K ?
1
2
?
2
d
=
p
f
[?
f
X
1
+?(X
2
?L
f
)] +L
f
?
1
2
?
2
f
Solving the indi?erence condition for D yields
D =
p
d
[?
d
X
1
+?(X
2
?L
d
)] ?p
f
[?
f
X
1
+?(X
2
?L
f
)]
. ¸¸ .
I
+(L
d
?L
f
)
. ¸¸ .
II
?
1
2
(?
2
d
??
2
f
)
. ¸¸ .
III
+K
(4.8)
17
One can easily see that both ?
d
and p
d
negatively depend on L
d
, as
??
d
?L
d
= ?
(1??)X
1
z+X
2
1
<
0 and
?p
d
?L
d
= ?
1??
z+X
2
1
< 0.
96
The ?rst part (I) of the RHS captures the e?ort and repayment e?ect
of choosing a domestic instead of a local foreign bank. The second part (II)
captures the direct e?ect of excess liquidation value when choosing a domestic
over a foreign bank. The third part (III) shows the e?ect on monitoring costs,
due to di?erent equilibrium monitoring levels chosen by the ?rm, for di?erent
bank types.
Obviously if the domestic bank generates the same liquidation value as
the local banks, it can at the maximum require a repayment that leads to
zero pro?ts for the bank in expectations D = K
18
.
As I later want to discuss the choice of the domestic bank between follow-
ing its customer or not, which will a?ect the liquidation e?ciency L
d
of the
bank, one ?rst has to understand how the maximum requirable repayment
for the bank depends on its liquidation value
19
.
Di?erentiating D with respect to L
d
, as is done in appendix 2, one ?nds
dD
dL
d
= 1 ?? (4.9)
with ? = p
d
? ?
?p
d
?L
d
[?
d
X
1
+?(X
2
?L
d
)]
Intuitively, the maximum repayment premium D?D
f
over foreign banks
the domestic bank can demand equals the investing ?rm’s willingness to pay
for the domestic bank’s di?erent liquidation e?ciency. The ?rm in general
looks for a bank generating a relatively high liquidation value if the resulting
bene?t of excess liquidation value and an induced higher feasible share in
payo?s exceeds the disadvantage of reducing the manager’s incentive to spend
e?ort. This is the case if ? < 1. If the higher liquidation value reduces the
manager’s e?ort too much, the parent ?rm would rather like to work with
a bank that achieves a relatively lower liquidation value. This is the case if
? > 1.
18
This can be checked easily: If L
d
= L
f
, then ?
i
and p
i
are also the same for both
types of banks. The equation boils down to D = K = D
f
.
19
The constraint that the parent ?rm has to get a transfer su?ciently high to be able to
pay back the ?rm could potentially be binding here, however only in a quantitative way
and only in the case of very large di?erences in the liquidation e?ciency of the domestic
bank and its foreign competitors. Qualitatively, given the assumption, that the liquidity
constraint holds for zero-pro?t making banks requiring D
f
, the constraint tendentially also
holds for D = D
f
+ . The constraint that the manager has su?cient funds from period
1 pro?ts to transfer back to the parent ?rm can be neglected here, as, as seen above, this
constraint just adds a lower bound to the required repayment D
d
.
97
Technically, the e?ect of the liquidation value on the e?ort problem is
captured by
?p
d
?L
d
< 0. The larger this negative e?ect of a high liquidation
value L
d
on the manager’s e?ort level p
d
, the larger (positive) ?
?p
d
?L
d
[?
d
X
1
+
?(X
2
? L
d
)] > 0 and therefore the larger Omega. So ? < 1 becomes less
likely and therefore it is more likely that the ?rm will prefer a bank with low
liquidation value.
What is going to be the situation in the market for this loan?
If the e?ort problem is su?ciently small, the parent ?rm has a positive
willingness to pay for a higher liquidation e?ciency of a bank. Therefore
the domestic bank can win the loan contract with a demanded repayment
of D > K and make positive expected pro?ts, if it can generate a higher
liquidation value L
d
> L
f
from the investment project than the local banks.
If, in that case, the domestic bank generates a lower liquidation value than
the local banks K < L
d
< L
f
, it could only charge D = D
f
= K at
the maximum, which would lead to zero pro?ts for the bank, whether the
contract is won or not (where the domestic bank will straightforwardly not
win the contract.).
If however the parent ?rm, due to the e?ort problem faced, has a positive
willingness to pay for a lower liquidation e?ciency, the domestic bank wins
the contract and makes positive pro?ts (D > K), only if it creates liquidation
value lower than the local foreign banks. With higher or equal liquidation
e?ciency the domestic bank will make zero pro?ts
20
.
Next up the case of a ?rm that needs to take up a large loan relative to
the size of the investment project is discussed.
The case of a severely cash-strapped ?rm
The di?erence between this and the former case is, that the size of the loan
K the ?rm has to take up to ?nance the project compared to project size is
now so large that K > L
i
, so the liquidation value of the complete project
for any bank does not cover the loan. Therefore, from the perspective of the
banks, the loan is risky now. In case of a project failure the bank will make
a loss K?L
i
. As all banks will only be willing to supply the loan if expected
pro?t is non-negative, they will all ask for a repayment D
i
> K to make
positive ex post pro?ts in the case the project is successful.
20
Note, that the discussion is restricted to cases where all banks have su?ciently high
immanent liquidation e?ciency, such that L
i
> K holds true.
98
Again the expected pro?t of the bank is ?
i
= ´ pD
i
+ (1 ? ´ p)L
i
?K.
The payo?s for parent ?rm and subsidiary manager can intuitively be
derived again. We will immediately see the di?erence to the small loan case.
If the project fails, it is again liquidated by the loan providing bank. As
L
i
< K there is no excess liquidation value left for the parent ?rm. So its
payo? in this case is now 0. If the project is successful, the outside option of
the investing ?rm in bargaining with the manager is 0 and the continuation
value is now (X
2
?D
i
).
The expected payo? of the investing parent ?rm in general therefore is
now
E
I
= K +p
i
[?
i
X
1
+?(X
2
?D
i
)] + (1 ?p
i
)0 ?
1
2
?
2
i
(4.10)
and the expected payo? of the manager is
E
M
= p
i
[(1 ??
i
)X
1
+ (1 ??)(X
2
?D
i
)] ?
1
2
zp
2
i
(4.11)
One again needs to check for the maximum repayment that can be de-
manded by the domestic bank. In this case, the analysis is very straightfor-
ward.
Financing by a foreign (local) bank
As local banks are symmetric, they will again compete themselves down
to zero pro?ts in expectations, so the demanded repayment D
f
is
D
f
=
1
p
f
[K ?(1 ?p
f
)L
f
] (4.12)
The payo? of the parent ?rm when choosing a local bank is now
E
I
(f) = K +p
f
[?
f
X
1
+?(X
2
?D
f
)] ?
1
2
?
2
f
One can again I derive equilibrium e?ort and monitoring levels (see ap-
pendix 3)
p
f
=
X
1
+(1??)(X
2
?D
f
)
z+X
2
1
and
?
f
=
X
1
[X
1
+(1??)(X
2
?D
f
)]
z+X
2
1
99
Note that the equilibrium e?ort and monitoring level are now directly
negatively depending on the demanded repayment of the respective loan-
providing bank.
Financing by the domestic bank
It is straightforward that under ?nancing by the domestic banks expected
parent ?rm payo?, e?ort and monitoring levels respectively, are simply
E
I
(d) = K +p
d
[?
d
X
1
+?(X
2
?D
d
)] ?
1
2
?
2
d
with
p
d
=
X
1
+(1??)(X
2
?D
d
)
z+X
2
1
and
?
d
=
X
1
[X
1
+(1??)(X
2
?D
d
)]
z+X
2
1
The optimal choice of the repayment size for the domestic bank
As
?p
i
?D
i
< 0 and
??
i
?D
i
< 0 it is obvious that the payo? of the parent ?rm
is strictly decreasing in the demanded repayment D
i
irrespective of a bank’s
liquidation e?ciency.
Therefore the maximum repayment the domestic bank can demand equals
the demanded repayment of local foreign banks D = D
f
.
As the local foreign banks compete themselves down to zero pro?ts, D =
D
f
=
1
p
f
[K ? (1 ? p
f
)L
f
] is such that the domestic bank would make zero
pro?ts in expectations, if it generates liquidation value L
d
= L
f
, as expected
pro?t of the domestic bank then is
p
d
D
d
+ (1 ?p
d
)L
d
?K = p
f
D
f
+ (1 ?p
f
)L
f
?K = 0
Therefore the domestic bank would make negative pro?ts if it generates
liquidation value lower than the local foreign banks. In that case the domestic
bank, in order to break even, would have to demand a larger repayment
than feasible D
d
> D
f
= D, therefore not winning the loan contract. If it
generates higher liquidation value it can demand repayment D
d
= D
f
and
make positive expected pro?ts from winning the loan contract as (1?p)L
d
>
(1 ?p)L
f
.
The results of subsection 2.1 can be subsumed by the following Lemma
that can be used in advance.
100
Lemma 1
The domestic bank will choose the liquidation value-maximizing loan pro-
vision mode, if its client is severely cash-strapped or moderately cash-strapped
without facing too large a problem of incentivising its subsidiary manager to
spend e?ort (? < 1). If the bank’s client is moderately cash-strapped and
faces a very large e?ort problem by its manager (? > 1), the bank will choose
the liquidation value-miminizing loan provision mode, however restricted to
L
d
> K.
4.2.2 Loan provision mode choice and endogenous liq-
uidation value
I now want to discuss, what the liquidation value of the domestic bank might
look like, depending on how the bank provides the loan and where it therefore
liquidates the project assets.
For starters, the ”‘basic bank-inherent liquidation value”’ of the project
is discussed. As the domestic bank already has an existing relationship with
the respective ?rm, all else equal, due to the bank already having extended
loans to the ?rm and therefore possibly having generated information about
potential asset takers in the industry in the home country, the domestic bank
should have a higher ”‘inherent liquidation”’ e?ciency in the home country
than the local foreign banks in the foreign country, formally L
0
d
> L
0
f
.
Let us assume that the domestic bank can not sell the liquidated project
assets in the foreign market, if it has no physical presence in this country, as it
can not ?nd potential takers in this market
21
. Therefore, if the bank wants to
provide the loan from its home o?ce, it has to transfer the liquidated assets
back to its home market, where it has knowledge about potential takers of
the assets. However, the bank occurs transportation costs by shipping the
asset cross-border, such that only a fraction of the value L
d
= (1 ? t)L
0
d
arrives in the home country, with 0 < t < 1.
If the domestic bank follows the customer abroad, establishing a physi-
cal presence, it is assumed, that the domestic bank does get to know about
potential asset takers in the foreign market aided by its general industry
knowledge. However, it does not have as good an idea about potential de-
21
This rather strict assumption is made for simplicity only. As long as it holds true, that
liquidation in a market can be done more e?ciently if the respective bank has a physical
presence, and therefore e.g. managers that can locally screen the market for asset-takers,
the following qualitative results hold true.
101
mand in the foreign market compared to the home market only generating
L
d
= ?L
0
d
with 0 < ? < 1 when selling the asset abroad.? can be interpreted
as a proxy for the local information requirement to ?nd the asset-taker with
the highest willingness to pay for the liquidated assets.
The domestic bank’s liquidation value of the project is therefore max-
imized by following its customer, establishing a physical presence abroad,
if
?L
0
d
> (1 ?t)L
0
d
or ? > (1 ?t)
If ? < (1 ?t), the liquidation value is maximized by shipping liquidated
assets back home. Therefore the maximum attainable liquidation value is
independent of the bank following its customer or not, as the advantage of
being present in the local market is not made use of in equilibrium.
Lemma 2
The domestic bank’s liquidation value for the project is maximized by
establishing a physical presence abroad, if transport costs t of repatriating the
physical assets to be liquidated are high relative to the bank’s disadvantage
when selling the assets in the local market (? > (1 ? t)). If ? > (1 ? t) the
maximum liquidation value achievable by the bank is independent of its loan
provision mode choice.
This result does not add too much to a discussion about bank’s contin-
gent liquidation e?ciencies, compared to the exogenous assumptions about
liquidation values made by Marin and Schnitzer (2006). Rather the above
setup just acts as a starting point for the further analysis.
Let us now discuss a further interesting endogenization of the transport
costs of shipping assets abroad from a political (foreign government) perspec-
tive.
Retaining ”‘local”’ assets: Political interference in cross-border as-
set transfer
Assume now that the project assets include human capital. Imagine the
assets of the project that are liquidated to be for example high-tech produc-
tion machinery lines. Therefore an amount of human capital H
FDI
is asset-
embedded. This notion is not wide-spread in the theoretical literature yet,
but a closer look at the intuition shows this concept makes sense. Consider a
subsidiary which operates an assembly line. It buys the single machines for
102
the line. However, to make the line work e?ciently the engineering produc-
tion manager has to align the machinery properly and specify them, so they
operate together in an optimal way. By doing so the engineering production
manager leverages his human capital onto the physical assembly line. Even
if this engineer now is strapped from these assets, part of his human capital
in form of his speci?cation/alignment skills are still present in the physical
assembly line. When a bank after liquidation decides to transfer these assets
back home, therefore the total stock of human capital in the foreign economy
decreases by this level.
Foreign’s economy-wide production function is assumed to be of the sim-
ple form
Y = H
?
K
1??
= (H
FDI
+H
0
)
?
K
1??
(4.13)
where H
0
is foreign’s own endowment in human capital, H
FDI
is the
human capital stock embedded in the assets of the discussed FDI project,
and K is the country’s capital stock
22
.
Assume now that the foreign government can restrict physical asset-
transfer out of the country such that a fraction tL
0
d
and therefore tH
FDI
can
be kept in the country. The government can actively choose t by de?ning
physical asset export restrictions. However, if the government does so, it faces
costs of C(t) = ?t, for example because it has to invest in border patrols, or
more generally because it looses reputation among potential following foreign
investors, or due to other reciprocal actions by other governments.
We assume that the government’s objective is simply to maximize gross
domestic product Y
23
over t, taking the costs of implementing t into consid-
22
Let us abstract from the fact that FDI also increases the capital in the country as it
does not yield additional insight, and taking this into account could also interfere with the
analysis of whether the subsidiary manager can, from a political point of view, actually
transfer period 1 pro?ts back to the mother company. Alternatively, one could distinguish
between portfolio capital (pro?t streams) and physical capital (liquidated assets), such
that the host country government would be able to block transfers of physical capital,
but not of portfolio capital. This assumption would also be absolutely sensible, when you
think about trying to smuggle a suitcase across border, compared to trying to smuggle an
assortment of large machines across border. In this case the following results on human
capital scarcity in the FDI host country would just as well apply to capital scarcity of this
country.
23
This objective function of the government could e.g. be rationalized if the government
levies a withholding tax on all value-added in the economy. In this case, tax income, all
103
eration. So the government maximizes over t
Y = (tH
FDI
+H
0
)
?
K
1??
?C(t) (4.14)
As shown in appendix 4 the optimal asset export restriction t then is
t
?
=
(
?
?K
1??
H
FDI
)
1
??1
?H
0
H
FDI
(4.15)
The simple point to be made is, that
dt
?
dH
0
= ?
1
H
FDI
< 0, so that due to
decreasing marginal productivity of human capital a country with a low hu-
man capital endowment will be less permissive concerning asset-repatriation
than a country with high human capital endowment.
So another intermediate result can be stated.
Lemma 3
”‘Transport costs”’ for repatriation of project assets from the FDI host
country to the (parent ?rm and bank) home country will be larger, the lower
the endowment of the host country in human capital.
So, taking into account Lemma 2, the domestic bank’s liquidation value
of the project is maximized by following its customer, establishing a physical
presence abroad, if the host country is endowed with little human capital,
such that (1 ? t
?
(H
0
))L
0
d
< ?L
0
d
. In this case the bank attains maximum
liquidation value by selling the assets in the host country market. If the re-
spective target country for the ?rm investment is richly endowed with human
capital, such that (1 ?t
?
(H
0
))L
0
d
> ?L
0
d
the domestic bank’s ex post liquida-
tion value is maximized by shipping the assets cross-border and sell them in
the home market. The threshold value for the human capital endowment in
this evaluation is
H
0
= (
?
?K
1??
H
FDI
)
1
??1
?(1 ??)H
FDI
(4.16)
4.2.3 When do banks follow their customer abroad?
Having derived the qualitative optimal liquidation value to be implemented
by a domestic bank, as well as the partially endogenized relationship between
type of loan provision and respective liquidation value for the domestic bank,
these results can now be combined to discuss when a domestic bank should
else equal, would be maximized by a maximization of country GDP.
104
follow its customer abroad
24
. Fixed costs of entry only act as a tiebreaker in
the model if, concerning the liquidation value, the bank is indi?erent whether
to follow the customer or provide the loan without a physical presence abroad.
The below results directly follow from the previous analysis. For starters,
the general structure of the bank decision to follow its customer abroad or not
depends on whether the respective loan project favours banks with high or
low achieved liquidation values, and how these liquidation values are shaped
by the bank’s decision to engage in FYC foreign direct investment itself or
not. If a maximized liquidation value is pro?t-maximizing for the bank, it
will choose the provision mode that maximizes the liquidation value it can
generate. If a (relatively) low liquidation value is pro?t-maximizing, the bank
will choose the loan provision mode that leads to the lowest liquidation value
it can generate. Additionally, the relative liquidation value for the domes-
tic bank, compared to the characteristics of local foreign banks, determines
whether the loan contract can be won by the domestic bank.
