Description
Since the release of the Horizontal Merger Guidelines, a body of literature has developed concerning various "price tests" routinely used in the delineation of markets in competition analysis. These empirical tests include price correlations, Granger causality, cointegration and stationarity analysis, and econometric models of price responses based on natural experiments.
______________________________________________________________________________
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
The Role of Price Tests in Market Definition
(A Comment Pertaining to #2b and #4 of the Questions for Public Comment)
HMG Review Project – Comment, Project No. P092900
James F. Nieberding, Ph.D.
LECG, LLC
1
November 9, 2009
Issue: Should The Guidelines Include A Discussion Of The Role That Price Tests Can Play In Identifying
Substitute Products To Complement the Hypothetical Monopolist Test For Market Definition?
*************************************************************
Recommendation: Since the release of the Horizontal Merger Guidelines
(“Guidelines”), a body of literature has developed concerning various “price tests” routinely used
in the delineation of markets in competition analysis. These empirical tests include price
correlations, Granger causality,
2
cointegration and stationarity analysis,
3
and econometric models
of price responses based on natural experiments.
4
These price tests are based on the premise that
if products are in the same market, then their prices ought to be systematically related. If not,
then shocks to one of the prices should not have an impact on the other prices. In instances
where reliable price data are available, and economic evidence indicates that price tests are
appropriate, price tests can assist with market definition in identifying substitute products while
1
Nieberding is a Principal at LECG in the global competition practice. The views expressed are those of the author
and should not be construed as representing the position of LECG or other experts in LECG.
2
The Granger test can determine how much of the variability in current prices in one region is explained by past (or
“lagged”) prices from that region and then determines how much additional variation can be explained by
controlling for past prices in other regions. If past values of prices from other regions significantly enhance the
prediction of the current price in the region at issue, then prices from the other regions are said to “Granger cause”
prices in the region at issue. A finding that prices from other regions affect those in the region at issue would
provide statistical evidence consistent with a broad geographic market.
3
Cointegration is a statistical technique that estimates whether various time series (e.g., prices in different regions)
systematically adjust back to an equilibrium relationship even in the presence of trending variables. If two or more
regional price series are found to be cointegrated, then their relative values do not move too far apart from their
longer-run equilibrium relationship and they systematically return to that relationship after temporary short-run
disturbances. Accordingly, cointegrated prices are consistent with a longer-run equilibrium relationship among the
prices in different regions, and support a conclusion that the regions are part of the same geographic market
4
The first three techniques are “pure” price tests in that they generally restrict the analysis to only the prices under
study, whereas a natural experiment may be done in the context of a broader econometric model.
1
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
the basic aspects of the Guidelines are retained. Price-based tests can provide empirical evidence
that, together with other economic evidence for the merger under review, could narrow, broaden,
or affirm candidate markets based upon the hypothetical monopolist test. Used in conjunction
with the hypothetical monopolist test, price-based tests could result in a more accurate and
efficient merger review process. As a result, the Guidelines would benefit from a discussion
about price tests and how they might inform the market definition process.
Analysis To Support Recommendation: As an example, when contemplating a
geographic market, price tests can be applied to price from different geographic areas to
empirically test the hypothesis that different regions are interrelated in an economic sense. If
two or more regions are part of an integrated market, then prices in these regions should follow a
similar pattern, or be systematically related in an equilibrium relationship, over time. However,
if price tests find persistent price differences among the regions under study, this provides
evidence consistent with separate, not integrated, markets. Suppose, for example, that a tsunami
significantly reduces production capacity for a given product in a certain region. As a result,
prices for that product increase in the affected area. An econometric analysis of the price
responses across different regions might allow for inferences about the geographic extent of the
market. That is, a price test based upon this natural experiment may also allow for inferences
about the extent of this geographic market. If the price effects are confined to the area affected
by the tsunami, this would be consistent with the notion of a regional market. However, if the
price effects of the unexpected capacity reduction are shown to spread from one geographic area
to other geographic areas, this would be consistent with a geographic market broader than just
the affected region. Such an analysis can empirically test whether the relatively higher prices in
the affected region generate a sufficient supply response from other geographic areas to constrain
2
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
prices in the affected region. If so, then this provides evidence of an broader market in that these
“outside” areas might reasonably be included in the geographic market.
