Economics Overview

Changes in the Horizontal Merger Guidelines and Market Definition in Monopolisation Cases

On 20 April 2010, the Federal Trade Commission and the Department of Justice (the Agencies) jointly released proposed updated Horizontal Merger Guidelines (PHMGs) for public comment. ¹ ,² The Agencies received many thoughtful comments that focused on the use of the PHMGs in evaluating mergers. ³ However, the effect of the PHMGs likely will reach beyond horizontal mergers. In particular, parties in monopolisation cases brought under section 2 of the Sherman Act often look to the hypothetical monopolist test outlined in the existing 1992 Horizontal Merger Guidelines (HMGs) as a guide for defining a relevant market. ⁴ Thus, the changes represented in the PHMGs, especially those related to the hypothetical monopolist test, would be expected to affect economic analysis in monopolisation cases as well.

This article provides a general discussion of the issues that can arise when the hypothetical monopolist test is used to define a relevant market in monopolisation cases and how the PHMGs may affect these issues.

Use of guidelines in mergers and monopolisation cases

The HMGs provide a clear conceptual guide to defining a relevant market in merger cases and assessing market power. In particular, the HMGs state that 'the Agency will delineate the product market to be a product or group of products such that a hypothetical profit-maximising firm that was the only present and future seller of those products ('monopolist') likely would impose at least a 'small but significant and nontransitory' increase in price.' ⁵ The Agency's procedure is referred to as the SSNIP test or the hypothetical monopolist test for defining a relevant market. ⁶

Antitrust merger analysis is concerned with identifying and prohibiting mergers if their effect 'may be substantially to lessen competition, or tend to create a monopoly.' ⁷ The HMGs note that 'mergers should not be permitted to create or enhance market power or to facilitate its exercise.' ⁸ Given this goal, merger analysis is necessarily forward-looking (ie, it is interested in what would happen to prices if the firms were to merge) and is concerned with any enhancement in market power, as expressed by higher market prices.

The issues are different in monopolisation cases, namely: whether or not the firm possesses market power in a properly defined market; and whether the firm obtained or maintained that power through anti-competitive conduct. ⁹ Both of these issues generally require defining a relevant market in which the firm's market power is assessed and anti-competitive conduct might have occurred.

Although the relevant market analysis is an important element of monopolisation cases, there is no specific guide for defining a relevant market in such cases. This often leads parties to look to the HMGs for guidance in defining the relevant market. ¹⁰ However, the difference in the questions being asked in monopolisation cases versus merger cases affects one's ability to apply the SSNIP test to market definition analysis in monopolisation cases, and may lead one to find either too much or too little market power.

First, in monopolisation cases where the firm is a monopolist or has significant market power, the SSNIP test could result in an overly wide market and thus understate the firm's market power. This will occur when the firm possesses significant market power and when actual prices are the base price against which the profitability of a small but significant non-transitory price increase is tested. Under these conditions, the SSNIP test will define an overly wide market because a rational monopolist would have increased its price to the point where further increases are not profitable. In other words, at the monopoly price the monopolist is on the elastic portion of its demand curve. The tendency for observed competition (and the cross-elasticity of demand at the observed prices) to overstate the amount of competition faced by a monopolist has been termed the 'cellophane fallacy' after the Supreme Court's decision in United States v EI DuPont De Nemours. ¹¹

It is worth highlighting that in contrast to monopolisation cases, the cellophane fallacy is not generally an issue in applying the hypothetical monopolist test to mergers. This is because the merging parties and possibly other firms are likely to compete prior to the merger, so the current market price is unlikely to be a monopoly price. ¹² Further, the goal of merger analysis is to assess whether the merger would enhance market power as opposed to identifying the extent of current market power. As a result, in the absence of collusion or an expected price change, the appropriate starting point of the analysis and the SSNIP test is the current state of competition.

Second, attempts to avoid the cellophane fallacy by applying the SSNIP test at an alternative 'competitive' price (ie, not the existing market price, but rather a price at which the alleged monopolist would earn 'competitive' profits) can lead to the opposite problem in which the SSNIP test points to an overly narrow market and thus overstates the firm's market power. The competitive price is defined as the price at which firms earn a 'normal' rate of return while covering their costs, including fixed costs and opportunity costs (ie, the competitive price is equal to the marginal 'economic' cost). In practice, it is rather difficult to estimate the competitive price because the information required to calculate economic costs, and thus the competitive price, is often not readily available. As a result, the firm's incremental cost is often used as proxy for the competitive price.

Applying the SSNIP test at the firm's incremental cost essentially assumes that incremental costs approximate marginal costs and that the competitive price is equal to the firm's marginal cost (ie, the firm earns zero profit from incremental sales at this point). If the firm's price exceeds its incremental cost by a small but significant amount (ie, the firm has positive profit), the test will imply that the firm's product is in a market of its own and that the firm has market power.