For the latter consideration, two other following threshold levels play a
role in the analysis, namely the threshold level for human capital endowment
H
0
, such that the liquidation value for the domestic bank, when selling the
liquidated asset in its home market, equals the liquidation value for the local
foreign banks, formally (1?t
?
(H
0
))L
0
d
= L
0
f
, and ?, such that the liquidation
value for the domestic bank when selling the liquidated asset in the host
country equals the liquidation value for the local foreign banks, formally
?L
0
d
= L
0
f
.
Solving for the respective threshold levels yields
H
0
= (
?
?K
1??
H
FDI
)
1
??1
?(1 ?
L
0
f
L
0
d
)H
FDI
(4.17)
and
? =
L
0
f
L
0
d
(4.18)
Taking above results and Lemmas 1-3 together, and taking into account
the ?xed costs of establishing a physical presence abroad as a tie-breaker,
the following ?nal summarizing proposition can be made
25
.
24
I focus on the ”‘political”’ interpretation of cross-border transport costs for assets.
Of course, one could just as well discuss results when having t depending on geographical
distance.
25
It is implicitly assumed that, even when the foreign government blocks away a rela-
tively large portion in a cross-border asset ?ow, the loan for a moderately cash-strapped
105
Proposition
The domestic bank will follow its customer abroad, establishing a physical
presence in the foreign market, and provide the loan to the ?rm if
a) its customer is either severely cash-strapped (K > L
i
) or moderately
cash-strapped (K < L
i
),while facing a not too large e?ort problem of the
subsidiary manager (? < 1), and the target country for the ?rm’s investment
is poorly endowed with human capital (H
0
< H
0
). Additionally the domestic
bank must not face too high an informational disadvantage in ?nding local
asset-takers, formally ? > ?.
b) its customer is moderately cash-strapped (K < L
i
),the ?rm faces a large
e?ort problem for the subsidiary manager (? > 1), and the host country is
richly endowed with human capital (H
0
> H
0
). Additionally the domestic
bank must face a relatively high informational disadvantage in ?nding local
asset-takers, formally K < ?L
0
d
< L
0
f
.
The domestic bank will supply the loan to its customer without a physical
presence in the subsidiary’s host market if
c) its customer is either severely cash-strapped or moderately cash-strapped,
while facing a small e?ort problem, and the host country is richly endowed
with human capital ((1 ?t(H
0
))L
0
d
> max[?L
0
d
; L
0
f
]).
d) its customer is moderately cash-strapped while facing a large e?ort
problem, and the host country is relatively poorly endowed with human capital
(K < (1 ?t(H
0
))L
0
d
< min[?L
0
d
; L
0
f
]).
The domestic bank will not supply the loan to its customer, leaving the
provision to a local foreign bank if
e) its customer is either severely cash-strapped or moderately cash-strapped,
while facing a small e?ort problem, the host country is poorly endowed with
human capital and the local information disadvantage for selling assets in the
host market is high for the domestic bank (L
0
f
> max[(1 ?t(H
0
))L
0
d
; ?L
0
d
]).
?rm stays non-risky for the domestic bank. This simply restricts the cases to be analyzed,
not changing the general qualitative following results. Allowing t to change the risk type
of a loan from a non-risky to a risky loan would simply add the result, that the lower the
human capital endowment of a host country, the more likely the FDI project is risky from
the point of view of the banks. Then unambiguously the optimal strategy of the domestic
bank is to follow its customer abroad, if this bank is able to overcome the informational
disadvantage of selling the potentially liquidated assets in the host country. This ?nding
simply reinforces the notion that a domestic bank’s physical presence in the FDI host
country maximizes its liquidation e?ciency on its client’s assets, if the host country is
characterized by human capital scarcity.
106
f ) its customer is moderately cash-strapped, while facing a large e?ort
problem, the host country is relatively richly endowed with human capital
and the local information disadvantage for selling assets in the host market
is low for the domestic bank (K < L
0
f
< min[(1 ?t(H
0
))L
0
d
; ?L
0
d
]).
Concerning cases where liquidation value should be minimized by the do-
mestic bank two remarks have to be made. For one, these cases are restricted
to moderately cash-strapped ?rms (K < L
i
), so e.g. minimizing liquida-
tion value by selling assets abroad without a physical presence is ruled out
(L
d
= 0). The domestic bank only wants to reduce liquidation value to the
lower bound L
d
= K in these cases. Below this threshold level the project
would be ex post (after choosing the provisioning mode) risky, which changes
the objectives of the bank.
Additionally it is straightforward that, down to this threshold level L
d
,
it is not ex post ine?cient for the bank to liquidate the assets in the low-
est value-yielding way, as ex post all excess liquidation value (L
d
? K) is
transferred to the client ?rm.
To sum up, in general the loan provision mode choice of the domestic
bank can be mapped by the type of client (?, K), the type of host country
(H
0
, ?, L
0
f
), and bank-speci?c characteristics L
0
d
.
The two following illustrations summarize the qualitative provision mode
outcomes, showing the equilibrium provisioning outcomes given client ?rm
and country/market characteristics.
107
Figure 4.2: Equilibrium Bank Strategies - The Case of Severely
Cash-Strapped or Moderately Cash-Strapped/Small E?ort
Problem Firms
?
H
0
1
FOLLOW YOUR CUSTOMER
STRATEGY
LOAN PROVISION
WITHOUT
LOCAL STRUCTURE
LOAN PROVIDED BY
LOCAL FOREIGN BANK
H
0
FDI
d
f
FDI
o H
L
L
H K
H ) 1 ( ) (
0
0
1
1
1
? ? =
?
?
?
?
?
?
0
0
d
f
L
L
= ?
Figure 4.3: Equilibrium Bank Strategies - The Case of
Moderately Cash-Strapped Firms with a Large E?ort Problem
?
H
0
1
LOAN PROVISION WITHOUT
LOCAL STRUCTURE
LOAN PROVIDED BY LOCAL
FOREIGN BANK
FOLLOW YOUR CUSTOMER
STRATEGY
H
0
FDI
d
f
FDI
o H
L
L
H K
H ) 1 ( ) (
0
0
1
1
1
? ? =
?
?
?
?
?
?
0
0
d
f
L
L
= ?
Basically the optimal provision mode (Follow your customer FDI ver-
sus staying at home) of the domestic bank is a non-unique function of the
108
provision mode-speci?c liquidation value attainable. The optimal and the liq-
uidation value-maximizing provision mode are identical in two out of three
project/?rm-characteristics cases. However, in the case of a moderately cash-
strapped ?rm facing a severe e?ort problem concerning the manager, this
result does not hold true. The liquidation-value maximizing mode is shown
to be a function of country-speci?c characteristics.
So, for the majority of cases discussed, the domestic bank will follow its
customer abroad if the latter enters a country relatively poorly endowed with
human capital.
4.3 Conclusion
I proposed a model yielding insight into the decision of banks to enter a
foreign market physically to follow an existing customer abroad or to serve
this customer from home.
The building stone for the analysis is the paper by Marin and Schnitzer
(2006)[97], who discuss the choice of ?nancing mode for an investor/parent
?rm engaging into foreign direct investment, taking into account a double
moral hazard problem on the side of the host country subsidiary manager
concerning e?ort and repayment. Building on their model, the possibility
of pro?t-making domestic banks, having the choice to follow their customer
abroad, has been added, as well as, in a stylized way, an endogenous provision
mode-dependent arising di?erence in this bank’s liquidation e?ciency. The
former remodelling is done to bring in a dimension of general strategic choice
for banks to be analyzed, while the latter uses a simple ”‘political”’ story to
specify the e?ect of host country characteristics on bank follow your customer
FDI.
Through these additions to the model I am able to discuss bank-,
project/?rm- as well as country speci?c determinants of whether a bank
follows its customer abroad or not.
It is ?rst shown that banks unambiguously choose the liquidation-value
maximizing strategy if the respective customer is severely cash-strapped,
making the loan risky from the perspective of banks. Then the liquidation-
value maximizing strategy given country characteristics is described.
I ?nd that, in this case, follow-your customer induced bank FDI tenden-
tially takes place if the host country is poorly endowed with human capital.
Principally it could also be shown, that follow your customer FDI takes place
if home and host country are distant from each other (high t exogenously)
or, in a slightly twisted setup, if the host country is poorly endowed with
109
capital. Additionally the client-following bank must have a su?ciently high
immanent liquidation advantage over local foreign banks to overcome its in-
formational disadvantage (? < 1) in the local market, or stated in another
way, this informational disadvantage must not be too high.
As could be expected from the basic model by Marin and Schnitzer
(2006)[97], the bank will not necessarily choose the liquidation value-maximizing
provision mode, if the respective customer is only moderately cash-strapped,
such that its FDI project can be considered non-risky from a bank’s perspec-
tive.
The intuitive di?erence to the former case, from the point of view of
banks, is, that an increased liquidation value does not directly a?ect bank
pro?ts. Ex post, the loan providing bank will always get K if the ?nanced
multinational ?rm subsidiary fails, if the loan is small compared to the size
of the project and its assets. Therefore the liquidation value only indirectly
determines the maximum repayment that a bank can require in case of a
successful project outcome, as well as whether the bank can actually win the
loan contract without making negative expected pro?ts.
In this case one additionally needs to distinguish between ?rms that, due
to the exogenous relative bargaining power of parent ?rm and subsidiary
manager, the structure of e?ort costs for the subsidiary manager and payo?
characteristics of the project, face a high or low problem in implementing
su?cient e?ort spent by the manager.
In the subcase of a small e?ort problem the domestic bank will choose
the liquidation value-maximizing provision mode, but not so if the e?ort
problem of its customer is large. In the latter subcase, the bank will choose
the liquidation value-minimizing mode (restricted to L
d
> K), as in this
case the negative e?ect of a high liquidation value on the manager’s e?ort
outweighs the positive e?ects on excess liquidation value and on the fraction
of pro?ts transferred from the manager to the parent ?rm.
I think the analysis helps understand what might be deemed to be a
small puzzle of follow your customer FDI by banks. Considering the bank-
ing sector, there does not seem to be a su?ciently satisfying reason why
banks should follow their customers abroad to extend loans to them and not
simply provide the loan cross-border (directly or indirectly) from home, as
”‘transport costs”’ in the classical sense do not really seem to apply to such
cross-border transactions. Also, considering the literature on multinational
banking, which discusses information asymmetries between local and po-
tential entrant banks (e.g. Dell’Arricia, Friedman and Marquez (1999)[47],
Dell’Arricia and Marquez (2004)[48]) and how to solve the problem of in-
formational disadvantages by physically entering the market compared to
110
cross-border lending (e.g. Lehner (2007)[92]), such considerations should
only play a minor role in supplying a loan to a customer the respective bank
already has an existing loan relationship with, therefore already possessing
knowledge about the ?rm.
Follow your customer-behaviour by banks here is explained by the incen-
tive to alter the liquidation e?ciency of the respective bank on a customer
project abroad in a pro?t-enhancing way. One argument in this reasoning is,
that even though transport costs should not play a large role in the provision-
ing of loans directly, they do play a role in the loan provision mode decision
when it comes to the liquidation of assets, if these assets are of physical na-
ture, e.g. an assembly line or similar machinery. Transport costs in various
forms might then occur when shipping these assets to the market where they
generate the highest market value for the liquidating bank.
Introducing country characteristics in the model might help explain why
we observe bank foreign direct investment in countries that at ?rst sight do
not seem to be too attractive for market-seeking bank FDI. Indeed, as shown
in the introduction of this thesis, bank foreign direct investment in developing
countries has increased signi?cantly over the last decade. In the proposed
model, underdeveloped countries with low human capital endowment would
attract follow your customer-induced bank FDI following some real sector
investment projects.
One could however argue, that these countries should also tendentially
not be likely targets for horizontal real sector FDI. But if these countries e.g.
are attractive targets for outsourcing, they do attract vertical real sector FDI.
Within my story, these FDI projects would then yield an incentive for banks
to engage in follow your customer bank FDI in this country themselves, if
the respective real sector ?rms would (weakly)
26
prefer to ?nance the project
with a ”‘high liquidation e?ciency”’ bank type.
Another implicit result, that is not openly discussed in this paper, stem-
ming from the model structure, is, that a parent ?rm starting up a subsidiary
abroad would always demand to be the only party allowed to repay the loan.
If the subsidiary manager itself would be allowed to repay the loan, the par-
ent ?rm would never be able to extract more than D from the manager,
because the latter could else simply repay D to the bank himself to prevent
liquidation of the project. In contrast, with exogenous bargaining power, the
parent ?rm could extract more than repayment R > D if it can threaten to
break o? renegotiation resulting in the liquidation of the project.
26
In the case of a severely cash-strapped ?rm, the ?rm is really indi?erent between banks
concerning their liquidation value.
111
Introducing follow your customer strategies in this model world also sheds
some light on the discussion of Marin and Schnitzer (2006)[97]. By introduc-
ing cases where debt ?nancing would be conducted by local banks, they dis-
cuss settings in which Foreign Direct Investment by ?rms does not constitute
a capital ?ow into the respective host country. They also show empirically
that the ?nancing of inward FDI by local banks is a signi?cant source of
?nancing for entering ?rms.
However, in their model, they do not allow for bank FDI. Empirically,
their dataset unfortunately restricts them to discuss ex post-local banks, so
that they can not distinguish between local foreign banks and local sub-
sidiaries of banks from other countries.
The above theory suggests, that a signi?cant part of these local inward
FDI ?nanciers might actually be subsidiaries of entrant ?rms’ home coun-
try banks. If these subsidiaries primarily re?nance themselves from home
country deposits and/or other domestic sources of ?nance, real sector FDI
?nanced by these subsidiaries, even when the former does not constitute a
direct capital ?ow, leads to an indirect capital ?ow via the bank structure.
So, even though in a strict sense, indeed some FDI projects would not con-
stitute a direct capital ?ow, they still might lead to capital ?ows via local
bank subsidiaries of home country banks ?nancing these projects.
The proposed model yields a variety of testable hypotheses. However,
numerous obstacles for empirical testing of the model exist. For one, bilat-
eral in-depth bank-?rm data for each loan project involving a multinational
?rm subsidiary is needed to observe the pattern of bank provision mode
for di?erent ?rm-/project characteristics. Even country characteristics can
not be discussed without information about ?rm characteristics, due to the
interaction of both types of characteristics.
Also, concerning the discussion of country characteristics on the level of
bank foreign direct investment, one would need explicit information about
whether an observed bank foreign direct investment project is motivated by
follow your customer-considerations, or whether the project is conducted for
local market-seeking reasons. In the latter case, one would expect a positive
in?uence of the level of host country development on the total volume of
bank FDI in?ows, whereas in the above follow your customer-story the e?ect
works in the opposite direction.
As I at this time do not have access to such a data set, empirical testing
of the proposed theory has to be left to future research.
112
APPENDIX
Appendix 1: Equilibrium e?ort and monitoring levels in case of
small loan and ?nancing by a local bank
The parent ?rm maximizes its expected payo? E
I
over the choice of the
monitoring level ? given the ?nancing is conducted by a local foreign bank.
Max
?
E
I
= p
f
[?
f
X
1
+L
f
+?(X
2
?L
f
)] + (1 ?p
f
)L
f
?
1
2
?
2
f
Simultaneously the subsidiary manager maximizes his expected payo?
E
M
over the choice of e?ort level p given the ?nancing is conducted by a
local foreign bank.
Max
p
f
E
M
= p
f
[(1 ??
f
)X
1
+ (1 ??)(X
2
?L
f
)] ?
1
2
zp
2
f
The reaction functions for the parent ?rm and manager respectively are
?
f
= p
f
X
1
and p
f
=
(1??
f
)X
1
+(1??)(X
2
?L
f
)
z
Inserting reaction functions into each other then yields
p
f
=
X
1
+(1??)(X
2
?L
f
)
z+X
2
1
and
?
f
=
X
1
[X
1
+(1??)(X
2
?L
f
)]
z+X
2
1
Appendix 2: Maximum feasible required repayment D and bank’s
liquidation value
I start with the equilibrium maximum feasible required repayment D,
which is
D = p
d
[?
d
X
1
+?(X
2
?L
d
)]?p
f
[?
f
X
1
+?(X
2
?L
f
)]+(L
d
?L
f
)?
1
2
(?
2
d
??
2
f
)+K
The derivative
?D
?L
d
is then
?D
?L
d
= 1 ??p
d
+
?p
d
?L
d
[?
d
X
1
+?(X
2
?L
d
)] +
??
d
?L
d
[p
d
X
1
??
d
]
113
As we know from the payo?-maximization problem of the parent ?rm
?
d
= p
d
X
1
the ?nal term is zero, yielding
?D
?L
d
= 1 ??p
d
+
?p
d
?L
d
[?
d
X
1
+?(X
2
?L
d
)]
Appendix 3: Equilibrium e?ort and monitoring levels in case of
severely cash-strapped ?rm and ?nancing by a local bank
The parent ?rm again maximizes its expected payo? E
I
over the choice
of the monitoring level ? given the ?nancing is conducted by a local foreign
bank.
Max
?
E
I
(f) = p
f
[?
f
X
1
+?(X
2
?D
f
)] ?
1
2
?
2
f
Simultaneously the subsidiary manager maximizes his expected payo?
E
M
(f) over the choice of e?ort level p given the ?nancing is conducted by a
local foreign bank.
Max
p
E
M
= p
f
[(1 ??
f
)X
1
+ (1 ??)(X
2
?D
f
)] ?
1
2
zp
2
f
The reaction functions for the parent ?rm and manager respectively are
?
f
= p
f
X
1
and p
f
=
(1??
f
)X
1
+(1??)(X
2
?D)
z+X
2
1
Inserting reaction functions into each other then yields
p
f
=
X
1
+(1??)(X
2
?D
f
)
z+X
2
1
and
?