The natural experiment described above, and the resulting price increase in the affected
region, is similar to a hypothetical monopolist in the affected region attempting to increase prices
above some pre-merger level. In fact, a price-based natural experiment similar to that described
was favorably considered by the European Commission in their review of the Blackstone /
Acetex merger. In that matter, Roeller and Stehmann (2006) discuss how the European
Commission found that the natural experiment presented by the merging parties (based upon
unexpected plant outages) was helpful in broadening the European Commission’s initially
contemplated relevant geographic market:
Such natural experiments can be a suitable empirical methodology to shed light on the source of
existing competitive constraints that are likely to impede the exercise of market power. …
Unexpected outages, though short-lived, may provide some indication about the source of the
competitive constraint faced by producers located in the EEA [European Economic Area]. …
These various empirical analyses have enabled the Commission to determine that the EEA did not
constitute a distinct geographic market.
The strengths and weaknesses of the various price tests used in market definition are well
documented in the literature.
5
For example, one weakness is that the results of price tests may be
driven in part (or in full) by the effect of common factors/influences affecting the prices under
study. If so, then a finding that prices from different regions are related might not necessarily
show that these regions are part of the same market, but might merely reflect parallel movements
of common influences in each region over time. Unless the influence of common factors is
controlled for, one cannot be sure whether the results of price tests are spurious or valid. A
recognized benefit to using price tests is that they can be conducted with information only on
prices. These data often are more readily available in a time-constrained merger investigation
5
This Comment lists some of this literature as references.
3
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
than are data on firms’ quantities, which would be necessary (along with prices) for estimating a
demand system from which own-price and cross-price elasticities could be obtained.
A more fundamental issue about price tests is whether or not they are able to identify
relevant markets for antitrust purposes. There is a well recognized distinction between
“economic markets” and “antitrust markets” particularly with respect to geographic boundaries.
The latter is what concerns antitrust practitioners but the former may be what price tests actually
are defining. The issue that has arisen about price tests is whether they are able to help define
antitrust markets in which competitive effects and market power can be analyzed.
The geographic extent of an economic market is determined by arbitrage and
transportation costs, though it may be somewhat difficult to find a clear “gap” between products
that would indicate whether or not a particular product is “in” or “out” of a contemplated
geographic market.
6
The concept of an antitrust market from the Guidelines involves the
delineation of a product and geographic space within which market power by a hypothetical
monopolist can be exercised. But, as many have noted, there may not be a clear correspondence
between what the Guidelines call antitrust markets and what economists usually understand
economic markets to be. Relevant antitrust markets can be smaller or larger than the
corresponding economic markets. For example, an antitrust market could be larger than the
economic market where, due to an absence of entry barriers, firms from outside an industry
would enter quickly in response to a SSNIP. Price tests alone would be unable to identify such a
competitive constraint. As noted in a report commissioned by the UK’s Office of Fair Trading
(“OFT”) (1999), “What needs to be determined in defining an antitrust market is whether two
6
As noted in Church, J. and Ware R. (2000), Cournot’s definition of an economic market is still the one accepted
today (An economic market is “the entire territory of which parts are so united by the relations of unrestricted
commerce that the prices there take the same level throughout, with ease and rapidity.”) A similar definition of an
economic markets was advanced by Alfred Marshall, who defined it as an area in which “prices of the same goods
tend to equality with due allowance for transportation costs” (as quoted in Scheffman, D. and Spiller P. (1987)).
4
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
areas that are in the same market at historical prices would still be in the same market if the
producers in one area would increase their price by some significant and non-transitory amount.
This question cannot be answered by looking at price movements alone.”
Despite the debate about the utility of price tests in defining antitrust markets, price tests
do appear in merger analysis, usually as a component of a broader antitrust market-delineation
study. As noted by Coe and Krause (2008), “[p]rice-based tests of market delineation remain
popular for preliminary work in antitrust cases despite existing criticisms.” Boshoff (2006)
states, “In sum, tests of price co-movement have been criticized in the literature, but continue to
be employed by competition analysts due to their simplicity and the data constraints that prohibit
more advanced analysis, such as estimation of elasticities.”
Competition authorities, and economists who work for them, have also recognized the
usefulness of price-based tests in market delineation. For example, the OFT (1999) states,
All techniques based on the analysis of prices alone, including co-integration techniques, are very
useful to define economic markets, but they should be used with care when establishing relevant
antitrust markets. … Although the analysis of prices alone is not sufficient to establish whether a
market is not an antitrust market, it is often the case that no other data but time series of prices are
available to the analyst. In that case, the techniques developed in this chapter [dynamic price
regressions and cointegration analysis] should be used, as they are the most correct ones from a
statistical point of view. (emphasis in original)
And, as stated by Haldrup (2003),
We are aware that the concept of an anti-trust market as defined by the SSNIP-methodology is
different from the concept of an economic market. … Even though we realize that the market
concepts are different, we will maintain that much useful information can be extracted from
[price tests] based on the economic market concept, that is relevant for the anti-trust market and
that cannot be extracted in any other way.