However, many firms earn positive profits even though they do not possess market power of antitrust concern. For example, when products are differentiated each firm faces a downward sloping demand and may earn positive profits without exercising market power of antitrust concern.

Implications of the PHMGs for monopolisation cases

The discussion of market definition and market power in the PHMGs differs from that in the HMGs in ways that suggest important implications for monopolisation cases. While the HMGs state that defining a market is a necessary step in evaluating a merger, the PHMGs downplay the role of market definition analysis. The PHMGs also propose changes to the SSNIP test and identify additional information that may be used when determining if a hypothetical monopolist would have an incentive to increase prices. At the same time, the PHMGs continue to treat supply-side substitution in a way that may be unnecessarily restrictive in monopolisation cases. In aggregate, if the changes outlined in the PHMGs were adopted in monopolisation cases they may increase the instances of firms being found to have market power of antitrust concern.

The role of market definition analysis

The HMGs state that defining a market is a necessary first step in evaluating a merger. ¹³ However, the PHMGs downplay the importance of market definition analysis by emphasising that market definition is only one of several tools and that defining relevant markets may not be necessary in some cases. ¹⁴ A number of commentators have questioned the move away from market definition, noting that the practice of defining a relevant market helps the Agencies to focus on key issues. ¹⁵

The move away from market definition may be more practical in merger cases than in monopolisation cases. The PHMGs suggest that market definition is being de-emphasised in part because '[s]ome of the analytical tools used by the Agencies to assess competitive effects do not rely on market definition.' ¹⁶ In particular, the PHMGs suggest the Agencies may use diversion ratios and upward pricing pressure to analyse the effects of a merger between firms that produce differentiated products. ¹⁷ At least in theory, these analytical techniques may be implemented without defining a relevant market. These techniques look only at the amount of substitution between the products of the merging firms and margins on these products to determine if a firm selling both firms' products would likely increase the price on at least one product. Regardless of whether these techniques are of practical value or not in the merger setting, such an approach is not applicable in monopolisation cases. This is because in a monopolisation case there is not a similar predefined pair of products or set of pairs to consider. For this reason, reliance on market definition is likely to remain more important in monopolisation cases than the new guidelines suggest it may become in mergers.

In addition, the PHMGs acknowledge that there may be more than one relevant market ¹⁸ and encourage the choice of the narrowest market. ¹⁹ Focusing on the narrowest possible market and ignoring 'significant' substitution may lead one to understate the extent of competition faced by a firm in a monopolisation case and subsequently to overstate the extent of the firm's market power.

The hypothetical monopolist test

The PHMGs propose changes to the SSNIP test and identify additional information that may be used when analysing whether a hypothetical monopolist would find a small but significant nontransitory price increase profitable.

The price examined in a SSNIP Test

The PHMGs change the focus of the SSNIP test from the prices of the merging firms' products to the 'value [the firms] contribute to the products or services used by customers,' or what economists call the 'value added' by the firm. ²⁰ The HMGs state that 'the price for which an increase will be postulated will be whatever is considered to be the price of the product at the stage of the industry being examined.' ²¹ In contrast, the PHMGs state that the 'SSNIP is intended to represent a 'small but significant' increase in the prices charged by firms in the candidate market for the value they contribute to the products or services used by customers.' ²² The PHMGs also propose a bar of 10 per cent as opposed to 5 per cent for a 'small but significant increase in price' in cases in which the value added is used, as opposed to the product price. ²³

The PHMGs do not describe how to calculate the value that a firm contributes to the products or services used by customers. Instead, the PHMGs offer two examples where the value a firm adds is calculated as the difference between the price of the product sold by the firm and the cost of the principal inputs. ²⁴

Focusing on the value added, instead of the price for the firm's product or service, ²⁵ means that the market definition will depend upon the degree of the firm's vertical integration. If the firm is more vertically integrated, then its value added will be a larger component of its product price. In this case, a change in the value added would translate into a larger change in its product price, and the SSNIP test is likely to have the same results regardless of whether it is performed using a product price or a value added.

However, if the firm is not vertically integrated then its value added may be small as a percentage of the price of its product. Hence, a small increase in its value added may translate into a very small change in its product price. Because few customers will switch to other suppliers in response to a very small change in the price of its product, the new SSNIP test on the change in the value added may find this firm to be in a market by itself even if the HMGs' version of the SSNIP would find a wider market.