?
=
X
1
[X
1
+(1??)(X
2
?D
f
)]
z+X
2
1
Appendix 4: The foreign government’s choice of t
The governments maximization problem is
max
t
Y = (tH
FDI
+H
0
)
?
K
1??
?C(t)
The First Order Condition then is
K
1??
?(tH
FDI
+H
0
)
??1
H
FDI
?? = 0
Solving for t then yields
t
?
=
(
?
?K
1??
H
FDI
)
1
??1
?H
0
H
FDI
114
Chapter 5
The E?ect of Bank Sector
Consolidation through M&A
on Credit Supply to Small and
Medium-Sized Enterprises
5.1 Introduction
In recent years the evolution of market structure in banking can be de-
scribed by two characteristics: An overall consolidation of the banking sector,
strongly driven by Mergers and Acquisitions, and an internationalization of
banking institutes driven by Foreign Direct Investment
1
. This evolution was
made possible by improvements in information technology, ?nancial deregula-
tion, globalization of real and ?nancial markets and, for Europe, the abolition
of exchange rate risks.
International consolidation in the 1990s was partly driven by multina-
tional banks establishing subsidiaries in foreign markets through the acqui-
sition of local incumbent banks. The volume of cross-border M&A involving
target banks in emerging economies for example rose from about 6 billion
USD in the period 1990-1996 to about 50 billion USD in the period 1997-
2000[63].
1
As is discussed in the introductory chapter of this thesis.
115
Still, most of M&A activity in the banking sector was on a national
level, possibly as a reaction to the threat of multinational bank entry, as the
following graph (taken from Berger et al.(2000)[12] illustrates.
Figure 5.1: Volume of M&A Activity in the Banking Sector
Source: Securities data company
In these surroundings two widely-voiced concerns about the changing
structure of the banking sector have come up in public discussion in recent
years.
For one, there is growing concern in OECD countries, that the consol-
idation of the banking sector might lead to reduced credit availability to
small ?rms. The BIS Group of Ten report on Consolidation [23] identi?es
two possible underlying hazardous processes. First, close to the theoretical
story by Stein (2002)[123], larger and more complex credit institutions arising
through consolidation might have a lower propensity to lend to small ?rms.
Second, as it is widely accepted that ”‘relationship lending”’ is an important
characteristic of credit contracts between banks and small ?rms (e.g. Berger
and Udell (2002)[19], a problem on the market level may arise. Relationship
banking is often described to be on the basis of soft, non-veri?able informa-
tion such as management character and this type of information by design is
hardly transferable between banks in contrast to hard, veri?able information
116
like balance sheets or other audited statements. Therefore, small ?rms loos-
ing their old loan relationship with a bank due to consolidation might face
di?culties ?nding new credit partners.
Indeed, Sapienza (2002)[117] ?nds indication that smaller borrowers are
at the loosing end of bank consolidation, as long-standing credit relations
are disrupted and client information transmission to other banks is hard to
process.
If this negative e?ect is for real, it constitutes a serious challenge to devel-
oped economies. First of all, SMEs should be more harmed by a contraction
of credit available to them than large ?rms, as the former tend to be predom-
inantly ?nanced by bank credit (see BIS G10 report)[23]. From the point of
view of the whole economy, small and medium-sized companies accounted
for 66% of total employment in Europe on average and above 50% in the
U.S. and Canada in 1996 (Source: Eurostat). Also SMEs are supposed to
be more ?exible than large ?rms making them key drivers of innovation and
sectoral change. So if possible concerns about small business ?nancing are
right, the problem is of severe economic magnitude.
At the same time there has been growing concern in Less Developed
Countries (LDCs), that foreign bank entry is non-bene?cial or even harmful
to small ?rms in host countries. Stiglitz
2
stated that
‘Foreign bank entry in Argentina............failed in terms of providing ade-
quate ?nancing for small and medium-sized enterprises”’.
One important thing to note here is, that ”‘cross-border market penetra-
tions are often performed via M&As, rather than via opening new branch
o?ces”’(Berger et al.(2000)[12].
Therefore one could argue that both concerns are tightly linked. As a
matter of fact, market entry through M&A is not fundamentally (qualita-
tively) di?erent from in-market or in-country M&A in the banking sector, at
least not in the sense proposed in the following. However, as will be brie?y
discussed in an informal extension of the following model, due to home coun-
try e?ects the impact of national versus multinational consolidation on SME
credit availability may di?er in its strength.
I follow the literature by Diamond (1984)[53] in that small ?rms in respect
to banking really di?er from large ?rms in that they are ”‘informationally
opaque”’. Like Stiglitz and Weiss (1981)[124] point out, ”‘the informational
wedge between insiders and outsiders tends to be more acute for small compa-
nies, which makes the provision of external ?nance particularly challenging”’.
In this chapter a theory is developed, incorporating both the notion of
2
in: El Pais, 10.1.2002
117
the in?uence of organisational characteristics of banks on their lending be-
haviour, using a twisted version of a general model on ?rm organization and
information processing proposed by Stein (2002)[123], as well as the idea of
relationship banking between banks and small ?rms, to give an explanation
for the potentially adverse e?ects of bank sector consolidation through M&A
on credit availability for SMEs. To that end, a fraction of the setup proposed
by Stein (2002)[123], who models the in?uence of internal capital markets on
management decision making, is used. I twist the model by changing its
objectives, introducing managerial choice in a bank on which type of loan
to specialise on, which allows for a direct discussion of the impact of or-
ganizational change on lending strategies, which also enables me to extend
the model to incorporate the idea of relationship banking. In contrast to
Stein (2002)[123], the model is therefore able to yield results concerning the
impact of organizational change within one bank on small ?rm credit avail-
ability on the market level. Through the notion of relationship banking one
can indirectly incorporate third bank behaviour into the model. Additionally,
the chapter discusses rather informal extensions usable to analyze optimal
(small) ?rm policy towards the banking sector, as well as the possible di?er-
ence between national and international M&A from the perspective of small
?rms.
Agreeing with Stein (2002)[123], that the simple notion of a technological
disadvantage of large banks in dealing with small ?rms in an overall sense is
rather too vague, this disadvantage is traced back to the organizational setup
of banks. Also I follow his idea ”‘that the key distinguishing characteristic
of small-business lending is that it relies heavily on information that is soft
..... that cannot be veri?ed by anyone other than the agent who produces
it.”’[123]. In contrast, banks extend credit to large ?rms based on hard-
information such as detailed income statements, balance sheets, etc.., which
in the following model can be learned about by other agents (CEO) inside the
bank but not by the bank’s outside investors. So, to sum up, it is assumed,
that transaction-based lending
3
is predominant in large ?rm credit supply,
while relationship-based lending dominates with small ?rm credit.
Large institutions typically show more layers of hierarchy than small ones.
Assuming that at least some of the decision power lies within higher levels
of the hierarchy, the importance of being able to pass on information to
the next level is more critical to the bottom (loan) manager than in small
institutions. This leads to the manager in a small bank being more likely to
3
Each transaction stands on its own such that information from the relationship is
irrelevant. Transaction-based lending can be further di?erentiated in ?nancial statement
lending, asset-based lending and credit scoring.
118
focus on projects generating soft information, small-?rm credit, than one in
a large bank. In the following model the e?ectiveness of soft information as
a means to get to know loan projects’ outcomes additionally depends on the
length of the bank-borrower relationship capturing the notion of relationship
banking. Due to this aspect, it is also less likely that small ?rms attain credit
from third banks, after their old relationship is cut. I will show that in the
following model, even if the dropped small ?rms attain credit from a third,
non-consolidated, bank, this will most likely come at the expense of other
small ?rms.
The chapter proceeds as follows. Section 2 gives a brief overview about
recent developments in the banking sector, as well as an overview of the
empirical and theoretical literature about possible motives of (cross-border)
consolidation through M&A. In section 3 the basic model about managers’
specialization decisions in small and large banks is laid out. Section 4 ana-
lyzes the market level e?ects of M&A on small ?rm ?nancing for a variety of
cases. Section 5 brie?y discusses the possible di?erence between inter- and
intranational bank sector consolidation. Section 6 deals with consequences
of M&A as described in the existing literature and other related facts about
banking that can get some new formal explanation from my model. Section
7 concludes.
5.2 Bank Sector Consolidation: An Overview
Before analyzing possible consequences of (cross-border) consolidation through
M&A, I ?rst want to give an overview about possible motives and underlying
factors for this kind of consolidation.
Why has consolidation in this industry picked up steam recently? For
sure, changes in the institutional environment like the Riegle-Neal and the
Gramm-Leach-Bliley Act in the U.S.
4
and the Single Market Program as well
as the monetary union in the EU, enabled banks to consolidate in a variety
of new ways and at lesser costs.
Concerning bank’s incentives to take an active part in consolidation, prac-
titioners interviewed in a study by the bank for international settlements
(BIS) mention revenue enhancement and cost savings as the primary motives
4
The Riegle-Neal Act lifted restrictions on interstate banking for U.S. banks, enabling
the industry to consolidate across U.S. states. The Gramm-Leach-Bliley Act allowed banks
to operate in both the commercial and investment banking segment, therefore making
consolidation across these segments legally feasible.
119
for such activity (BIS 2001)[23].Motives unmentioned, quite understandably,
are managerial motives for M&A.
5.2.1 Revenue enhancement as a motive for bank sec-
tor M&A
Revenue enhancement through M&A could come through the following forms
• Economies of Scale and (Geographic) Scope
• Increase in market power
Revenue economies of scale and (geographic) scope
One of the possible motives for/drivers of international bank expansion
overall, and for international M&A as a means of that, is the increased
demand for international ?nancial services by multinational corporations.
Trade in goods increased from 21% of world GDP in 1987 to 40% by 1997
(see World Bank (2004)[7]. Besides that, the volume of FDI and therefore the
geographic separation of production also increased. Such internationally op-
erating ?rms may be in need of an established banking partner in each of the
places they produce or sell their goods. Several empirical studies underline
this ”‘follow your customer”’-strategy of banks. For example, Goldberg and
Gross (1994) found, that foreign direct investment in a U.S. state was strongly
positively linked with foreign banking assets in this state[68](see chapters 2
and 4 for further studies). Revenue economies of geographic scope therefore
arise in the sense, that ?rms might be willing to pay premia for a bank’s
services if the same bank can provide services to the ?rm in other regions of
operation, too.
Special to universal banks providing a large variety of ?nancial services,
scope economies could be at hand through consumers’ willingness to pay a
premium for one-stop shopping, maybe also driven by the consumer’s un-
willingness to share his private information with more than one ?nancial
institution
5
, and through ”‘reputation economies”’. The latter may arise if
a universal bank is able to transfer its superior reputation in one banking
service to another by collective branding (Rajan, 1996)[114].
5
A simple line of reasoning would be transaction costs of documenting information each
time, another reasoning could be along the line of sharing market information with banks
that also provide services to competitors of the respective ?rm.
120
However there might also be diseconomies of scope in the (banking) in-
dustry, arising from less specialisation leading to less tailor-made products
and therefore lower prices chargeable, or due to customer worries that com-
bining services might lead to con?icts of interest within the bank (e.g. Berger
et al. (2000)[12].
Increase in market power
This motive is most probably a prevalent one for national or even regional
bank M&A, to a lesser degree for cross-border consolidation. In general, the
market level price e?ect of M&A depends on the induced increase in local
market concentration and the general demand structure.
In-market M&As in small (local) markets, with the demand side having
few outside options (e.g. small businesses who seem to strongly depend on
local banks for ?nancing)
6
, could most probably enable a consolidating insti-
tution to charge higher prices from their customers through e.g. lower deposit
rates and higher small business loan rates (see e.g. Berger et al.(2000))[12].
Empirical studies back up this conjecture. Banks tend to have better and
more permanent margins in more concentrated markets
7
.
From a market perspective cross-border consolidation, or even the threat
of it, could however decrease the exercise of market power because of in-
creased market contestability in any given country. However, the empirical
results on this are mixed (e.g. Molyneux et al. (1994)[101], Bikker and
Groeneveld (1998)[22] and Cerasi et al.(1998)[32]).
5.2.2 Cost saving as a motive for bank sector M&A
Possible cost savings theoretically arise through
• Economies of Scale
• Economies of Scope
• Increase in Cost X-E?ciency
6
Kwast, Starr-McCluer and Wolken (1997) ?nd that households and small businesses
almost always choose a local ?nancial institution[90].
7
They charge higher rates on small business loans and pay lower deposit rates (Berger
and Hannan (1997)[13] and react slower to changes in open-market interest rates (Jackson
(1997)[83].
121
Cost economies of scale
Practitioners often mention scale as an important means of reducing av-
erage costs in the banking industry [12]. However several empirical studies
for the U.S. found a rather U-shaped relationship between scale and aver-
age costs, suggesting that medium-sized banks ($100 million – $10 billion in
assets) are slightly more cost e?cient than either large or small banks
8
.
But as Berger et al.(2000)[12], as well as several other studies (see The
Economist (2006)[127], mention, most of this empirical literature relies on
data from the 1980s. The former authors argue, that, due to both techno-
logical process (Automated Teller machines (ATMs, Internet Banking, Risk
Management IT) as well as new dimensions of ?nancial engineering (inter-
national placement of bonds, larger scale economies may have arisen. In
retail business, the emergence of internet banking is a classic example for
conducting banking services in an environment of high ?xed costs (such as
the development of the portal) and low variable costs due to less sta?ng
required per transaction. Other possible sources of cost scale economies are
call centers and payment processing.
However from a broader perspective these latter scale economies, due to
the fact that these processes hardly function as USPs
9
of a bank, could prob-
ably be just as well be achieved by small banks outsourcing parts of these
processes or building networks
10
. If banks in general outsource the parts of
their value chain, that feature cost economies of scale, such e?ency consid-
erations concerning the size of the respective bank vanish. One example for
such outsourcing in Germany is Postbank taking over transaction services
for both Deutsche Bank and Dresdner Bank.
Cost economies of scope
Theoretically there are two main contradicting arguments concerning cost
economies of scope in the banking sector. On the one hand re-usability of
customer information for many products may lead to scope e?ciency (e.g.
Greenbaum et al. (1989)[72]), as duplication of e?ort in information research
is impeded. On the other hand a shift away from core competencies always
may lead to additional administrative costs as well as foregone cost reductions
along the learning curve (see Winton (1999)[132]).
8
see e.g. Bauer, Berger and Humphrey (1993)[9] and Clark (1996)[36].
9
Unique selling positions
10
Therefore one could expect banks to rather downsize by vertical disintegration of
operations, outsourcing e.g. call centers to specialised provider.
122
Cost X-E?ciency
Improvements in X-e?ciency through M&A can be achieved, if the ac-
quiring bank has superior managerial skill or organizational practice which
spills over to the target bank.
Simulations by Savage (1991)[118] and Sha?er (1993)[120] lead to the
conclusion, that X-e?ciency can signi?cantly be raised if ine?cient targets
are restructured by X-e?cient acquirers.
However empirical research yields weaker results on whether actual M&As
increased costs X-e?ciency
11
.
5.2.3 E?cient Risk diversi?cation as a motive for bank
sector M&A
Scale and scope economics in risk management
One very plausible motive for (international) consolidation via M&A for
a bank is to improve the risk-expected return tradeo?.
Under the modern theory of ?nancial intermediation (e.g. Diamond
(1984)[53], Diamond (1991) [54]) this argument holds against the traditional
view of capital markets, that investors optimize their portfolio in the risk-
return dimension themselves.
It is hard to di?erentiate however, whether observed risk-adverse behav-
iour of large U.S. banks (see e.g. Hughes and Mester (1998)[81] is for the
bene?t of the shareholder or due to managerial objectives.
Literature does not ?nd that scale plays a very important role for the
tradeo? (except when you think about banks getting so big that they ”‘can
not fail”’ because of state intervention (see Berger et al. (2000)[12]), but en-
hancing scope through geographic and service portfolio diversi?cation might
very well reduce risk without a similar decrease in expected returns.
The possibility for international risk diversi?cation in the banking indus-
try can be observed in Figure 5.2, which is taken from Berger et al.(2000)[12].
One striking example for such diversi?cation possibilities is the correlation
between banks’ Return on Equity (ROE) in France and the U.S. in the span
between 1979 and 1996 which is -0,815.
11
For the U.S. most studies ?nd positive, but small e?ects (e.g. De Young (1997)[50]).
Vander Vennet (1996,1998)[128][129] for Europe found that cross-border consolidation
increased X-e?ciency, but national M&A often failed to do so.
123
Figure 5.2: Correlation Analysis of Bank ROE among Nations
5.2.4 Managerial motives for bank sector M&A
The existing principal-agency literature gives a broad range of reasons why
managers might want to pursue M&As, both nationally and internationally,
as acquirers.
So-called empire-building tendencies of managemers
12
have received am-
ple interest in theoretical literature. For all kinds of M&As, the sphere of
control of the acquiring institution’s manager becomes larger which increases
manager’s utility mostly through reputational e?ects as well as to a lesser
extent through possible increased compensation (e.g. Chevalier and Avery
(1998)[33]).
Cross-border consolidation especially might be motivated from a manage-
rial point of view by two additional points.
First, with shareholders on average still mostly stemming from a ?rm’s
home country
13
, establishing additional business abroad instead of at home
12
e.g. Jensen and Murphy (1990)[84]
13
This home bias in equity has ?rst been ?rst discussed by French and Poterba
124
might enable a manager to enjoy more perquisites or slack o?. The rea-
soning would be, that home country shareholders normally know less about
economic conditions abroad than at home adding more uncertainty about
payo?s which the manager may use to increase his perks or reduce his e?ort.