However, as the OFT (1999) cautions,
The fact that the price of a certain product or area is found to affect prices of other products or
areas is not sufficient proof that those products or areas are in the same antitrust market. … This
question cannot be answered by looking at price movements alone. (emphasis in original).
Hosken and Taylor (2004), both FTC economists, state “there are clear advantages to
using a cointegration analysis to look at market definition.” These authors believe that despite
5
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
the recognized shortcomings of certain price tests, “price relationship tests of Forni and others
can be useful complements to more traditional inputs to the market definition exercise.” That is,
price tests can be most informative when accompanied by other analyses based on the
Guidelines’ approach to market definition, and by a thorough investigation of the economics of
industry under study. This sentiment regarding the usefulness of quantitative techniques in
market definition is echoed by the FTC and DOJ in their Commentary on the Horizontal Merger
Guidelines (2006):
To be probative, of course, such data analyses must be based on accepted economic principles,
valid statistical techniques, and reliable data. Moreover, the Agencies accord weight to such
analyses only within the context of the full investigatory record, including information and
testimony received from customers and other industry participants and from business documents.
Roeller and Stehmann (2006) state that in the Blackstone / Acetex merger review, the European
Commission determined that the price-based tests submitted on behalf of the merging parties
were useful in helping them form an opinion on the relevant geographic market:
7
The investigation became an interesting example for the role of quantitative analysis to delineate
antitrust markets. While the definition of product markets did not pose any challenge, one of the
key issues was the delineation of the relevant geographic market for each product affected by the
transaction. … These various empirical analyses have enabled the Commission to determine that
the EEA did not constitute a distinct geographic market. The relevant geographic market had to
include at least North America as well.
Durand and Rabassa (2005), two economists at the European Commission who worked on
Blackstone / Acetex, attest to the usefulness of the price-based tests in helping define the relevant
geographic market :
The econometric studies submitted by the parties were reviewed and extended by the member of
the CET [Chief Economist Team]. These various empirical analyses have enabled the
Commission to determine that the EEA did not constitute a distinct geographic market. The
relevant geographic market had to include at least North America as well.
7
The author, part of the economics team working on behalf of the merging parties, co-authored a 2005 submission
to DG COMP in Blackstone / Acetex entitled, “Price tests for geographic market definition: Cointegration analysis
and Granger causality applied to VAM and acetic acid prices,” by James Langenfeld, Mary Coleman, and James
Nieberding. See Commission Decision Case No COMP/M.3625 - Blackstone/Acetex, available at http://ec.europa.eu/competition/mergers/cases/decisions/m3625_20050713_20682_en.pdf (at footnote 14).
6
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
While some price tests might be more informative than others in assisting with market
definition, none are individually determinative. However, when properly used as a supplement
to the hypothetical monopolist test, price tests can aid in setting the framework for defining
markets for antitrust analysis in which competitive effects can be analyzed. For example, as in
Blackstone / Acetex, price tests can be used to empirically “check” whether a contemplated
geographic market is too small (i.e., supply shocks appear to transmit more widely than expected
under the candidate market definition). They also can check if a proposed geographic market is
too large (i.e., price effects are confined to a regional market and do not propagate to the larger
proposed area). Price tests can also inform the analysis on how to expand a market if a proposed
market is believed to be too narrow. That is, if shocks appear to propagate to prices more widely
than expected given an initial market definition, price tests might reveal where to look for
substitution. For example, suppose two additional markets (A and B) are believed to belong in a
candidate market. If price tests reveal that supply shocks transmit from the candidate market to
B (and perhaps vice versa) but not A, then B is the next best alternative.
A statement by Mncube et. al. (2006) of the Competition Commission of South Africa
provides an appropriate summary regarding the use of price-based tests in defining relevant
markets:
In spite of all the limitations, price tests can yield valuable information. … The use of price tests
can bring confidence to an analyst especially if the different tests point in the same direction and
support the documentary evidence and industry practices.
7
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
References
Audy, E. and Erutku, C. (2005) “Price Tests to Define Markets: An Application to Wholesale Gasoline in
Canada,” Journal of Industry, Competition and Trade, 5(2), pp. 137-154. Downloadable at http://www.springerlink.com/content/q50168v15487l228/
Boshoff, W. H. (2006), “Quantitative Competition Analysis: Stationarity Tests in Geographic Market
Definition,” Stellenbosch Economic Working Papers: 17/06. Available at http://www.ekon.sun.ac.za/wpapers/2006/wp172006/wp-17-2006.pdf
Cartwright, P. A., Kamerschen, D. R., and M. Huang (1989) “Price Correlation and Granger Causality
Tests for Market Definition,” Review of Industrial Organization 4, pp. 79-98. Downloadable at http://www.springerlink.com/content/ar5033p6lk458258/
Coe, P.J. and Krause, D. (2008) “An Analysis of Price-Based Tests of Antitrust Market Delineation,”
Journal of Competition Law & Economics, 4(4), pp. 983-1007. Downloadable at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1318125
Coleman, M. and Langenfeld, J. (2008) “Natural Experiment Analysis in Antitrust Analysis: The Use of
Daubert Criteria to Evaluate its Relevance and Reliability,” in ABA Section of Antitrust Law, Issues
in Competition Law and Policy, Wayne Dale Collins, et al, eds.
Commentary on the Horizontal Merger Guidelines (2006), U.S. Department of Justice and Federal Trade
Commission. Available athttp://www.justice.gov/atr/public/guidelines/215247.htm
Church, J. and Ware R. (2000) Industrial Organization, A Strategic Approach. Irwin McGraw-Hill.
Durand, B. and Rabassa, V. (2005) “The Role of Quantitative Analysis to Delineate Antitrust Markets:
An Example. Blackstone/Acetex,” Competition Policy News 3, pp. 118–122. Available at http://ec.europa.eu/competition/publications/cpn/cpn2005_3.pdf
Engle, R. and Granger, W. (1987) “Co-Integration and Error Correction: Representation, Estimation, and
Testing,” Econometrics, 55(2), pp. 251-276.
Forni, M. (2004) “Using Stationarity Tests in Antitrust Market Definition,” American Law and Economics
Review, 6(2), pp. 441-464. Downloadable athttp://papers.ssrn.com/sol3/papers.cfm?abstract_id=305420
Haldrup, N, (2003) “Empirical Analysis of Price Data in the Delineation of the Relevant Geographical
Market in Competition Analysis.” University of Aarhus, Economics Working Paper No. 2003-09.
Downloadable athttp://papers.ssrn.com/sol3/papers.cfm?abstract_id=429120
Hosken, D. and Taylor, C. (2004) “Discussion of ‘Using Stationarity Tests in Antitrust Market
Definition,’” American Law and Economics Review, 6(2), pp. 465-475. Downloadable at http://aler.oxfordjournals.org/cgi/content/extract/6/2/465
Mncube, L., Khumalo, J., Mokolo, R., and Njisane, Y. (2008) “Use of Price Correlation and Stationarity
Analysis in Market Definition” Lessons From a Recent Merger.” Policy and Research Division,
Competition Commission of South Africa. Available athttp://web.wits.ac.za/NR/rdonlyres/37777DEF-
BA14-432C-A1DD-085889BC7BBB/0/Mncube_PricetestsConferencepaper_3_.pdf
8
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
Quantitative Techniques in Competition Analysis, (1999) Prepared for the Office of Fair Trading by
LECG Ltd. Available athttp://www.oft.gov.uk/shared_oft/reports/comp_policy/oft266.pdf
Roeller, L.H. and Stehmann, O. (2006) “The Year 2005 at DG Competition: The Trend Toward a More
Effects-Based Approach,” Review of Industrial Organization, 29, pp. 281-304. Available at http://ec.europa.eu/dgs/competition/economist/competition2005.pdf
Slade, M. (1986) “Exogeneity Tests of Market Boundaries Applied to Petroleum Products,” Journal of
Industrial Economics, 34(3), pp. 291-303. Downloadable at http://econpapers.repec.org/article/blajindec/v_3a34_3ay_3a1986_3ai_3a3_3ap_3a291-303.htm
Scheffman, D. and Spiller P. (1987) “Geographic Market Definition Under the U.S. Department of
Merger Guidelines,” Journal of Law and Economics, 30(1), pp. 123-147. Downloadable at http://econpapers.repec.org/RePEc:ucp:jlawec:v:30:y:1987:i:1
:123-47
Warell, L. (2005) “Defining Geographic Coal Markets Using Price Data and Shipments Data,” Energy
Policy, 33(17), pp. 2216-2230. Available at http://www.kkv.se/upload/Filer/Forskare-
studenter/projekt/2001/proj122-2001_1.pdf
Werden, G. and Froeb, L. (1993) “Correlation, Causality, and All that Jazz: The Inherent Shortcomings of
Price Tests for Antitrust Market Delineation,” Review of Industrial Organization, 8, pp. 329-353.