Shareholders would then hardly know whether reduced pro?t (e?ciency) is
due to conditions in the new host country or due to changed behaviour of
management.
Also, cross-border consolidation might be comfortable for managers in
that it can reduce risk (see subsection above) increasing their job security
14
,
even if this risk reduction is ine?cient from the shareholder’s point of view,
in that yields too low a level of risk-adjusted expected return.
After discussing some motives for banks engaging in (cross-border) M&As,
the focus is now on the main topic of the chapter, namely how such activity
e?ects a speci?c ?rm segment in the market.
5.3 The Model
The setup di?erentiates between large and small banks by their organiza-
tional characteristics. A small bank is assumed to consist of a single loan
manager, whereas a large bank consists of two such loan managers plus a
CEO on top of the organization.
Both kind of banks are exclusively funded by risk-neutral outside investors
with an outside option of zero. Whereas in a small bank the ?nancing relation
is directly between loan manager and outside investor, in a large bank the
capital runs through the hands of the CEO, who receives capital from the
outside investor and subsequently allocates it among his two loan managers.
After a round of lending activity the investor gets back his initial investment
plus all monetary return on investment.
In both banks loan managers have to decide whether to specialise on sup-
plying credit to large ?rms (L) or small ?rms (S) ex ante. Let us assume that
the only di?erence between these two types of ?rms is the type of information
the bank loan manager can extract from them. To be precise, loan managers
in both type of banks can only extract ”‘soft information”’ from small ?rms,
which is non-veri?able to investors and other agents inside the bank, whereas
dealing with large ?rms yields ”‘hard information”’ that is still non-veri?able
to outside investors, but veri?able to other agents inside the bank.
(1991)[64].
14
see e.g. Morck, Shleifer and Vishny (1990)[102]
125
No matter what kind of credit the loan manager specialises on, he is
assumed to always be able to choose between two potential loan projects.
Each loan can take on size K = {0, 1, 2}. The projects can either be in a
good state of the world (G) yielding return g(K) or in a bad state of the
world (B) yielding return b(K). Both states have ex ante probability 1/2.
The respective projects’ states are non-correlated. Total returns for loan
projects controlled by manager i are denoted c
i
.
To structurize the problem and later reduce notational clutter, the fol-
lowing assumptions about project returns are made
15
.
1. ?2 < b(2) < ?1 < 0 < g(1) < 1
2. g(2) = 2g(1)
3. b(2) < min[3b(1) ?g(1); 2b(1)] = 3b(1) ?g(1)
4. g(1) > ?b(1)
Each loan manager can learn about the actual state of either both of his
possible projects or none of them before deciding how to allocate his capital
under control. The probability of the manager learning about the projects’
actual state of the world is
µ =
_
?
1/t
for small ?rms
? for large ?rms
_
with 0 < ?, ? < 1 (5.1)
? is a general e?ciency parameter for a loan manager generating informa-
tion about a small ?rm project. It is assumed, that this signal becomes more
informative, though with decreasing marginalities, over the length of rela-
tionship t ? [1; ?[ between the bank and the respective small ?rm client
16
.
This seems to be a very intuitive setup as soft facts like management charac-
ter usually take some time to explore. To further simplify the problem it is
assumed that there are always two possible small ?rm loan projects at hand
15
Assumption 1 gives us a well-constrained problem to work on. Assumption 2 is made
just to reduce notational clutter. Assumption 3 leads to simpli?ed equilibria later on and
will be discussed in advance. However, this assumption doesn’t change the qualitative
results. Intuitively it states that funding a bad project with 2 units of capital yields very
bad results, therefore such funding is tried to be ruled out by decision makers in the model.
Assumption 4 is necessary for investors funding the bank with, as it guarantees positive
expected pro?ts for investors.
16
I assume without a prior relationship between bank and ?rm that t = 1.
126
that have same relationship length with the respective bank. In the basic
model analysis each small ?rm has only on existing banking relationship.
The chapter discusses the multiple relationship case in an extension.? is the
respective e?ciency parameter for a manager generating information about
a large ?rm loan. As hard information is mostly passed on through stan-
dardized ?nancial statements the signal has a constant value of information
independent of the speci?c relationship.
To make this problem interesting, ? is assumed to increase over time (or
for that matter, ? to have fallen), else there will hardly have any interesting
equilibria on changing loan strategies in banks
17
.
Finally, as in contrast to Stein (2002) this model is less one of mechanism
design, but rather of a combined Nash-Bayesian game, one needs to charac-
terize prior believes of agents in the model. It is assumed, that all players
in the game ex ante always believe that other players will not be successful
in their research ex ante. However, except the extreme case, that players
believe with probability 1 that others will be successful in their research, the
qualitative results are unchanged.
All agents, loan managers and a possible CEO, have utility functions
U
i
= K
i
+c
i
(5.2)
where K
i
is the amount of capital and c
i
the net cash ?ow of projects under
control of agent i. Agents therefore act like ”‘e?cient empire-builders”’, so
agents’ are interested in both getting as much capital as possible under their
control as well as use the allocated capital e?ciently. These preferences,
together with the non-veri?ability of project information to outside investors,
leads to ?nancing constraints for banks in equilibrium, as discussed in Lemma
1.
Lemma 1
Investors will ?nance small banks with two units of capital in equilibrium
if g(1) + b(1) > 0 and b(2) < 2b(1), where the latter is ful?lled by assump-
tion.(see Appendix for proof)
Lemma 2
If Lemma 1 holds investors will ?nance large banks with four units of
capital
17
The rise of ? can very intuitively be explained by e.g. progress in auditing technology
over time. For transition economies a fall in ? could easily be explained by a loss of social
capital in the transition process.
127
I will only discuss this intuitively, as this Lemma obviously follows from
Lemma 1 in combination with assumed prior beliefs of players in the model.
In such a setup expected payo?s for outside investors’ funding a large bank
with K units are exactly like funding two small banks with K/2 units, there-
fore Lemma 2 holds.
Having laid out the general problem, one can now discuss the specialisa-
tion decision in both types of banks.
5.3.1 Inside the small bank
I start out with the much easier case of a small bank. The single loan manager
of this bank will choose his specialisation based on his expected utility level.
As shown above he will always have two units of capital to allocate on his
projects. His expected payo?, depending on allocative action and state of
information, can be summarized by the following payo? matrix (with chosen
allocation on the left and information on project 1 and 2 respectively received
by the manager at the top).
A — I {GG} {GB} {BG} {BB} {None}
(2;0) 2g(1) 2g(1) b(2) b(2) g(1) +
b(2)
2
(1;1) 2g(1) g(1) +b(1) g(1) +b(1) 2b(1) g(1) +b(1)
(0;2) 2g(1) b(2) 2g(1) b(2) g(1) +
b(2)
2
With the assumptions on project return structures made, one can straight-
forwardly see that the manager will choose
• Allocation (1;1) if he does not receive any information or if information
is {BB}
• Allocation (2;0) if he receives information {GB}
• Allocation (0;2) if he receives information {BG}
• Any feasible allocation if he receives information {GG}
His expected utility specialising on small ?rm loans will then be
EU(S) = ?
1/t
×
_
3g(1)
2
+
b(1)
2
_
+ (1 ??
1/t
) ×[g(1) +b(1)] + 2 (5.3)
His expected utility specialising on large ?rm loans will then be
128
EU(L) = ? ×
_
3g(1)
2
+
b(1)
2
_
+ (1 ??) ×[g(1) +b(1)] + 2 (5.4)
By comparing the utility levels one can come up with a causal relation
between the specialisation decision and the length of established relationships
at the point of decision between the loan manager and his existing small ?rm
client base.
Proposition 1
A small bank loan manager will specialise on small ?rm loans if he has
long-standing relationships with his existing small ?rm base. Else he will
specialise on large ?rm loans. The less e?cient general research about small
?rms is relative to research e?ciency about large ?rms, the longer the critical
length of relationship between small ?rm and bank to ensure further ?nancing
of the small ?rm. The critical length of relationship is t = ln ?/ ln ?. (Proof:
See Appendix)
The following graph illustrates the specialisation decision.
129
Figure 5.3: Relationship Length and Specialisation Decision
18
1 2 3 4 5 6 7 8 9 10
Relationship lengths t with small firms
A
c
c
u
r
a
c
y
o
f
I
n
f
o
r
m
a
t
i
o
n
?
1/t
?
Specialisation on
large firms
Specialisation on
small firms
In this case small banks that have existing relationships with small ?rms
with length t > 5 will specialise on small ?rm credit and small banks with
t < 5 will specialise on large ?rm credit.
5.3.2 Inside the large bank
The case of the large bank is far more interesting, as it includes strategic
interaction.
There are four di?erent stages to be analyzed to ?nd the equilibrium
specialisation decision.
1. Unit manager specialisation decision
2. Unit manager decision on whether to report information to CEO or not
3. CEO decision on capital allocation to loan managers
4. Manager decision on capital allocation to loan projects
18
For ? = 0, 1 and ? = 0, 6
130
The model can be solved by backward induction.
Stage 4
From the small bank case we already know the equilibrium allocation strategy
when a loan manager controls two units of capital (Case 1).
However in the case of a large bank managers might have anywhere from
1–4 units of capital to work with.
Case 2) Manager has one unit of capital
The manager’s problem can be summarized by the following payo? matrix
A — I {GG} {GB} {BG} {BB} {None}
(1;0) g(1) g(1) b(1) b(1)
g(1)
2
+
b(1)
2
(0;1) g(1) b(1) g(1) b(1)
g(1)
2
+
b(1)
2
So maximizing his expected payo? the manager will choose allocation
(1;0) in case he receives information {GB} and allocation (0;1) in case he
receives information {BG}. For any other information he will be indi?erent
between feasible allocations.
Case 3) Manager has three units of capital
19
A — I {GG} {GB} {BG} {BB} {None}
(2;1) 3g(1) 2g(1) +b(1) b(2) +g(1) b(2) +b(1)
3g(1)
2
+
b(2)
2
+
b(1)
2
(1;2) 3g(1) b(2) +g(1) 2g(1) +b(1) b(2) +b(1)
3g(1)
2
+
b(2)
2
+
b(1)
2
Maximizing his expected payo? the manager will choose allocation (2;1)
in case he receives information {GB} and allocation (1;2) if he receives infor-
mation {BG}. For any other information he is indi?erent between feasible
allocations.
Case 4) Manager has four units of capital
Here the only feasible allocation is (2;2),as loan volume per project is
restricted to K = 2.
131
Stage 3
Given that the CEO knows the optimal allocation decision of managers at
the loan level he will assign capital to the loan managers such as to maxi-
mize his expected utility. His allocation will therefore be determined by the
information he receives from his loan managers. Remember that the large
bank will be funded with four units of capital in equilibrium. Due to the
non-veri?ability of soft information, the manager’s preference structure and
the research success of loan managers being insecure, the CEO will not be
able to distinguish between good and bad soft information and ”‘silence”’ on
the side of the manager due to ?nding bad hard information
20
.
With the CEO prior probability belief about managers’ research success
being zero
21
, his contingent payo? matrix can be subsumed as in the table
in Appendix A
22
.
We can then identify the optimal contingent capital allocation strategy
for the CEO, as shown in the following table.
20
Non-documentable information as non-veri?able to the CEO will not be taken into
consideration by the CEO, as he understands the manager to always claim to have found
his projects to be in the good state of the world.
21
Therefore, if the manager does not report any documented information, the CEO will
assume that the manager failed in learning about the projects.
22
We do not need to di?er between cases (GB) and (BG) so the later is subsumed in
the former.
132
Information received CEO utility maximizing allocations
{GG}{GG} (4;0);(3;1);(2;2);(1;3);(0;4)
{GG}{none} (4;0)
{GG}{GB} (4;0);(3;1);(2;2)
{GG}{BB} (4;0)
{GB}{GG} (2;2);(1;3);(0;4)
{GB}{none} (2;2)
{GB}{GB} (2;2)
{GB}{BB} (3;1);(2;2)
{BB}{GG} (0;4)
{BB}{none} (2;2)
{BB}{GB} (2;2);(1;3)
{BB}{BB} (2;2)
{none}{GG} (0;4)
{none}{none} (2;2)
{none}{GB} (2;2)
{none}{BB} (2;2)
These optimal allocations follow directly from the analysis of expected
payo?s of the CEO given the information received.
23
Stage 2
One can now analyze the optimal information strategy of a manager towards
the CEO. At this stage it must be di?erentiated between managers who have
chosen to specialise (at stage 1) on small ?rms, possibly generating soft in-
formation, and those who have focused on large ?rms, possibly generating
hard information. It is of no matter whether soft information gets passed on
to the CEO because the latter won’t value this information at all. Also the
manager can not pass on any kind of information if he hasn’t gathered any.
So we can restrict analysis to managers who have gathered hard information
from large ?rms. These managers can choose whether to report their docu-
mented information to the CEO or keep quiet/just state that their projects
are in a good state. The capital allocation to this manager i, conditional on
23
Here is where assumption 3 plays a quantitative role. If not for this assumption, there
would be unclear optimal allocations for information tuples {none}{BB} and {BB}{none},
as the optimal allocation in these cases depends on the relative size ofb(1), b(2) and g(1).
However, the qualitative results are not changed by the assumption.
133
reporting or not, can be analyzed with the help of the following table
24
.
The following table gives the expected capital allocation for each manager
contingent on the information he and the other manager j passes on to the
CEO.
i — j {GG} {GB} {BG} {BB} {None}
{GG} (2;2) (3;1) (3;1) (4;0) (4;0)
{GB} (1;3) (2;2) (2;2) (2,5:1,5) (2;2)
{BG} (1;3) (2;2) (2;2) (2,5:1,5) (2;2)
{BB} (0;4) (1,5;2,5) (1,5;2,5) (2:2) (2;2)
{none} (0;4) (2;2) (2;2) (2:2) (2;2)
The resulting optimal information strategy towards the CEO has the
following main characteristics. First, passing on information to the CEO
when information about both projects is bad is weakly dominated by not
reporting this documented information (”‘none”’). Therefore the manager
will always not report the documented information in that case. Second, for
all other successfully gathered information reporting the actual information
to the CEO always weakly dominates not reporting.
One obvious advantage of specialising on large ?rm loans potentially yield-
ing hard information is evident here. Positive hard information passed on to
the CEO can increase the capital available to the respective loan manager to
2+ units. At the same time negative hard information can be hidden from
the CEO (not passed on to him), therefore no counteracting negative e?ect
of generating hard information exists.
Stage 1
With the help of results from stages 2-4 a manager’s decision on whether to
specialise on small or large ?rms can now be derived. Note that in this ?nal
step one not only has to worry about how the specialisation decision a?ects
capital allocation among managers but also about the way the specialisation
decision changes the projects’ expected net cash ?ows due to di?erences in
research e?ciency and capital allocation.
I analyze a normal game between the two loan managers in a large bank
to ?nd conditional Nash Equilibria for specialising on small ?rms. In order
to do so, ?rst each manager’s expected utility contingent on his own and the
24
For multiple allocation equilibria (see stage 3).I assume that the manager assigns same
probability to all of those optimal allocations and furthermore use the expected allocation
for the analysis.
134
other managers specialisation decision is analyzed
25
.
Given that manager j chooses to specialise on large ?rms, manager i’s
expected utility specialising on small ?rms is:
EU
i,j=L
(S) = E(K
i
|i = S, j = L) +E(c
i
|i = S, j = L) (5.5)
Using the results from stages 2 and 3 (and weighing them with respective
probabilities) yields
E(K
i
|i = S, j = L) = 2 ?
?
2
(5.6)
Note that the expected amount of capital received in this case depends
negatively on research e?ciency for hard information. This is very intuitive:
The more likely the other manager, specialising on large ?rms, attains hard
information about his project the more likely he will get more funding from
the CEO at the cost of the manager specialising on small ?rms.
Adding the results from stage 4 yields
EU
i,j=L
(S) = 2 ?
?
2
+(1 ?
?
4
)
_
(g(1) +b(1) +?
1/t
(0, 5g(1) ?0, 5b(1))
_
(5.7)
Given that manager j chooses to specialise on large ?rms ,manager i’s
expected utility specialising on large ?rms is:
EU
i,j=L
(L) = E(K
i
|i = L, j = L) +E(c
i
|i = L, j = L) (5.8)
Again, using results from stages 2-4 yields
EU
i,j=L
(L) = 2 + (?
2
?6?) [1/8(b(1) ?g(1))] +g(1) +b(1) (5.9)
Once more this shows two intuitive characteristics. For one, capital allo-
cation to each manager is independent of research e?ciency, as both man-
agers specialise on hard information with same e?ciency. Expected utility as
a whole is increasing in ? as allocation among loan projects becomes better,
therefore the expected cash ?ows of the pool of both projects under control
are higher
26
.
25
Note that by the structure of the game the specialisation decision is made before the
manager learns about the then-available loan projects’ state of the world.
26
When e.g. one project is in the good state and the other in the bad state, c
i
will be
g(1) + b(1) when the manager does not get information about the project states and has
2 units of capital to work with and 2g(1) when the manager learns about the states and
can therefore allocate all capital to the ”‘good”’ loan project.
135
The other two conditional expected utilities are constructed in the same
way and are
EU
i,j=S
(S) = 2 +g(1) +b(1) +?
1/t
(0, 5g(1) ?0, 5b(1)) (5.10)
and
EU
i,j=S
(L) = 2 +
?
2
+g(1) +b(1) +?[g(1) ?0, 5b(1)] (5.11)
We can put these results in the following simpli?ed standard game form
De?ning
X ? EU
i,j=S
(S)
Z ? EU
i,j=L
(S)
V ? EU
i,j=S
(L)
Y ? EU
i,j=L
(L)
yields the following normal game form of the problem.