9
doc_835589817.pdf
Since the release of the Horizontal Merger Guidelines, a body of literature has developed concerning various "price tests" routinely used in the delineation of markets in competition analysis. These empirical tests include price correlations, Granger causality, cointegration and stationarity analysis, and econometric models of price responses based on natural experiments.
______________________________________________________________________________
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
The Role of Price Tests in Market Definition
(A Comment Pertaining to #2b and #4 of the Questions for Public Comment)
HMG Review Project – Comment, Project No. P092900
James F. Nieberding, Ph.D.
LECG, LLC
1
November 9, 2009
Issue: Should The Guidelines Include A Discussion Of The Role That Price Tests Can Play In Identifying
Substitute Products To Complement the Hypothetical Monopolist Test For Market Definition?
*************************************************************
Recommendation: Since the release of the Horizontal Merger Guidelines
(“Guidelines”), a body of literature has developed concerning various “price tests” routinely used
in the delineation of markets in competition analysis. These empirical tests include price
correlations, Granger causality,
2
cointegration and stationarity analysis,
3
and econometric models
of price responses based on natural experiments.
4
These price tests are based on the premise that
if products are in the same market, then their prices ought to be systematically related. If not,
then shocks to one of the prices should not have an impact on the other prices. In instances
where reliable price data are available, and economic evidence indicates that price tests are
appropriate, price tests can assist with market definition in identifying substitute products while
1
Nieberding is a Principal at LECG in the global competition practice. The views expressed are those of the author
and should not be construed as representing the position of LECG or other experts in LECG.
2
The Granger test can determine how much of the variability in current prices in one region is explained by past (or
“lagged”) prices from that region and then determines how much additional variation can be explained by
controlling for past prices in other regions. If past values of prices from other regions significantly enhance the
prediction of the current price in the region at issue, then prices from the other regions are said to “Granger cause”
prices in the region at issue. A finding that prices from other regions affect those in the region at issue would
provide statistical evidence consistent with a broad geographic market.
3
Cointegration is a statistical technique that estimates whether various time series (e.g., prices in different regions)
systematically adjust back to an equilibrium relationship even in the presence of trending variables. If two or more
regional price series are found to be cointegrated, then their relative values do not move too far apart from their
longer-run equilibrium relationship and they systematically return to that relationship after temporary short-run
disturbances. Accordingly, cointegrated prices are consistent with a longer-run equilibrium relationship among the
prices in different regions, and support a conclusion that the regions are part of the same geographic market
4
The first three techniques are “pure” price tests in that they generally restrict the analysis to only the prices under
study, whereas a natural experiment may be done in the context of a broader econometric model.
1
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
the basic aspects of the Guidelines are retained. Price-based tests can provide empirical evidence
that, together with other economic evidence for the merger under review, could narrow, broaden,
or affirm candidate markets based upon the hypothetical monopolist test. Used in conjunction
with the hypothetical monopolist test, price-based tests could result in a more accurate and
efficient merger review process. As a result, the Guidelines would benefit from a discussion
about price tests and how they might inform the market definition process.
Analysis To Support Recommendation: As an example, when contemplating a
geographic market, price tests can be applied to price from different geographic areas to
empirically test the hypothesis that different regions are interrelated in an economic sense. If
two or more regions are part of an integrated market, then prices in these regions should follow a
similar pattern, or be systematically related in an equilibrium relationship, over time. However,
if price tests find persistent price differences among the regions under study, this provides
evidence consistent with separate, not integrated, markets. Suppose, for example, that a tsunami
significantly reduces production capacity for a given product in a certain region. As a result,
prices for that product increase in the affected area. An econometric analysis of the price
responses across different regions might allow for inferences about the geographic extent of the
market. That is, a price test based upon this natural experiment may also allow for inferences
about the extent of this geographic market. If the price effects are confined to the area affected
by the tsunami, this would be consistent with the notion of a regional market. However, if the
price effects of the unexpected capacity reduction are shown to spread from one geographic area
to other geographic areas, this would be consistent with a geographic market broader than just
the affected region. Such an analysis can empirically test whether the relatively higher prices in
the affected region generate a sufficient supply response from other geographic areas to constrain
2
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
prices in the affected region. If so, then this provides evidence of an broader market in that these
“outside” areas might reasonably be included in the geographic market.
The natural experiment described above, and the resulting price increase in the affected
region, is similar to a hypothetical monopolist in the affected region attempting to increase prices
above some pre-merger level. In fact, a price-based natural experiment similar to that described
was favorably considered by the European Commission in their review of the Blackstone /
Acetex merger. In that matter, Roeller and Stehmann (2006) discuss how the European
Commission found that the natural experiment presented by the merging parties (based upon
unexpected plant outages) was helpful in broadening the European Commission’s initially
contemplated relevant geographic market:
Such natural experiments can be a suitable empirical methodology to shed light on the source of
existing competitive constraints that are likely to impede the exercise of market power. …
Unexpected outages, though short-lived, may provide some indication about the source of the
competitive constraint faced by producers located in the EEA [European Economic Area]. …
These various empirical analyses have enabled the Commission to determine that the EEA did not
constitute a distinct geographic market.
The strengths and weaknesses of the various price tests used in market definition are well
documented in the literature.
5
For example, one weakness is that the results of price tests may be
driven in part (or in full) by the effect of common factors/influences affecting the prices under
study. If so, then a finding that prices from different regions are related might not necessarily
show that these regions are part of the same market, but might merely reflect parallel movements
of common influences in each region over time. Unless the influence of common factors is
controlled for, one cannot be sure whether the results of price tests are spurious or valid. A
recognized benefit to using price tests is that they can be conducted with information only on
prices. These data often are more readily available in a time-constrained merger investigation
5
This Comment lists some of this literature as references.
3
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
than are data on firms’ quantities, which would be necessary (along with prices) for estimating a
demand system from which own-price and cross-price elasticities could be obtained.
A more fundamental issue about price tests is whether or not they are able to identify
relevant markets for antitrust purposes. There is a well recognized distinction between
“economic markets” and “antitrust markets” particularly with respect to geographic boundaries.
The latter is what concerns antitrust practitioners but the former may be what price tests actually
are defining. The issue that has arisen about price tests is whether they are able to help define
antitrust markets in which competitive effects and market power can be analyzed.
The geographic extent of an economic market is determined by arbitrage and
transportation costs, though it may be somewhat difficult to find a clear “gap” between products
that would indicate whether or not a particular product is “in” or “out” of a contemplated
geographic market.
6
The concept of an antitrust market from the Guidelines involves the
delineation of a product and geographic space within which market power by a hypothetical
monopolist can be exercised. But, as many have noted, there may not be a clear correspondence
between what the Guidelines call antitrust markets and what economists usually understand
economic markets to be. Relevant antitrust markets can be smaller or larger than the
corresponding economic markets. For example, an antitrust market could be larger than the
economic market where, due to an absence of entry barriers, firms from outside an industry
would enter quickly in response to a SSNIP. Price tests alone would be unable to identify such a
competitive constraint. As noted in a report commissioned by the UK’s Office of Fair Trading
(“OFT”) (1999), “What needs to be determined in defining an antitrust market is whether two
6
As noted in Church, J. and Ware R. (2000), Cournot’s definition of an economic market is still the one accepted
today (An economic market is “the entire territory of which parts are so united by the relations of unrestricted
commerce that the prices there take the same level throughout, with ease and rapidity.”) A similar definition of an
economic markets was advanced by Alfred Marshall, who defined it as an area in which “prices of the same goods
tend to equality with due allowance for transportation costs” (as quoted in Scheffman, D. and Spiller P. (1987)).
4
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
areas that are in the same market at historical prices would still be in the same market if the
producers in one area would increase their price by some significant and non-transitory amount.
This question cannot be answered by looking at price movements alone.”
Despite the debate about the utility of price tests in defining antitrust markets, price tests
do appear in merger analysis, usually as a component of a broader antitrust market-delineation
study. As noted by Coe and Krause (2008), “[p]rice-based tests of market delineation remain
popular for preliminary work in antitrust cases despite existing criticisms.” Boshoff (2006)
states, “In sum, tests of price co-movement have been criticized in the literature, but continue to
be employed by competition analysts due to their simplicity and the data constraints that prohibit
more advanced analysis, such as estimation of elasticities.”
Competition authorities, and economists who work for them, have also recognized the
usefulness of price-based tests in market delineation. For example, the OFT (1999) states,
All techniques based on the analysis of prices alone, including co-integration techniques, are very
useful to define economic markets, but they should be used with care when establishing relevant
antitrust markets. … Although the analysis of prices alone is not sufficient to establish whether a
market is not an antitrust market, it is often the case that no other data but time series of prices are
available to the analyst. In that case, the techniques developed in this chapter [dynamic price
regressions and cointegration analysis] should be used, as they are the most correct ones from a
statistical point of view. (emphasis in original)
And, as stated by Haldrup (2003),
We are aware that the concept of an anti-trust market as defined by the SSNIP-methodology is
different from the concept of an economic market. … Even though we realize that the market
concepts are different, we will maintain that much useful information can be extracted from
[price tests] based on the economic market concept, that is relevant for the anti-trust market and
that cannot be extracted in any other way.