Manager i – Manager j Small Firms Large Firms
Small Firms X;X Z;V
Large Firms V;Z Y;Y
One can now analyze the critical length of relationship t which leads to
both managers specialising on small ?rms.
Obviously, specialisation on small ?rms (S; S) is a Nash Equilibrium if
X > V and it is unique if Z > Y . As managers in large banks are homogenous
in this model it is actually pretty intuitive that the only reasonable equilibria
are (S; S) and (L; L).
Inequality 1 (X > V ) is ful?lled for
2 +g(1) +b(1) +?
1/t
(
1
2
g(1) ?
1
2
b(1))
>
2 +
?
2
+g(1) +b(1) +?(g(1) ?
1
2
b(1))
Rearranging yields
?
1/t
> ?
1 + 2g(1) ?b(1)
g(1) ?b(1)
(5.12)
136
Solving for t yields two case-dependent results:
t >
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.13)
for ? <
g(1)?b(1)
1+2g(1)?b(1)
and
t <
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.14)
for ? >
g(1)?b(1)
1+2g(1)?b(1)
Case 1 (? not too large) yields a feasible threshold level for relationship
length
t
1
=
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.15)
above which (S; S) is a Nash-Equilibrium. Case 2 would yield a negative
threshold level meaning that for large values of ? (S; S) is never a Nash-
Equilibrium. In this case ?
1+2g(1)?b(1)
g(1)?b(1)
is larger than one, so no relationship
length, leading to an increased allocation e?ciency of the manager for a given
volume of capital to work with, is su?cient to o?set the possible capital
allocation advantage (getting more capital from the CEO to work with) of
choosing to specialise on hard information projects.
Inequality 2 (Z > V ) is ful?lled for
2 ?
?
2
+ (1 ?
?
4
)
_
(g(1) +b(1) +?
1/t
_
1
2
g(1) ?
1
2
b(1)
__
>
2 + (?
2
?6?)
_
1
8
(b(1) ?g(1))
_
+g(1) +b(1)
Rearranging yields
?
1/t
> ?[1 +
1 +g(1)
1 ?
?
4
(g(1) ?b(1))
] (5.16)
Again solving for t, we got to distinguish between two cases, RHS < 1
yielding feasible relationship lengths and RHS > 1, where in the latter case
no t can be su?ciently large.
I ?nd, that for ? > ??
_
(??)
2
?4 with ? =
4+9g(1)?5b(1)
2(g(1)?b(1))
no t exists such
that inequality 2 holds.
137
For ? > ? ?
_
(??)
2
?4 we get
t >
ln?
ln[?[1 +
1+g(1)
1?
?
4
(g(1)?b(1))
]
(5.17)
So inequality 2 is ful?lled for t > t
2
(and low levels of ?) with
t
2
=
ln?
ln[?[1 +
1+g(1)
1?
?
4
(g(1)?b(1))
]
(5.18)
Summing up the two results, for low enough levels of e?ciency in gener-
ating hard information and for su?ciently high relationship lengths to small
?rms at hand, loan managers in large banks will specialise on loan provision
to small and medium-sized enterprises
27
.
The following proposition summarizes the above ?ndings on stage 1.
Proposition 2
Loan managers in large banks will always specialise on large ?rms, if
research e?ciency when dealing with hard information is su?ciently large
(? >
1
1+
1+g(1)
g(1)?b(1)
), independent of available relationships to small ?rms.
They will also de?nitely specialise on large ?rms, even if ? is small,
if they do not have very long standing relationships with small ?rms (t <
ln ?
ln
[
?(1+
1+g(1)
g(1)?b(1)
)
]
).
Specialising on small ?rms in a large bank is an equilibrium strategy i?
? <
1
1+
1+g(1)
g(1)?b(1)
and t > t
1
=
ln ?
ln
[
?(1+
1+g(1)
g(1)?b(1)
)
]
. It is an unique equilibrium i?
? < ? ?
_
(??)
2
?4 and t > t
2
=
ln?
ln[?[1+
1+g(1)
1?
?
4
)(g(1)?b(1))
]
.
5.3.3 Small business ?nancing in small and large banks
To sum up, small banks will lend to small ?rms if they have existing rela-
tionships with small ?rms with length at least
t =
ln ?
ln ?
(5.19)
27
This is a results that di?ers from the line of arguing of Stein (2002)[123], where from
his theory one can derive, that large banks would always choose to ?nance large ?rms.
138
In comparison, even in the best possible case for small ?rms (low ? and
managers coordinate on equilibrium (S; S)), large banks will only lend to
small ?rms with which they have at least relationship length
t
1
=
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.20)
It can easily be shown that t
1
> t (see proof in Appendix 3). This means,
that small ?rms having a relationship with their respective bank of length t,
with t < t < t
1
, will get a positive expected loan volume
28
in the period of
interest, if their bank stays small, but zero loan volume for sure if their bank
becomes part of a larger bank structure.
Proposition 3
Small banks are more likely to extend credit to small ?rms than large
banks. Large banks are more likely to extend credit to large ?rms than small
banks. Small business clients, in the period of interest, with relationship
lengths t < t < t
1
with their respective banks will attain an expected loan
volume of 1 if their bank is small, but no loan if their bank is large.(Directly
follows from proof in Appendix 3)
In the following, let us focus on ?rms with such relationship lengths t
with the consolidated institution as the basis of further analysis, as these are
the interesting cases to be studied.
5.4 The Role of Consolidation
Now when one thinks of consolidation as a merger or acquisition between two
small banks, leading to the evolution of a large bank with the respective loan
managers still in place but now headed by an additional CEO, one can easily
discuss the consequences of consolidation on small ?rm credit on a bank and
market level. In the proposed basic model setup the organisational change
has no impact on bank technological parameters ? and ? and all banks share
the same technology.
28
If a client of a small bank, the small ?rm with respective t as above will get an expected
loan volume of (1 ??
1/t
) ×1 +?
1/t
[0, 25 ×4 +0, 25 ×0 +0, 5 ×1] = 1, which can directly
be seen from the expected capital allocation of the manager as laid out in the section on
small banks.
139
There is one, though hardly interesting, ?rst result, namely consolidation
will never have any adverse e?ect on small ?rms if ? = 0. Let us rule out
that possibility for now.
5.4.1 Changing credit supply within the merged bank
Proposition 4
Small ?rms with relationship length t with t < t < t
1
with their respective
small bank will not be supplied with loans in the analyzed period, if their bank
merges or is acquired, while they would have been ?nanced with an expected
loan volume of 1 by this bank if it had stayed independent.
This result directly follows from Proposition 3, as consolidation via M&A
simply changes the structure of the respective banks from small to large. As
was shown in the case of a small bank, small ?rms with relationship length
t < t would not have received credit from their respective small bank even
if it stayed independent. Even in the best possible case small ?rms with
relationship length t, with t < t < t
1
, will not receive credit from the new
banking structure their old bank is consolidated into, but would have done
so, if the small bank had stayed independent.
The following ?gure subsumes, which bank-?rm relationships are a?ected
by the respective bank being involved in M&A activity.
Figure 5.4: Bank-Firm Relationships A?ected by M&A
t
No Loan,
unaffected
by M&A
No Loan,
negatively effected
by M&A
Positive expected
loan volume,
unaffected by M&A
t t
1
140
5.4.2 SME credit supply at the market level
Whether small ?rms are in the end adversely a?ected by consolidation de-
pends on whether other small banks make up for the lost loan supply from
the consolidated bank.
For starters let us focus on the case where small ?rms only have an existing
relationship with just one bank.
If small ?rms have only one standing relationship with a small bank, then
the ?rms de?nitely adversely a?ected by consolidation will be those who had
relationship length t with t < t < t
1
with one of the consolidated small banks.
Again this is straightforward. The small ?rms identi?ed are those who had
a relationship with one of the consolidated banks but will not get ?nanced
by it again in the period of interest.
With any other bank, small or large, they have relationships with length
t = 1.
As a direct consequence from the results of the basic model above, it must
be true, that these ?rms will never receive credit from another large bank.
Whether they stand any chance to receive credit from another non-consolidated
small bank depends on whether this small bank has existing relationships to
small ?rms and how research e?ciency parameters look like.
Small ?rms in general attain credit from small banks in the period of
interest, if they have at least relationship length t = ln ?/ ln ? with this
bank. As t = 1 for the small ?rm it will only possibly get funded by this
third small bank if ? > ?, whereas it would have received credit if its old
relationship bank had stayed independent if ?
1/t
> ?, which is a less binding
constraint for t > 1 with the old bank.
If this third small bank had formerly specialised on small ?rms it will have
relationship length t > 1 with at least two small ?rms. In this case the small
?rm set free by a consolidated institution will not receive credit from the
third bank, even if ? > ?, as the loan manager in this bank achieves higher
expected utility from sticking with supplying loans to its incumbent small
?rm clients, due to better knowledge about them leading to more e?cient
allocations of loans.
So one can conclude that small ?rms will not only be a?ected by the
merger through the direct e?ect, that it is less likely that the newly merged
bank, in which at least one former part was their credit partner, will extend
credit to them, but also by the fact that it is a lot less likely that they receive
credit from a third non-consolidated bank, which constitutes a real problem
at the market level
29
.
Proposition 5
29
To be precise the only potential source of credit for the analyzed ?rm are new entrant
141
Small ?rms with a single relationship with a consolidating bank will su?er
in overall credit availability in expectations through this consolidation. They
will be less likely to get credit from the consolidated institution, as well as be
less likely to receive credit from other banks in the market, due to their lack
of a relationship with other banks. The only potential source of ?nance are
small new entrant banks in a setting where ? > ?.
5.4.3 Small ?rms with multiple bank relationships –
The odd ?rm out
The impact of consolidation on small ?rms above was concerned with small
?rms only having an existing relationship with one bank.
Let us now consider the case where these ?rms have n > 1 relationships
with at least one of the banks not being part of a M&A process, so the
respective ?rm had taken up a loan from 2 banks in its history at some
di?ering times.
In this case there exist di?erent possible scenarios for the overall credit
availability to small ?rms. The easiest case in thinking about the problem is
an economy with three banks i,j,k where i and j are small banks that merger
while k stays independent and the loan market situation for small ?rm s is
analyzed
30
. It should be clear, that a small ?rm s with relations to banks i
and j will not fare any better than if the ?rm only had a relationship with
either i or j, as both banks merge together.
The respective length of relationships between the small ?rm and the
banks are t
i,s
,t
j,s
and t
k,s
.
Already knowing that, if bank k is a large bank it will likely not supply
credit to ?rm s, we focus on the more interesting case of bank k being small.
The benchmark for the following analysis is a single relationship small
?rm who will not attain credit from the consolidated institution. So, for
additional insights compared to the case of a single ?rm-bank relationship,
the focus is on the case t
i,s
< t
1
.
Case 1: Bank k has been specialising on large ?rms so far and t
k,s
> t
banks. Because if the other type of incumbent small banks, that have specialized on large
?rms to date, exists, these will not supply positive expected loan volume to the analyzed
small ?rm in the period of interest, given the assumption of rising ? over time. This is
shown in appendix 4.
30
I do not discuss bank entry here.
142
Due to the assumption of rising ? over time this case does not exist
31
.
Case 2: Bank k has been specialising on large ?rms so far and t
k,s
< t <
t
i,s
If so, the small ?rm will de?nitely not receive any credit from any incum-
bent bank whereas it would with positive probability, if bank i had stayed
independent.
Case 3: Bank k has specialised on small ?rms serving ?rms l,m and
t
k,s
< t
k,l
= t
k,m
.
Even if t
k,s
> t the small ?rm will not receive credit from the third bank
k because the loan manager of the bank has longer standing relationships
at hand, therefore better knowledge about these ?rms and therefore higher
expected utility serving ?rms l,m instead of s.
Case 4:Bank k has specialised on small ?rms serving ?rms l,m and t
k,s
>
t
k,l
= t
k,m
> t
This is the case, where the ?rm su?ering from M&A activity is not the
one set free by a consolidated institution, but other small ?rms are negatively
a?ected by such a development.
Here ?rm s will attain positive expected loan volume from bank k but
only at the expense of either ?rm l or m. Remember that small banks in
the model are restricted to only screen two possible loan projects. So even
though in this case ?rm s still possibly attains credit, at least one other small
?rm will be adversely a?ected by the ripple e?ect of consolidation through
M&A.
The case analysis can be subsumed in the following proposition.
Proposition 6
Keeping up multiple relationships with banks decreases the probability of a
small ?rm being negatively a?ected by consolidation through M&A. However,
as a group, small ?rms overall su?er from consolidation. Small ?rms nega-
tively a?ected by consolidation need not be direct clients of the consolidated
institution.
31
With rising ? if t
k,s
> t in the period of interest it must be true that t
k,s
> t in
the period before. Therefore bank k would have specialised on small ?rms in the earlier
period.
143
5.5 National versus Multinational Consolida-
tion through M&A and heterogeneous coun-
tries
Up to this point I have established a model giving insights into potential
e?ects of overall M&A in the banking sector on small ?rm credit availability.
Let us now discuss an international perspective namely whether in the eye
of small ?rms consolidation involving a foreign bank is better or worse than
pure intranational active consolidation.
How could these two types di?er in a non-trivial way
32
?
One potential di?erence might come in the form of a home country spe-
ci?c bank heterogeneity in research e?ciency. As stated in the basic setup,
the research e?ciency on hard information should intuitively be rather in-
dependent of length of relationship between bank and large ?rm, due to
the standardization of research on hard information. But one factor of how
good managers inside a bank do research on hard information should be the
amount of times they have done that which means the population of large
?rms the manager has dealt with. This is simply the idea of ”‘learning by
doing”’(e.g. Krugman (1987)[89]. As the globalization of ?nancial services is
a rather new phenomenon the size of the home country large ?rm population
should therefore have a positive e?ect on the respective banks’ research e?-
ciency concerning hard information
33
. One possible functional form, similar
to the one used for learning in an international environment by Krugman
(1987)[89], for research e?ciency of bank i is
?
i
= min(?
P
T=1
[X
i,T
+?X
j,T
] ; 1) (5.21)
with 0 < ? < 1;X ? 0;0 < ? < 1
32
Of course intranational M&A, besides changing the organizational structure of banks,
also reduces the number of banks in the market whereas international M&A only changes
the structure of one local bank, leaving the number of banks operating in the market
constant. Let us abstract from this simple di?erence.
33
One could of course also argue, that the di?erence between national and foreign banks
might also depend on the e?ciency ? of generating soft information. As Berger and
Udell (2002)[19] put it ”‘Cross-border consolidation may create additional problems for
relationship lending because a foreign-owned bank may come from a very di?erent market
environment, with a di?erent language, culture,....”’.
144
where X
i,T
denotes the size of the large ?rm population in the bank’s
home country at time T, X
j,T
is the former in all other countries, ? gives the
level of learning by international spillovers and P is the period of interest.
At the maximum the manager always learns about the actual state of
large ?rm credit projects, so learning is bounded at maximum 1.
If bank research e?ciency on hard information depends on the size of
the population of large ?rms in their respective home country, banks from
”‘larger”’ countries will have higher research e?ciency on hard information
than banks from ”‘small”’ countries, as long as ? = 1 in all the respective
countries, as
? theta
i
?X
i,T
> 0 in that case.
What follows from this for the analysis on bank sector consolidation?
If a country is small in the above sense, all else equal, national banks will
have a rather small degree of research e?ciency on hard information ?
medium
.
A foreign bank from an even smaller country would show even smaller ?
low
. In contrast a foreign bank stemming from a larger country would feature
a large ?
high
. Now assume that any of these banks take over or merge with
another domestic bank, with hard information research e?ciency equal to
the maximum of these e?ciencies of both merging banks.
Remember, that in the best case from the point of view of small ?rms
this bank will only lend to small ?rms if
t > t
1
=
ln ?
ln
_
?(1 +
1+g(1)
g(1)?b(1)
)
_
(5.22)
It can easily be shown for bank i that
?t
1
?X
i
> 0 for ?
i
< 1
34
.
So even if a higher ?, induced in the consolidated institution by a merger
with a foreign bank from a large country, does not lead to the bank not
providing loans to any small client ?rm anymore, no matter the lengths
of relationships at hand, it does increase the minimum relationship lengths
required to keep up ?nancing of small enterprise customers.
Summarized, the higher the research e?ciency of the consolidated bank
the less likely the bank keeps serving its existing small ?rm client base.
Proposition 7
Result 1: Bank sector consolidation should have worse e?ects on small
?rm credit supply, the larger the respective country’s large ?rm population is.
34 ?t
1
?X
i
=
?t
1
??
×
??
?X
1
.
It is easy to show that
?t
1
??
=
ln ?
??(ln ?)
2
> 0.
As
??
? X
1
> 0 it follows that
?t
1
?X
i
> 0.
145
Result 2: Consolidation through M&A involving foreign banks is more
harmful to small ?rms than pure national consolidation, if the foreign ac-
quiring/merging bank stems from a country ”‘larger”’ than the host country,
whereas national consolidation is more harmful if the foreign banks come from
a ”‘smaller”’ country.
5.6 The Empirics of Small Business Lending
and Consequences of Bank M&A
What does existing literature have to say about the characteristics of bank
lending to SMEs and consequences of M&A activity in the banking sector?
5.6.1 Characteristics of Small Business Lending
As is assumed in the setup of the model, one very important feature of
small business lending is that it is quite substantially relationship-based.
Numerous studies employing a variety of methods to measure relationship
strength between bank and borrower (e.g. length of relationship, exclusivity
of credit relationship, service scope of relationship) ?nd, that strength of
relationships has positive implications for the respective borrower, in that the
latter for example pay lower interest rates (e.g. Berger and Udell (1995)[17])
and are more likely to get funded by a respective bank (e.g. Scott and
Dunkelberg (1999)[119]).