However, as the OFT (1999) cautions,
The fact that the price of a certain product or area is found to affect prices of other products or
areas is not sufficient proof that those products or areas are in the same antitrust market. … This
question cannot be answered by looking at price movements alone. (emphasis in original).
Hosken and Taylor (2004), both FTC economists, state “there are clear advantages to
using a cointegration analysis to look at market definition.” These authors believe that despite
5
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
the recognized shortcomings of certain price tests, “price relationship tests of Forni and others
can be useful complements to more traditional inputs to the market definition exercise.” That is,
price tests can be most informative when accompanied by other analyses based on the
Guidelines’ approach to market definition, and by a thorough investigation of the economics of
industry under study. This sentiment regarding the usefulness of quantitative techniques in
market definition is echoed by the FTC and DOJ in their Commentary on the Horizontal Merger
Guidelines (2006):
To be probative, of course, such data analyses must be based on accepted economic principles,
valid statistical techniques, and reliable data. Moreover, the Agencies accord weight to such
analyses only within the context of the full investigatory record, including information and
testimony received from customers and other industry participants and from business documents.
Roeller and Stehmann (2006) state that in the Blackstone / Acetex merger review, the European
Commission determined that the price-based tests submitted on behalf of the merging parties
were useful in helping them form an opinion on the relevant geographic market:
7
The investigation became an interesting example for the role of quantitative analysis to delineate
antitrust markets. While the definition of product markets did not pose any challenge, one of the
key issues was the delineation of the relevant geographic market for each product affected by the
transaction. … These various empirical analyses have enabled the Commission to determine that
the EEA did not constitute a distinct geographic market. The relevant geographic market had to
include at least North America as well.
Durand and Rabassa (2005), two economists at the European Commission who worked on
Blackstone / Acetex, attest to the usefulness of the price-based tests in helping define the relevant
geographic market :
The econometric studies submitted by the parties were reviewed and extended by the member of
the CET [Chief Economist Team]. These various empirical analyses have enabled the
Commission to determine that the EEA did not constitute a distinct geographic market. The
relevant geographic market had to include at least North America as well.
7
The author, part of the economics team working on behalf of the merging parties, co-authored a 2005 submission
to DG COMP in Blackstone / Acetex entitled, “Price tests for geographic market definition: Cointegration analysis
and Granger causality applied to VAM and acetic acid prices,” by James Langenfeld, Mary Coleman, and James
Nieberding. See Commission Decision Case No COMP/M.3625 - Blackstone/Acetex, available at http://ec.europa.eu/competition/mergers/cases/decisions/m3625_20050713_20682_en.pdf (at footnote 14).
6
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
While some price tests might be more informative than others in assisting with market
definition, none are individually determinative. However, when properly used as a supplement
to the hypothetical monopolist test, price tests can aid in setting the framework for defining
markets for antitrust analysis in which competitive effects can be analyzed. For example, as in
Blackstone / Acetex, price tests can be used to empirically “check” whether a contemplated
geographic market is too small (i.e., supply shocks appear to transmit more widely than expected
under the candidate market definition). They also can check if a proposed geographic market is
too large (i.e., price effects are confined to a regional market and do not propagate to the larger
proposed area). Price tests can also inform the analysis on how to expand a market if a proposed
market is believed to be too narrow. That is, if shocks appear to propagate to prices more widely
than expected given an initial market definition, price tests might reveal where to look for
substitution. For example, suppose two additional markets (A and B) are believed to belong in a
candidate market. If price tests reveal that supply shocks transmit from the candidate market to
B (and perhaps vice versa) but not A, then B is the next best alternative.
A statement by Mncube et. al. (2006) of the Competition Commission of South Africa
provides an appropriate summary regarding the use of price-based tests in defining relevant
markets:
In spite of all the limitations, price tests can yield valuable information. … The use of price tests
can bring confidence to an analyst especially if the different tests point in the same direction and
support the documentary evidence and industry practices.