5.6.2 Consequences of (Cross-border) Consolidation on
Small Business Lending: The Existing Literature
What does the existing literature have to say about the consequences of
(cross-border) M&A in the banking sector on small ?rms?
For in-market M&A, leading to increased market concentration, loan in-
terest rates and service fees might increase due to the increased market power
of banks in this market. Theoretically, small ?rms might be most a?ected by
this development, due to the fact that they are found to be rather immobile
concerning their source of funding (Kwast et al.(1997)[90].
146
Indeed, studies (e.g. Akhavein, Berger and Humphrey (1997)[2]) ?nd that
in-market M&As, substantially increasing market concentration, empower
banks in this market to charge higher prices.
In terms of cross-border consolidation, this is typically not a kind of in-
market M&A from the perspective of small ?rms, even though it might be for
large wholesale customers in investment banking, so it seems that national
M&A in this respect is more likely to be harmful to small ?rms.
Besides the market power implications, M&As also change three charac-
teristics of the involved banks, which, after reviewing the literature, might
a?ect small business credit supply (see Berger et al.(2000)[12], where the
second and third characteristics are the one the proposed model deals with:
• Increase in consolidated bank scale (and scope)
• Increase in consolidated bank organizational complexity
• External e?ects on the lending behaviour of other banks in the market
In general an increase in bank size might lead banks to shift their strategic
focus from small business loans to wholesale services such as underwriting and
other investment banking activities. An explanation could be, that capital
market services can only be provided by large banks due to e.g. a critical
mass of depositors. If banks face an upward sloping supply curve of capital
this possible new market, now feasible to be served, would reduce the capital
of this bank allocated to small business ?nancing.
For bank scope e?ects, as Berger et al.(2000)[12] mention Williamson-
type organizational diseconomies of scale, large banks might be ill-equipped
to conduct relationship-based lending.
Empirical studies concerning the U.S., like Berger and Udell (1996)[18]
and Berger et al.(1998)[16], indeed ?nd a negative relationship between a
bank’s size and the proportion of its assets employed in providing small
business credit.
35
Raising the topic of large banks’ relationship lending-capabilities, Cole,
Goldberg and White (1999)[38] and Berger et al.(2002)[15] ?nd, that large
banks tendencially engage transaction-based lending to small ?rms, whereas
relationships play a much larger role in lending between small banks and
their respective small business clients.
However, whereas ex ante bank scale seems to be a negative indicator on
whether small ?rms receive credit from the respective bank at all, quite a
35
The latter study observed, that small banks (below $100 million assets) use 9% of
assets on small business lending, very large banks (above $10 billion assets) only 2%.
147
few studies show evidence, that large banks seem to be safe havens for the
small ?rms they after all serve in times of ?nancial distress. Hancock and
Wilson (1998)[75] for the U.S. ?nd that ?nancial distress reduced small ?rm
credit in large banks way less than in small banks. DeHaas and van Lelyveld
(2002)[46] come up with the same results for Eastern Europe for large foreign
banks compared to relatively smaller national banks. For the latter it is quite
unclear whether this in an e?ect of bank scale or rather (geographic) scope.
Houston et al.(1997)[79] for the U.S. ?nd that loan growth of banks in U.S.,
who are part of a Bank Holding Company (BHC) was less depending on
those banks own than on the holding’s ?nancial health, cementing the idea
that these holdings serve as internal capital markets.
The e?ect of an increase in organizational complexity on small business
credit is theoretically analyzed in my model. Numerous studies investigate
this relation also looking on the di?erence between pure national and multi-
national banks.
For example out-of-state ownership of a bank in the U.S. predominantly
had a negative e?ect on small business credit in the respective banks (e.g.
Berger et al.(1998)[16]. Being part of a bank holding had an e?ect in the
same direction (e.g. Berger and Udell (1996)[18].
In order to understand the e?ect of consolidation on the market level, one
has to analyze how other banks react to M&As in their market in respect to
small business lending.
The empirical literature gives mixed results. Berger et al.(1998)[16] point
to indication, that other banks more than o?set the negative e?ect on small
business lending in a consolidated institution by expanding their business in
this ?eld. However, contrary to this point, Berger et al.(1999)[11] ?nd a ne-
glectable external e?ect of M&A on other banks’ small business lending, with
only mature small banks being a?ected positively in their lending activity in
this segment.
Furthermore, Berger et al.(1999)[11] discover a positive e?ect of consoli-
dation on market entry in the banking industry. Combined with the generally
found empirical results, that de novo banks tend to have the highest share
of assets invested in small business loans
36
, this induced (additional) entry
might help o?set bank-level negative e?ects of consolidation on the credit
availability for small and medium sized enterprises.
In a recent contribution, Bonaccorsi di Patti and Gobbi (2007)[52] study
the e?ect of bank mergers and acquisitions on credit supply to Italian ?rms,
36
This has been found by several studies e.g. Goldberg and White (1998)[91], Berger,
Bonime, Goldberg and White (1999)[11] and DeYoung, Goldberg and White (1999)[115].
148
with the ?rm dataset including mainly small and medium sized enterprises.
Supporting the above model’s propositions the authors ?nd, that these ?rms
are at least in the short-run negatively a?ected by such change in the market.
To be precise, such reorganization within the banking sector leads to a drop
in credit of 9%. However, not captured by our model, they ?nd that credit
levels for these ?rms revert to old volume after three years.
5.6.3 Bank sector consolidation and relationship man-
agement strategies of small businesses
One indirect result of the model is, that for the single small company the
threat of bank sector consolidation, and therefore the fear of loosing its credit
partner, should give an incentive to have lending relationships with multiple
banks. Even though such multi-sourcing is theoretically expensive (even
more so than for transparent ?rm) as soft information has to be transferred
a number of times
37
, the empirical literature does support the idea that ?rms
use multiple banks as sources of ?nance (e.g. Detragiache et al.(2000)[49].
Berger et al.(2001)[14] argue more or less exactly to my point, that ”‘in-
formationally opaque ?rms are more likely to have multiple lenders......This
is because after being cut o? by the primary bank, opaque ?rms are likely
to encounter more di?culty in ?nding additional lenders and/or have less
favourable loan terms until their new relationships mature.”’
This is very close to the theoretical story suggested above, as the major
problem of small businesses in ?nding new credit partners is here, that the
former do not have an existing relationship with the latter.
5.7 Findings and Shortcomings
In this paper I have developed a model taking up concerns about the e?ect
of consolidation in the banking sector through M&A on credit availability
for small ?rms, combining the notion that banks di?er in their lending strat-
egy because of di?erent organizational setup with the idea that relationship
lending plays a large role in small business lending. I come up with the result
that consolidation via M&A indeed reduces the availability of small business
37
Costs could come in the form of simple transaction costs, duplicated e?ort and free-
rider problems. Also, ?rms might be reluctant to share con?dential information to multiple
banks who also have relationships with their competitors (e.g. Chiesa et al.(1995)[21].
149
lending in the economy. A reasoning on why ?rms might want to keep up
lending relationships with several banks can be derived in consequence.
Unfortunately a discussion on welfare e?ects of such changes can not
be made, as the loss in potential loan provision for SMEs is the gain in
potential loan provision for large informationally non-opaque ?rms in this
model. Welfare discussion would therefore need a further assumption on
which type of ?rm should be the favoured main recipient of bank ?nance. As
it is often argued, that small, new ?rms are the backbone of technological
development, whereas large ?rms tend to operate in saturated markets, it
might be true form an economy-perspective that loan provision to SMEs
should be the priority. This would then tentatively suggest that welfare is
reduced by such a shift in loan supply from small to large ?rms.
The model probably gives a good ?rst formal market-level insight into
the topic, but might fail to fully explain the reality because of a few short-
comings
38
.
One might be, that consolidated banks might often organize in a decen-
tralized way, for example by having each unit perform like a ?nancial center,
which would not lead to the evolution of an internal capital market. This
in turn would keep the manager’s specialisation decision in the old incen-
tive surroundings of a small bank. Also banks might organize in a way to
completely disentangle large and small ?rm credits personnel-wise.
So it looks like their might be scope for organizational strategy reducing
the hazardous e?ect of consolidation on small ?rm lending. It seems like
the main point of bank organization in the view of small ?rms would be
decentralized ?nancing negating the e?ect of the existence of internal capital
markets on lending strategy.
Also, I implicitly employ the idea, that soft, relationship information is
rather bank-speci?c than manager-speci?c. If, realistically the information
is with the manager, the manager of a merged bank might simply leave the
consolidated bank taking his relationships with him and starting up a new
small bank. Whether they do so would then depend on their expected utility
in the new bank compared to the consolidated institution.
Sticking with the manager’s objectives, one though vague argument might
also be a misspeci?cation of manager’s preferences. Managers might have
38
Of course, the result that there simply will be no more provision of loans to some
SMEs in some circumstances has to be seen as rather qualitative. Realistically small ?rms
might not be completely excluded from loans, but rather would have to pay higher interest
rates than without consolidation. Still the result that consolidation in the way discussed
in my model is harmful to SMEs would hold.
150
a really strong preference to keep up existing relationships due to person-
nel friendships with small ?rm clients’ management, social standing in the
community, etc. o?setting potential empire-building tendencies. Changing
preferences that way would lead to the conclusion that small business lend-
ing might actually be of larger volume than e?cient, as even non-performing
loans in expectations could be handed out due to these kind of preferences.
Another point to notice is that I implicitly assume that both types of
banks, small and large, are ex ante capable of supplying both loans to large
?rms and small ?rms. However due to risk-regulation and the possibly dif-
fering nature of bank relationships with small and large ?rms this might not
be the case
39
.
For all of the above reasoning, a large variety of possible avenues for fur-
ther theoretical research exists. Also, discussing entry motives of multina-
tional banks or motives for intranational M&A and how they are intertwined
with consequences of such actions should be very interesting, because motives
should shape strategies which should determine the e?ects of such actions.
Concerning empirics, the literature on the e?ects of in-market consolidation
or foreign bank entry via M&A on SME loan supply still needs further work,
as results seem ambiguous on quite a lot of questions about the topic. One
important point is an international comparison of e?ects of M&A on small
business lending. At the moment most of the research is based on U.S. data,
so there is not much leeway to understand whether and how the underlying
processes di?er between countries.
One ?nal important line of research to us would also be how the Basel II
rules change the picture on small ?rm lending. On the one hand, Basel II is
not in accordance with relationship lending threatening small ?rms relying
on this type of credit. On the other hand with the clear knowledge about
information requirements through the communication of Basel II small ?rms
will be pushed to increase their level of hard information, which should enable
them to be more competitive with large ?rms in the market for loans.
Within this thesis one ?nal point that can be made is that, stemming
from the above analysis, one would expect foreign bank entry via Green?eld
Investment to be the optimal entry mode from the perspective of small and
medium-sized enterprises, because this mode of entry, contrary to entry via
acquisition, does not directly lead to one small bank in the market becoming
integrated in a large bank structure, therefore the negative e?ect of such ?rms
39
For example a large ?rm might need multiple services from one bank such as wholesale
banking. Small banks might simply not have the scale to compete in these lines of business
thereby loosing businesses like credit procurement from large ?rms as well.
151
set free by the new organizational structure and not ?nding a new bank for
loan provision, does not apply to entry via Green?eld Investment.
152
APPENDIX
Appendix 1: Proof of Lemma 1
The expected net payo? of investor ?nancing a small bank with K units
of capital is respectively
• EP(1) =
g(1)
2
+
b(1)
2
• EP(2) = g(1) +b(1)
• EP(3) =
3g(1)
2
+
b(1)
2
+
b(2)
2
Overall, a ?nancing volume K is feasible if EP(K) > 0 (outside option)
For g(1)+b(1) > 0, which is ful?lled by assumption 4, banks are therefore
able to obtain two units of capital.
For b(2) < 2b(1) expected net payo? of investing three units of capital is
negative:
To show:
3g(1)
2
+
b(1)
2
+
b(2)
2
< 0 ?b(2) < ?b(1) ?3g(1)
for g(1) +b(1) > 0 follows
?b(1) ?3g(1) < 2b(1)
by assumption 3 b(2) < 2b(1) therefore
b(2) < ?b(1) ?3g(1) ?EP(3) < 0
Appendix 2: Proof of Proposition 1
The small bank loan manager will strongly prefer specialisation on small
?rm loans over specialisation on large ?rm loans i?
EU(S) > EU(L)
or
153
?
1/t
×
_
3g(1)
2
+
b(1)
2
_
+ (1 ??
1/t
) ×[g(1) +b(1)] + 2
>
? ×
_
3g(1)
2
+
b(1)
2
_
+ (1 ??) ×[g(1) +b(1)] + 2
Rearranging yields
?
1/t
> ?
and, using the log
t >
ln?
ln?
Therefore the minimum relationship length inducing the manager to again
lend to small ?rms in the period of interest is
t >
ln?
ln?
Appendix 3: Proof of di?erence in SME loan supply small
banks/large banks
I want to show t < t
1
Inserting yields
t =
ln ?
ln ?
< t
1
=
ln ?
ln
[
?(1+
1+g(1)
g(1)?b(1)
)
]
which can be simpli?ed to
1+g(1)
g(1)?b(1)
> 0
which is ful?lled as by the assumption necessary for bank operations to
be funded g(1) ?b(1) > 0/g(1) > 0.
Appendix 4: Third banks behaviour formerly specialised on
large ?rms
I proof by contradiction that banks formerly specialised on large ?rms
will not supply a positive expected loan volume to the analyzed ?rm set free
by a consolidated institution.
De?ne ?
?1
as the hard information research e?ciency in the former period
and ?
0
as the parameter in the period of analysis, with by assumption ?
?1
<
?
0
.
A bank will only supply positive expected loan volume to a ?rm with it
has relationship lengths t = 1 in the period of interest i?
154
? > ?
0
The bank specialises on large ?rms in the earlier period i?
? < ?
?1
So a bank specialised on large ?rms will provide positive expected loan
volume to the small ?rm in the analyzed period i?
?
0
< ? < ?
?1
which is a contradiction to
?
0
> ?
?1
155
A
p
p
e
n
d
i
x
A
:
C
E
O
p
a
y
o
?
s
g
i
v
e
n
i
n
f
o
r
m
a
t
i
o
n
r
e
c
e
i
v
e
d
a
n
d
s
u
b
s
e
q
u
e
n
t
c
a
p
i
t
a
l
a
l
l
o
c
a
t
i
o
n
I
—
A
(
4
;
0
)
(
3
;
1
)
(
2
;
2
)
(
1
;
3
)
(
0
;
4
)
{
G
G
}
{
G
G
}
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
{
G
G
}
{
n
o
n
e
}
4
g
(
1
)
7
g
(
1
)
2
+
b
(
1
)
2
3
g
(
1
)
+
b
(
1
)
b
(
2
)
2
+
5
g
(
1
)
2
+
b
(
1
)
2
2
g
(
1
)
+
b
(
2
)
{
G
G
}
{
G
B
}
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
3
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
b
(
2
)
{
G
G
}
{
B
B
}
4
g
(
1
)
3
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
2
b
(
1
)
b
(
2
)
+
g
(
1
)
+
b
(
1
)
2
b
(
2
)
{
G
B
}
{
G
G
}
2
g
(
1
)
+
b
(
2
)
3
g
(
1
)
+
b
(
1
)
4
g
(
1
)
4
g
(
1
)
4
g
(
1
)
{
G
B
}
{
n
o
n
e
}
2
g
(
1
)
+
b
(
2
)
5
g
(
1
)
2
+
3
b
(
1
)
2
3
g
(
1
)
+
b
(
1
)
b
(
2
)
2
+
5
g
(
1
)
2
+
b
(
1
)
2
2
g
(
1
)
+
b
(
2
)
{
G
B
}
{
G
B
}
2
g
(
1
)
+
b
(
2
)
3
g
(
1
)
+
b
(
1
)
4
g
(
1
)
3
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
b
(
2
)
{
G
B
}
{
B
B
}
2
g
(
1
)
+
b
(
2
)
2
g
(
1
)
+
2
b
(
1
)
2
g
(
1
)
+
2
b
(
1
)
g
(
1
)
+
b
(
1
)
+
b
(
2
)
2
b
(
2
)
{
B
B
}
{
G
G
}
2
b
(
2
)
b
(
2
)
+
b
(
1
)
+
g
(
1
)
2
g
(
1
)
+
2
b
(
1
)
3
g
(
1
)
+
b
(
1
)
4
g
(
1
)
{
B
B
}
{
n
o
n
e
}
2
b
(
2
)
b
(
2
)
+
3
b
(
1
)
2
+
g
(
1
)
2
3
b
(
1
)
+
g
(
1
)
b
(
2
)
2
+
3
g
(
1
)
2
+
3
b
(
1
)
2
2
g
(
1
)
+
b
(
2
)
{
B
B
}
{
G
B
}
2
b
(
2
)
b
(
2
)
+
b
(
1
)
+
g
(
1
)
2
g
(
1
)
+
2
b
(
1
)
2
g
(
1
)
+
2
b
(
1
)
2
g
(
1
)
+
b
(
2
)
{
B
B
}
{
B
B
}
2
b
(
2
)
b
(
2
)
+
2
b
(
1
)
4
b
(
1
)
b
(
2
)
+
2
b
(
1
)
2
b
(
2
)
{
n
o
n
e
}
{
G
G
}
2
g
(
1
)
+
b
(
2
)
b
(
2
)
2
+
5
g
(
1
)
2
+
b
(
1
)
2
3
g
(
1
)
+
b
(
1
)
7
g
(
1
)
2
+
b
(
1
)
2
4
g
(
1
)
{
n
o
n
e
}
{
n
o
n
e
}
2
g
(
1
)
+
b
(
2
)
b
(
2
)
2
+
2
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
2
b
(
1
)
b
(
2
)
2
+
2
g
(
1
)
+
b
(
1
)
2
g
(
1
)
+
b
(
2
)
{
n
o
n
e
}
{
G
B
}
2
g
(
1
)
+
b
(
2
)
b
(
2
)
2
+
5
g
(
1
)
2
+
b
(
1
)
2
3
g
(
1
)
+
b
(
1
)
5
g
(
1
)
2
+
3
b
(
1
)
2
2
g
(
1
)
+
b
(
2
)
{
n
o
n
e
}
{
B
B
}
2
g
(
1
)
+
b
(
2
)
b
(
2
)
2
+
3
g
(
1
)
2
+
3
b
(
1
)
2
g
(
1
)
+
3
b
(
1
)
b
(
2
)
+
g
(
1
)
2
+
3
b
(
1
)
2
2
b
(
2
)
156
Chapter 6
Conclusion and Outlook
The ongoing globalization of the banking industry has raised a number of
questions on the entry motives and strategies of banks going abroad, as well
as the impact of the change of market structures in the banking industry,
that has accompanied this process.