7
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
References
Audy, E. and Erutku, C. (2005) “Price Tests to Define Markets: An Application to Wholesale Gasoline in
Canada,” Journal of Industry, Competition and Trade, 5(2), pp. 137-154. Downloadable at http://www.springerlink.com/content/q50168v15487l228/
Boshoff, W. H. (2006), “Quantitative Competition Analysis: Stationarity Tests in Geographic Market
Definition,” Stellenbosch Economic Working Papers: 17/06. Available at http://www.ekon.sun.ac.za/wpapers/2006/wp172006/wp-17-2006.pdf
Cartwright, P. A., Kamerschen, D. R., and M. Huang (1989) “Price Correlation and Granger Causality
Tests for Market Definition,” Review of Industrial Organization 4, pp. 79-98. Downloadable at http://www.springerlink.com/content/ar5033p6lk458258/
Coe, P.J. and Krause, D. (2008) “An Analysis of Price-Based Tests of Antitrust Market Delineation,”
Journal of Competition Law & Economics, 4(4), pp. 983-1007. Downloadable at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1318125
Coleman, M. and Langenfeld, J. (2008) “Natural Experiment Analysis in Antitrust Analysis: The Use of
Daubert Criteria to Evaluate its Relevance and Reliability,” in ABA Section of Antitrust Law, Issues
in Competition Law and Policy, Wayne Dale Collins, et al, eds.
Commentary on the Horizontal Merger Guidelines (2006), U.S. Department of Justice and Federal Trade
Commission. Available athttp://www.justice.gov/atr/public/guidelines/215247.htm
Church, J. and Ware R. (2000) Industrial Organization, A Strategic Approach. Irwin McGraw-Hill.
Durand, B. and Rabassa, V. (2005) “The Role of Quantitative Analysis to Delineate Antitrust Markets:
An Example. Blackstone/Acetex,” Competition Policy News 3, pp. 118–122. Available at http://ec.europa.eu/competition/publications/cpn/cpn2005_3.pdf
Engle, R. and Granger, W. (1987) “Co-Integration and Error Correction: Representation, Estimation, and
Testing,” Econometrics, 55(2), pp. 251-276.
Forni, M. (2004) “Using Stationarity Tests in Antitrust Market Definition,” American Law and Economics
Review, 6(2), pp. 441-464. Downloadable athttp://papers.ssrn.com/sol3/papers.cfm?abstract_id=305420
Haldrup, N, (2003) “Empirical Analysis of Price Data in the Delineation of the Relevant Geographical
Market in Competition Analysis.” University of Aarhus, Economics Working Paper No. 2003-09.
Downloadable athttp://papers.ssrn.com/sol3/papers.cfm?abstract_id=429120
Hosken, D. and Taylor, C. (2004) “Discussion of ‘Using Stationarity Tests in Antitrust Market
Definition,’” American Law and Economics Review, 6(2), pp. 465-475. Downloadable at http://aler.oxfordjournals.org/cgi/content/extract/6/2/465
Mncube, L., Khumalo, J., Mokolo, R., and Njisane, Y. (2008) “Use of Price Correlation and Stationarity
Analysis in Market Definition” Lessons From a Recent Merger.” Policy and Research Division,
Competition Commission of South Africa. Available athttp://web.wits.ac.za/NR/rdonlyres/37777DEF-
BA14-432C-A1DD-085889BC7BBB/0/Mncube_PricetestsConferencepaper_3_.pdf
8
Nieberding: The Role of Price Tests in Market Definition HMG Review Project – Comment, Project No. P092900
Quantitative Techniques in Competition Analysis, (1999) Prepared for the Office of Fair Trading by
LECG Ltd. Available athttp://www.oft.gov.uk/shared_oft/reports/comp_policy/oft266.pdf
Roeller, L.H. and Stehmann, O. (2006) “The Year 2005 at DG Competition: The Trend Toward a More
Effects-Based Approach,” Review of Industrial Organization, 29, pp. 281-304. Available at http://ec.europa.eu/dgs/competition/economist/competition2005.pdf
Slade, M. (1986) “Exogeneity Tests of Market Boundaries Applied to Petroleum Products,” Journal of
Industrial Economics, 34(3), pp. 291-303. Downloadable at http://econpapers.repec.org/article/blajindec/v_3a34_3ay_3a1986_3ai_3a3_3ap_3a291-303.htm
Scheffman, D. and Spiller P. (1987) “Geographic Market Definition Under the U.S. Department of
Merger Guidelines,” Journal of Law and Economics, 30(1), pp. 123-147. Downloadable at http://econpapers.repec.org/RePEc:ucp:jlawec:v:30:y:1987:i:1
:123-47 Warell, L. (2005) “Defining Geographic Coal Markets Using Price Data and Shipments Data,” Energy
Policy, 33(17), pp. 2216-2230. Available at http://www.kkv.se/upload/Filer/Forskare-
studenter/projekt/2001/proj122-2001_1.pdf
Werden, G. and Froeb, L. (1993) “Correlation, Causality, and All that Jazz: The Inherent Shortcomings of
Price Tests for Antitrust Market Delineation,” Review of Industrial Organization, 8, pp. 329-353.
9
doc_835589817.pdf