This thesis has started out with giving an overview about what existing
economic literature has to say about a large array of questions arising in
the discussion about multinational banks and then has o?ered an insightful
practitioners’ view on strategic decisions within multinational banks via the
case study of Bank Austria Creditanstalt in Central and Eastern Europe.
In the following this body of work has focused on, directly and indirectly,
shedding new light on three speci?c issues that have arisen in the analysis of
multinational bank behaviour and the consequences of associated changes in
banking markets’ structure.
Chapter 2, after giving an overview about the existing literature on multi-
national banking, yielded the following main insights via the case study con-
ducted on the Austrian market leader in one of the core regions of multina-
tional bank expansion world-wide, Central and Eastern Europe. In general,
a highly complex relationship between a bank’s core business strategy and its
home market, its mode and timing for entry into foreign markets, its strat-
egy in building business in these markets as well as the geographic pattern
of a banks geographic expansion, was found. Retail-oriented and wholesale-
oriented banks seem to di?er signi?cantly in their timing and choice of entry
mode into foreign markets, with wholesale-oriented institutions seeming far
more inclined to grow foreign business organically via Green?eld Investment,
whereas successful retail banking in foreign markets seems to require non-
mobile assets only available via the acquisition of local incumbents. Addi-
157
tionally, having a su?ciently large commercial client customer base to follow
abroad seems to enable wholesale-oriented banks to enter markets at earlier
stages of economic development, as entry into such a market can be pro?table
almost from the beginning with such sources of revenue at hand. Concerning
empirical identi?cation strategies of factors underlying the location decision
of multinational banks, it is concluded from the ?ndings, that, at least for
regions economically integrating and with similar law and bank regulation,
focusing on single host country characteristics might be ine?cient, as location
strategies might a broader regional, not country-speci?c, type
1
.
Chapter 3 theoretically discussed the choice of entry mode of multina-
tional institutions into foreign markets in general, not restricted to the bank-
ing industry. Whereas the existing literature discusses the choice of com-
panies whether to enter a foreign market via the setup of a completely new
structure (Green?eld Investment) or via the acquisition of an incumbent
company in the host market in a completely static way, my proposed model
incorporates a more ”‘dynamic”’ view. In a market that is entered sequen-
tially by foreign companies the choice of entry mode of the early mover
a?ects the entry decision of potential subsequent entrants, where early entry
via Green?eld Investment has the strategic advantage compared to entry via
acquisition, that pro?t-reducing sequential entry via Green?eld Investment
becomes less likely. The basic model thus yields a reasoning, why Green?eld
Investment makes up such a substantial share of foreign direct investment
in general, especially concerning the number of occurrences, whereas recent
theoretical literature has focused on adding further explanations for entry
via M&A, especially focusing on the analysis of asset-complementarity of ac-
quirer and target. It is also found, that a perverse e?ect of limited takeover
possibilities in markets to be entered further increases the attractiveness of
Green?eld Investment for early market entrants. Whereas such a basic model
can explain why Green?eld Investment should be an important entry path
for e.g. the wholesale banking industry, an extension is additionally pro-
posed that yields the tentative prediction, that entry via M&A should be
a favoured entry mode into foreign markets in the retail banking industry.
Incorporating the notion of market-speci?c learning by doing e?ects we ?nd
M&A to be the early entry mode deterring harmful sequential entry, if the
respective industry is characterized by a low degree product di?erentiation
between ?rms in the market. Due to generally assumed low ex ante product
di?erentiation and heterogeneous regulation across countries, we deem this
1
A variety of further results, new or con?rming existing literature, can be found in the
conclusion of the case study.
158
extension to be a good ?t for the analysis of the retail banking industry.
The theory proposed in Chapter 4 o?ered new insights into a type of
banking foreign direct investment, that has not received su?cient formal
treatment in the existing literature though being heavily discussed both in
general business literature as well as empirically, namely banks going abroad
to provide services locally to its existing multinational ?rm customer base.
In my model, which bases on a recent theoretical contribution by Marin
and Schnitzer (2006)[97], a bank’s decision whether to follow its customer
abroad or not for loan provision is discussed. The model yields that this
location decision of banks is shaped by an interaction of client-speci?c and
host country-speci?c characteristics. Tendentially
2
banks will follow their
customers abroad if their multinational customer enters what might be called
a non-fully developed country characterized by a low endowment in human
capital. The argument runs along the line of collateral, where it is derived,
that governments’ of such countries might be more inclined to block out?ow
of valuable asset-embedded human capital, creating the need for banks to
sell such assets of a non-performing loan project locally in the host country.
Doing so e?ciently however requires a physical presence abroad to be able
to identify potential local asset-takers. This line of argument is close to
concerns of practitioners that mentioned, that issues arising about the cross-
border liquidation of collateral are an important reason for banks to follow
customers abroad
3
.
Chapter 5 again did not primarily focus on multinational banking per se,
but uses a theoretical model to analyze the e?ect of general re-organization of
the banking industry on loan supply for small and medium-sized enterprises
characterized by their informational opacity. To be precise the e?ect of M&A
in the banking sector on the former is analyzed theoretically, where the gen-
eral process applies to in-market consolidation as well as foreign bank entry
via M&A. It is shown, that the likelihood of loan provision to small ?rms is
reduced by M&A activity in the banking sector due to two modelled reasons.
For one, a small ?rm is likely to loose its loan relationship with its small
bank when this bank integrates into a large bank structure via either being
a target or an acquirer in an M&A deal, due to loan managers incentives
for specialising on large ?rm loans being larger in large, hierarchical banking
institutions. Incorporating the notion of relationship-based banking for the
SME sector, such that soft, non-veri?able, information becomes better the
2
As the following argument holds true for two out of three types of clients discussed.
3
I’d like to thank Jana Schwarze (Commercial Clients, Stadtsparkasse Muenchen) and
Christoph Schropp (formerly Dresdner Bank AG) for pointing out this issue to me.
159
longer the respective bank-client relationship, I ?nd that these ?rms set free
by the consolidated institution are unlikely to attain loans from other small
banks in the market due to a missing relationship with these institutions.
Concerning the general focus of this thesis, it can tentatively be concluded,
that foreign bank entry via Green?eld Investment should not have an adverse
e?ect on the availability of loan ?nancing for SMEs, whereas entry via M&A
should be harmful to this client group.
Numerous unanswered questions concerning future developments in the
multinationalization of the banking industry persist.
For one, the basic business strategies of banks are undergoing a major
change. Banks increasingly focus on o?-balance sheet activity, abandoning
their role as risk-takers in their ?nancing business by passing on securitized
debt to other ?nancial market participants such as hedge funds. Increas-
ingly leaving such ?nancial services of the balance sheet banks do not act
as ?nancial intermediaries in the classic sense anymore, rather acting as in-
formation brokers and sales channels for other ?nancial market participants.
As a result of this development, banks are in less need of long-term re?nanc-
ing via deposits, shifting their retail strategies to act as a sales channel for
higher-margin investment products such as certi?cates and investment funds.
Instead of re?nancing via deposits banks therefore increasingly re?nance via
the capital market as well as more short-term the interbanking market. As
the classic intermediation role between depositor and borrower therefore be-
comes less important, and services such as asset management and investment
banking services such as bond underwriting seem to require su?cient scale,
an arising specialisation of banks on subsegments of the value chain is proba-
bly just in its beginning. Such specialisation is already found in Central and
Eastern Europe in a recent study by Dinger and von Hagen (2007)[55], who
?nd that old established banks in the region use their large existing branch
network to focus on raising deposits and subsequently transfer these assets to
new (mostly foreign) banks via the interbanking market, whereas the latter
new banks focus on the provision of loans, backed by re?nancing through
the interbanking market. How such changing business strategies of banks
will in?uence the further globalization of the banking industry remains to be
seen.
At the same time new political discussions about restricting entry (espe-
cially via the acquisition of incumbent national ?rms) in key industries have
gained momentum, with examples being the prevented takeover of Span-
ish energy national champion Endesa by German E.ON or recent discussion
about prohibiting majority stakes of foreign investment funds in core in-
dustries in Germany. With the banking industry still perceived to be of
160
strategic importance for the economy, further globalization of the industry
might be made infeasible by political interventionism
4
. Even without such
political barriers, according to the economic literature (e.g. see Berger et
al.(2000)[12]) the globalization of the banking industry might be restricted
to an incomplete level compare to other industries, as informal barriers to
entry in the commercial SME and retail banking sector coupled with a multi-
national banks’ potential partial inability to successfully integrate acquired
smaller local incumbents, might leave at least some banking services markets
in the hands of smaller local players.
Finally, the internationalisation of the banking industry might also come
under scrutiny by typical home countries of such multinational banks, as the
international exposure of home banks might be cause for concern of home
country governments. The latest development in the sub-prime mortgage
loan crisis in the U.S. has left numerous foreign ?nancial institutions
5
in
bad condition. As a result of this exposure of foreign banks to U.S. market
risks, indication has pointed to these institutions also restricting loan business
in their home countries. Such development might lead to political claims
towards domestic institutions to decrease their scope of business in (risky)
foreign markets.
4
Which, at least for the case of foreign bank entry via M&A, could possibly be ratio-
nalized by the ?ndings in Chapter 5 of this thesis.
5
Banks that have publicly been heavily discussed to be severely a?ected by this crisis
are British Northern Rock and German Banks IKB and Sachsen LB.
161
Bibliography
[1] Erste Bank AG. Corporate website.http://www.sparkasse.at/erstebank, 2007.
[2] Jalal Akhavein, Allen Berger, and David Humphrey. The e?ects of
bank megamergers on e?ciency and prices: Evidence from the pro?t
function. Review of Industrial Organization, Vol. 12:95–139, 1997.
[3] Robert Aliber. International banking: A survey. Journal of Money,
Credit and Banking, Vol. 16 (4):661–678, 1984.
[4] Thomas Andersson and Roger Svensson. Entry modes for direct in-
vestment determined by the composition of ?rm-speci?c skills. Scan-
dinavian Journal of Economics, 96(4):551–560, 1994.
[5] Elisa Bain, Justin Fung, and Ian Harper. Multinational banking: His-
torical, empirical and case perspectives. Melbourne Business School
Working Paper 99-05, 1999.
[6] Cli?ord Ball and Adrian Tschoegl. The decision to establish a foreign
bank branch or subsidiary: An application of binary classi?cation pro-
cedures. Journal of Financial and Qualitative Analysis, 17 (3):411–424,
1982.
[7] World Bank. Integration with the global economy. Technical report,
World Development Indicators, 2004.
[8] P.G. Baudino, G. Caviglia, E. Dorrucci, and G. Pineau. Financial FDI
to the eu accession countries. ECB Directorate General International
and European Relations, 19 March 2004, 2004.
[9] Paul Bauer, Allen Berger, and David Humphrey. E?ciency and produc-
tivity growth in u.s. banking. In Harold Fried, C.A. Knox Lovell, and
Shelton Schmidt, editors, The measurement of productive e?ciency:
162
Techniques and Applications, pages 386–413, Oxford, 1993. Oxford
University Press.
[10] William J. Baumol. Business behaviour, value, and growth. Mac Mil-
lan, New York, 1959.
[11] Allen Berger, Carl Bonime, Lawrence G. Goldberg, and Lawrence
White. The dynamics of market entry: The e?ects of mergers and
aquisitions on de novo entry and small business lending in the banking
industry. Technical report, Board of Governors of the Federal Reserve
System, 1999.
[12] Allen Berger, Robert DeYoung, Hesna Genay, and Gregory Udell.
Globalization of ?nancial institutions: Evidence from cross-border
banking performance. Brookings-Wharton Papers on Financial Ser-
vices, Vol. 3:23–158, 2000.
[13] Allen Berger and Timothy Hannan. Using measures of ?rm e?-
ciency to distinguish among alternative explanations of the structure-
performance relationship. Managerial Finance, 23:6–31, 1997.
[14] Allen Berger, Klapper, and Gregory Udell. The ability of banks to
lend to informationally opaque small businesses. Journal of Banking
and Finance, Vol. 25:2127 – 2167, 2001.
[15] Allen Berger, Nathan Miller, Michael Petersen, Raghuram Rajan, and
Jeremy Stein. Does function follow organizational form? evidence from
the lending practice of large and small banks. NBER working paper
No. 8752, February 2002.
[16] Allen Berger, Anthony Saunders, Joseph Scalise, and Gregory Udell.
The e?ects of bank mergers and aquisitions on small business lending.
Journal of Financial Economics, Vol. 50(2):187–229, 1998.
[17] Allen Berger and Gregory Udell. Relationship lending and lines of
credit in small ?rm ?nance. Journal of Business, 68:351–382, 1995.
[18] Allen Berger and Gregory Udell. Universal banking and the future
of small business lending. In Saunders and Walter, editors, Universal
Banking: Financial System Design Rediscovered, pages 558–627, Burr
Ridge, IL, 1996. Irwin Publishing.
[19] Allen Berger and Gregory Udell. Small business credit availability and
relationship lending: The importance of bank organisational structure.
Economic Journal, 112 (477):32–53, 2002.
163
[20] D. Bernheim. Strategic entry deterrence of sequential entry into an
industry. Rand Journal of Economics, 15:1–11, 1984.
[21] S. Bhattacharya and G. Chiesa. Proprietary information, ?nancial in-
termediation and research incentives. Journal of Financial Intermedi-
ation, 4:328–357, 1995.
[22] Jacob Bikker and Johannes Groeneveld. Competition and concentra-
tion in the eu banking sector. De Nederlandsche Bank, 1998.
[23] BIS. Group of ten report on consolidation in the ?nancial sector. Tech-
nical report, Bank for International Settlements, 2001.
[24] Hanneke Bol, Jakob de Haan, Bert Scholtens, and Ralph de Haas.
How important are foreign banks in european transition countries? a
comparative analysis. EconWPA International Finance Series, No.
0209005, 2002.
[25] J.P. Bonin, I. Hasan, and P. Wachtel. Bank performance, e?ciency and
ownership in transition countries. Journal of Banking and Finance,
29:31–53, 2005.
[26] R.A. Brealey and E.C. Kaplanis. The determination of foreign banking
location. Journal of International Money and Finance, Vol. 15 (4):577–
597, 1996.
[27] Claudia Buch. Why do banks go abroad? - evidence from german data.
Kiel Working Papers, No.948, 1999.
[28] Claudia Buch and Gayle de Long. Cross-border bank mergers: What
lures the rare animal? Journal of Banking and Finance, 28:2077–2102,
2004.
[29] Claudia Buch and Alexander Lipponer. FDI versus cross-border ?nan-
cial services: The globalization of german banks. Deutsche Bundesbank
Discussion Paper, No 05/2004, 2004.
[30] G.A. Calvo, L. Leiderman, and C.M. Reinhart. Capital in?ows and
real exchange rates appreciation in latin america - the role of external
factors. IMF Sta? Papers, 40 (1):108–151, 1993.
[31] Mark Casson. Evolution of multinational banks: A theoretical perspec-
tive. In Geo?rey Jones, editor, Banks as Multinationals, pages 14–29,
London, 1990. Routledge.
164
[32] Vittoria Cerasi, Barbara Chizzolini, and Marc Ivaldi. Sunk costs and
competitiveness of european banks after deregulation. LSE Financial
Market Group Discussion Paper 290, 1998.
[33] Judith Chevalier and Christopher Avery. Why do managers undertake
aquisitions? an analysis of internal and external rewards for aquisive-
ness. Journal of Law, Economics and Organization, 14:24–43, 1998.
[34] S. Claessens, A. Demirguec-Kunt, and H. Huizinga. How does foreign
entry a?ect the domestic banking market. Journal of Banking and
Finance, 25:891–911, 2001.
[35] Sophie Claeys and Christa Hainz. Acquisition versus green?eld: The
impact of the mode of foreign bank entry on information and bank
lending rates. ECB Working Paper, No. 653, 2006.
[36] Je?rey Clark. Economic cost, scale e?ciency and competitive viability
in banking. Journal of Money, Credit and Banking, 28:342–364, 1996.
[37] George Clarke, Robert Cull, and Maria Martinez-Peria. Does foreign
bank penetration reduce access to credit in developing countries: Evi-
dence from asking borrowers. World Bank Group Working Paper 2716,
2001.
[38] Rebel A. Cole, Lawrence G. Goldberg, and Lawrence J. Wright.
Cookie-cutter versus character: The micro structure of small business
lending by large and small banks. In Richard W. Lang, editor, Business
Access to Capital and Credit. Federal Reserve Bank of Dallas, 1999.
[39] Bank Austria Creditanstalt. Annual report 2006.
[40] Bank Austria Creditanstalt. 2007.
[41] J.S. Crystal, B.G. Dages, and L. Goldberg. Has foreign bank entry led
to sounder banks in latin america? Current Issues in Economics and
Finance, 8 (1):1–6, 2002.
[42] Elisa Curry, Justin Fung, and Ian Harper. Multinational banking:
Historical, empirical and case perspectives. In A.W. Mullineux and
V. Murinde, editors, Handbook of International Banking, pages 27–59,
UK and USA, 2003. Edward Elgar.
165
[43] B.G. Dages, L. Goldberg, , and D. Kinney. Foreign and domestic
bank participation in emerging markets: Lesson from mexico and ar-
gentina. Federal Reserve Bank of New York Economic Policy Review,
September:17–36, 2000.
[44] Ralph de Haas and Iman van Lelyveld. Foreign banks and credit sta-
bility in central and eastern europe. a panel data analysis. Journal of
Banking and Finance, 30 Issue 7:1927–1952, 2006.
[45] Ralph de Haas and Iman van Lelyveld. Internal capital markets and
lending by multinational bank subsidiaries. De Nederlandsche Bank
Working Paper, 101/June, 2006.
[46] R.T.A. deHaas and I.P.P. vanLelyveld. Foreign bank penetration and
private sector credit in central and eastern europe. Technical report,
De Nederlandsche Bank Sta? Reports No.91, 2002.
[47] Giovanni Dell’Ariccia, Ezra Friedman, and Robert Marquez. Adverse
selection as a barrier to entry in the banking industry. RAND Journal
of Economics, 30 (3):515–534, 1999.
[48] Giovanni Dell’Ariccia and Robert Marquez. Information and bank
credit allocation. Journal of Financial Economics, 72 (1):185–214,
2004.
[49] E. Detragiache, P. Garella, and L. Guiso. Multiple versus single bank-
ing relationships: Theory and evidence. Journal of Finance, 55:1133–
1161, 2000.
[50] Robert DeYoung. Bank mergers, x-e?ciency, and the market for cor-
porate control. Managerial Finance, 23:32–47, 1997.
[51] Robert deYoung and Daniel Nolle. Foreign-owned banks in the us:
Earning market share or buying it? Journal of Money, Credit and
Banking, 28:622–636, 1996.
[52] Emilia Bonaccorsi di Patti and Giorgio Gobbi. Winners or losers? the
e?ects of banking consolidation on corporate borrowers. Journal of
Finance, LXII (2):669–695, 2007.
[53] Douglas W. Diamond. Financial intermediation and delegated moni-
toring. Review of Economic Studies, Vol. 51, No. 3:393 – 414, 1984.
166
[54] Douglas W. Diamond. Monitoring and reputation: The choice between
bank loans and directly placed debt. Journal of Political Economy, 99
(4):689–721, 1991.
[55] Valeriya Dinger and Juergen von Hagen. Risk e?ects of bank special-
ization in central and eastern europe. European Review of Finance and
Economics, forthcoming, 2007.
[56] Marc Melitz Elhanan Helpman and Stephen Yeaple. Export versus
FDI with heterogeneous ?rms. American Economic Review, 94:300–
316, 2004.
[57] Liz Feely and Dave Compare. When worlds collide: culture clash.
Journal of Business Strategy, 24 No.4:9–13, 2003.
[58] Norman Fieleke. The growth of u.s. banking abroad: An analytical
survey. Federal Reserve Bank of Boston Conference Series, October
18:9–40, 1977.
[59] A. Fisher and P. Molyneux. A note on the determinants of foreign
bank activity in london between 1980 and 1989. Applied Financial
Economics, 6:271–277, 1996.
[60] D. Focarelli and A.F. Pozzolo. The determinants of cross-border bank
share holdings: an analysis with bank-level data from oecd countries.
Banca D’Italia Temi di Discussione, 381 - October 2000, 2000.
[61] D. Focarelli and A.F. Pozzolo. The patterns of cross-border bank merg-
ers and shareholding in oecd countries. Journal of Banking and Fi-
nance, 25:2305–2337, 2001.
[62] D. Focarelli and A.F. Pozzolo. Where do banks expand abroad? an
empirical analysis. Universita degli Studi del Molise Economics and
Statistics Discussion Paper, No. 9/03, 2003.
[63] Bank for International Settlements Committee on the Global Finan-
cial System. Foreign direct investment in the ?nancial sector of emerg-
ing market economies. Technical report, BIS, March 2004.
[64] Kenneth French and James Poterba. Investor diversi?cation and in-
ternational equity markets. American Economic Review, 81:222–226,
1991.
167
[65] Robert H. Gertner, David S. Scharfstein, and Jeremy C. Stein. Inter-
nal versus external capital markets. Quarterly Journal of Economics,
109:1211–1230, 1994.
[66] Mariassunta Giannetti and Steven Ongena. Financial integration and
entrepeneurial activity: Evidence from foreign bank entry in emerging
markets. Stockholm School of Economics, mimeo, 2005.
[67] R. Gilbert and X. Vives. Entry deterrence and the free rider problem.
Review of Economic Studies, 53:71–83, 1986.
[68] Lawrence Goldberg and Robert Grosse. Location choice of foreign
banks in the united states. Journal of Economics and Business, 46:367–
379, 1994.
[69] Linda Goldberg. When is u.s. bank lending to emerging markets
volatile? NBER Working paper, 8209, 2001.
[70] Holger G¨org. Analysing foreign market entry: The choice between
green?eld investment and acquisitions. Journal of Economics Studies,
27(3):165–181, 2000.
[71] Christopher Green, Victor Murinde, and Ivaylo Nikolov. The e?ciency
of foreign banks in central and eastern europe: Evidence on economies
of scale and scope. Journal of Emerging Markets Finance, 3 No. 2:175–
205, 2004.
[72] Stuart Greenbaum, George Kanatas, and Itzhak Venezia. Equilibrium
loan pricing under the bank client relationship. Journal of Banking
and Finance, 13:221–235, 1989.
[73] Herbert Grubel. A theory of multinational banking. Banca Nazionale
del Lavoro Quarterly Review, 123:349–363, 1977.
[74] Mauro Guillen and Adrian Tschoegl. At last the internationalization of
retail banking? the case of the spanish banks in latin america. Wharton
School Center for Financial Institutions, University of Pennsylvania,
Working paper 99-41, 1999.
[75] Diana Hancock and James A. Wilcox. The ”credit crunch” and the
availability of credit to small business. Journal of Banking and Finance,
22 (6):983–1014, 1998.
168
[76] Olena Havrylchyk and Emilia Jurzyk. Pro?tability of foreign banks in
central and eastern europe: Does the mode of entry matter? LICOS
Discussion Paper, 16606, 2006.
[77] Jean-Francois Hennart and Young-Ryeol Park. Green?eld vs. acquisi-
tion: The strategy of japanese investors in the united states. Manage-
ment Science, 39 No. 9:1054–1070, 1993.
[78] Bengt Holmstroem and Jean Tirole. Financial intermediation, loanable
funds, and the real sector. Quarterly Journal of Economics, CXII:663–
691, 1997.
[79] Joel Houston, Christopher James, and David Marcus. Do bank internal
capital markets promote lending? Journal of Banking and Finance, 46
(2):899–918, 1997.
[80] Thomas F. Huertas. US multinational banking: History and prospects.
In Geo?rey Jones, editor, Banks as Multinationals, pages 248–267, Lon-
don, 1990. Routledge.
[81] Joseph Hughes and Loretta Mester. Bank capitalization and cost: Ev-
idence of scale economies in risk management and signalling. Review
of Economics and Statistics, 80 (2):314–325, 1998.
[82] W. Hultman and Randolph McGee. Factors a?ecting the foreign bank-
ing presence in the u.s. Journal of Banking and Finance, 13:383–396,
1989.
[83] Wiliam Jackson. Market structure and the speed of price adjustments:
Evidence of non-monotonicity. Review of Industrial Organization, 12
(1):37–57, 1997.
[84] Michael Jensen and Kevin Murphy. Performance pay and top-
management incentives. Journal of Political Economy, 98:225–264,
1990.
[85] Michael C. Jensen. Agency costs of free cash ?ow, corporate ?nance,
and takeovers. American Economic Review, 76(2):323–329, 1986.
[86] Fatih Kasirye Joe Peek, Eric S. Rosengren. The poor performance of
foreign bank subsidies: Were the problems aquired or created? Journal
of Banking and Finance, 23 (2-4):579–604, 1999.
[87] Geo?rey Jones. Banks as multinationals. In Geo?rey Jones, editor,
Banks as Multinationals, pages 1–13, London, 1990. Routledge.
169
[88] Michael Klein, Joe Peek, and Eric Rosengren. Troubled banks, im-
paired foreign direct investment: The role of relative access to credit.
American Economic Review, 92 (3):664–682, 2002.
[89] Paul Krugman. The narrow moving band, the dutch disease and the
competitive consequences of mrs. thatcher: Notes on trade in the pres-
ence of dynamic scale economies. Journal of Development Economics,
27:41–55, 1987.
[90] Myron Kwast, Martha Starr-McCluer, and John Wolken. Market def-
inition and the analysis of antitrust in banking. Antitrust Bulletin, 42
(4):973–995, 1997.
[91] Lawrence J. White Lawrence G. Goldberg. De novo banks and lending
to small businesses. Journal of Banking and Finance, 22 (6-8):851–867,
1998.
[92] Maria Lehner. Entry mode choice of multinational banks. Munich
Economics Discussion Paper, 2007 forthcoming, 2007.
[93] Maria Lehner and Monika Schnitzer. Entry of foreign banks and their
impact on host countries. Munich Economics Discussion Paper, No.
152, 2006.
[94] Ross Levine. Denying foreign bank entry: Implications for bank interest
margins. Working Papers Central Bank of Chile, 222, 2003.
[95] Andzelika Lorentowicz, Dalia Marin, and Alexander Raubold. Owner-
ship, capital or outsourcing: What drive german investment to eastern
europe. Munich Economics Discussion Paper, 2002-03, 2002.
[96] Giovanni Majnoni, Rashmi Shankar, and Eva Varhegyi. The dynamics
of foreign bank ownership - evidence from hungary. The World Bank
Policy Research Working Paper, 3114, 2003.
[97] Dalia Marin and Monika Schnitzer. When is FDI a capital ?ow. Munich
Economics Discussion Paper, 2006-22, 2006.
[98] James Markusen and Anthony Venables. The theory of endowment,
intra-industry and multinational trade. Journal of International Eco-
nomics, 52 (2):209–234, 2000.
[99] Marc Melitz. The impact of trade on intra-industry allocations and
aggregate industry productivity. Econometrica, 71:1695–1725, 2003.
170
[100] Atif Mian. Distance constraints: The limits of foreign lending in poor
economies. Journal of Finance, 61 (3):1465–1505, 2006.
[101] Philip Molyneux, D.M. Lloyd-Williams, and John Thornton. Compet-
itive conditions in european banking. Journal of Banking and Finance,
18 (3):445–459, 1994.
[102] Randal Morck, Andrei Shleifer, and Robert Vishny. Do managerial
objectives drive bad aquisitions? Journal of Finance, 45:31–49, 1990.
[103] D. Morgan, B. Rime, and P.E. Strahan. Bank integration and state
business volatility. Quarterly Journal of Economics, 119 (4):1555–1585,
2004.
[104] F. Moshirian. International investment in ?nancial services. Journal
of Banking and Finance, 25:317–337, 2001.
[105] Thomas M¨ uller. Analyzing modes of foreign entry: Green?eld invest-
ment versus acquisition. Munich Economics Discussion Paper 2001-01,
January 2001, 2001.
[106] Douglas Nigh, Kang Rae Choh, and Suresh Krishnan. The role of
location-related factors in u.s. banking involvement abroad: An empir-
ical examination. Journal of International Business Studies, Fall:59–72,
1986.
[107] Volker Nocke and Stephen Yeaple. Cross-border mergers and acquisi-
tions versus green?eld foreign direct investment: The role of ?rm het-
erogeneity. Journal of International Economics, 72 (2):336–365, 2007.
[108] Daniel Nolle and Rama Seth. Do banks follow their customers abroad?
Federal Reserve Bank of New York, Research Paper 9620, 1996.
[109] Gerhard Smoley Head of Investor Relations Bank Austria Group. Tele-
phone interview 22.01.2007.
[110] U.S. General Accounting O?ce. Foreign banks: Assessing their role
in the u.s. banking system. Report to the Ranking Minority Mem-
ber; Committee on Banking, Housing and Urban A?airs; U.S. Senate,
GAO/GGD-96-26, 1996.
[111] Oliver Pastre. International bank-industry relations: An empirical as-
sessment. Journal of Banking and Finance, 5:65–76, 1981.
171
[112] J. Peek and E.S. Rosengren. Implications of the globalization of the
banking sector: the latin american experience. New England Economic
Review, September/October:45–63, 2000.
[113] Horst Ra?, Michael Ryan, and Frank Staehler. Asset ownership and
foreign-market entry. CESifo Working Paper No. 1676, February 2006,
2006.
[114] Raghuram Rajan. The entry of commercial banks into the securities
business: A selective survey of theories and evidence. In Anthony
Saunders and Ingo Walter, editors, Universal Banking: Financial Sys-
tem Design Reconsidered, pages 282–302. Irwin, 1996.
[115] Lawrence J. White Robert DeYoung, Lawrence G. Goldberg. Youth,
adolescence and maturity at banks: Credit availability to small busi-
nesses in an era of banking consolidation. Journal of Banking and
Finance, 23 (2-4):463–492, 1999.
[116] M. Sabi. An application of the theory of foreign direct investment
to multinational banking in LDCs. Journal of International Business
Studies, 19 (3):433–447, 1987.
[117] Paolo Sapienza. The e?ects of banking mergers on loan contracts.
Journal of Finance, 57 (1):329–367, 2002.
[118] D.T. Savage. Mergers, branch closings and cost savings. Board of
Governors of the Federal Reserve System, 1991.
[119] Jonathan Scott and William Dunkelberg. Bank consolidation and small
business lending: A small ?rm perspective. In Richard W. Lang, editor,
Business Access to Capital and Credit, pages 328–361. Federal Reserve
Bank of Dallas, 1999.
[120] Sherrill Sha?er. Can megamergers improve bank e?ciency? Journal
of Banking and Finance, 17:423–436, 1993.
[121] Farouk Soussa. A note on banking FDI in emerging markets: Literature
review and evidence from m&a data. BIS CGFS Publications, 22, 2004.
[122] Heinrich Stackelberg. Marktform und gleichgewicht. Julius Springer,
Vienna, 1934.
[123] Jeremy Stein. Information production and capital allocation: Decen-
tralized versus hierarchical ?rms. Journal of Finance, Vol. LVIII, No.
5:1891 – 1921, 2002.
172
[124] Joseph Stiglitz and Andrew Weiss. Credit rationing in markets with
imperfect information. American Economic Review, 71 (3):393–410,
1981.
[125] Jean Tirole. The theory of industrial organization. MIT Press, Cam-
bridge, MA, 1988.
[126] Adrian Tschoegl. Who owns the major us subsidiaries of foreign banks?
a note. Wharton Financial Institutions Center Discussion Paper, 03-
11, 2003.
[127] Various. Thinking big (survey: International banking). The Economist,
May 20th, 2006, 2006.
[128] Rudi Vander Vennet. The e?ect of mergers and aquisitions on the
e?ciency and pro?tability of ec credit institutions. Journal of Banking
and Finance, 20 (9):1531–1558, 1996.
[129] Rudi Vander Vennet. Causes and consequences of eu bank takeovers.
In Richard Portes Kees Koedijk, Marco Pagano, editor, The Changing
European Landscape, pages 45–61, Brussel, Belgium, 1998. Centre for
European Policy Research.
[130] Barry Williams. Positive theories of multinational banking: Eclectic
theory versus internalisation theory. Journal of Economic Surveys, 11
(1):71–100, 1997.
[131] Oliver Williamson. Economics of discretionary behaviour: Managerial
objectives in a theory of the ?rm. Kershaw Publishing, London, 1974.
[132] Andrew Winton. Don’t put all your eggs in one basket? diversi?cation
and specialisation in lending. University of Minnesota, 1999.
173
Curriculum Vitae Peter Beermann
Personal Data
Name Peter Beermann
Date of birth January 20th, 1977
Place of birth Munich, Germany
Working Career
Since Nov. 2007 L.E.K. Consulting GmbH, Munich
Associate
Dec. 2002 – Sept. 2007 LMU University of Munich
Chair for International Economics,, Prof. Dr. Dalia Marin
Research and Teaching Assistant
Apr. 2001 – Oct. 2002 ifo Institute for Economic Research, Munich
Directorate of social policy and labour markets
Student worker
May 2001 – Aug. 2001 Logistics Gateway Inc , Toronto (CAN)
Intern Business Development
Apr. 1998 – Apr. 2001 The Boston Consulting Group, Munich
Student worker
June 1999 – Aug. 2000 tecis FDL AG, Munich
Trainee / Junior Consultant
Military Service
09/1996 – 06/1997 Military service in Lenggries and White Sands, Texas (USA)
Academic Career
Apr. 2003 – Jan. 2008 LMU University of Munich
Doctoral Student in Economics
11/1997 – 11/2002 LMU University of Munich
Diploma-Student of Economics
09/1987 – 06/1996 Maria Theresia Gymnasium, Munich
High School Student
doc_465717889.pdf