For example, consider a monopolisation case involving the retail price of gasoline at gas stations. If the gas stations are owned by a company that is not vertically integrated, then the 10 per cent increase in value added would be based on a 10 per cent increase in the difference between wholesale and retail gas prices. If the gas stations are owned by a vertically integrated company that refines crude oil into gasoline, then the 10 per cent increase in value added would be based on a 10 per cent increase in the difference between retail gas prices and the price of the crude oil contained in the gasoline and other inputs. ²⁶ In dollar terms, the price increase for the vertically integrated firm will be larger than for the non-vertically integrated firm and therefore less likely to be sustainable. As a result, all else equal, it is likely that the firm that is not vertically integrated will be found to be in a narrower market than the vertically integrated firm. Therefore a different market may be defined entirely due to the degree of vertical differentiation and not market power. Note that the product price SSNIP test (both under the PHMGs and the HMGs) would give the same result for these two companies regardless of their degrees of vertical integration.

To offset what would otherwise be purely a move toward smaller SSNIPs, the PHMGs proposes a bar of 10 per cent as opposed to 5 per cent for a 'small but significant increase in price' in cases in which the value added is used as opposed to the product price. ²⁷ However, even with the higher bar, the value added SSNIP will likely define a narrower market than the firm's product price SSNIP in instances in which the firm's value added is small relative to the price of its product.

While the PHMGs suggest using value added, they also state that a SSNIP test based on product prices may be used '[w]here such implicit [value-added] prices cannot be identified with reasonable clarity.' ²⁸ Remember that the value added SSNIP test uses a 10 per cent price increase, while the product price SSNIP test uses an increase of 5 per cent. Now suppose a firm's value added accounts for 20 per cent of its product price (ie the cost of inputs accounts for 80 per cent of the product price). In this case a 10 per cent SSNIP on the value added translates into a 2 per cent change in the firm's product price. However, if the SSNIP test is conducted using the product price, it would use a 5 per cent change in the firm's product price. If consumer substitution to competing products in response to a 2 per cent product price increase is sufficiently smaller than substitution in response to a 5 per cent change in the product price, then the value added SSNIP test will find a narrower market than the product price SSNIP test for the same firm. This creates an incentive for the firm to argue in favour of using the product price SSNIP test over the value added SSNIP test. The opposite is true for a firm where the value added accounts for more than 50 per cent of the firm's product price.

Evidence relevant to the profitability of a SSNIP

Whether or not a hypothetical monopolist would find a SSNIP to be profitable depends upon how customers react to the price increase. In order to answer this question directly, one needs to estimate the elasticity of demand. However, as a result of limited data and lack of exogenous price variation, reliable elasticity estimates cannot be obtained in many cases. To obtain such estimates, one would ideally observe multiple instances of demand responses to price changes that were not associated with other changes that would affect supply and demand, such as marketing promotions or quality changes.

Reflecting the difficulty of estimating elasticity directly, both the HMGs and PHMGs identify potential evidence, besides a direct estimate of elasticity, that may be relevant in determining the profitability of a SSNIP, with the PHMGs adding additional evidence that could be considered. In particular, the PHMGs endorse the use of margins and critical loss analysis, which evaluates how a SSNIP on one product would affect the hypothetical monopolist's profits, as part of the hypothetical monopolist test. ²⁹

The PHMGs state that profit margins may be relevant in determining whether or not a hypothetical monopolist would find a SSNIP profitable. ³⁰ In particular, the PHMGs state: 'Higher pre-merger margins_ make it more likely that_ the candidate market satisfies the hypothetical monopolist test.' ³¹ This statement is based on economic theory where, under certain conditions, higher product margins are associated with less elastic demand. In particular, a profit-maximising firm sets its prices so that the ratio between the firm's profit margin and price (known as the Lerner Index) is inversely proportional to the product's demand elasticity. ³² Therefore, one can infer that products with higher margins have fewer close substitutes.

Theoretically, analysis of profit margins may be valuable in monopolisation cases. Recall that a true monopolist would be charging a profit-maximising price and therefore would not be able to sustain any further increase in price. The SSNIP test will fail for this monopolist because it is already receiving a monopoly price. In this case, high profit margins may correctly identify a monopolist where the SSNIP test does not.

On the other hand, high profit margins often do not reflect market power. For example, in the presence of fixed costs one will generally find high margins without any adverse acts on the part of the firm. This occurs because only high margins will ensure an adequate return on capital to provide incentives for entry into this industry. Similarly, in innovative industries with large spending on R&D, high profit margins simply enable the firms to recoup their R&D investments. A relatively low-cost, capacity-constrained firm that faces significant competition from higher-cost firms may also have high margins that reflect the firm's cost advantage as opposed to its having a market power of antitrust concern.

Furthermore, accounting margins likely to be used in monopolisation cases or merger analysis are not economic margins. They are more likely to capture variable costs than the cost of a marginal unit. Finally, the use of margins is an especially crude proxy for demand elasticity in markets where products are differentiated. In differentiated products, a firm will face a downward sloping demand and may well enjoy a positive margin. Therefore, the PHMGs' emphasis on margins may lead one to find market power where margins are high for the aforementioned reasons rather than because the firm has successfully monopolised the market.

The PHMGs state that the 'Agencies also may consider a 'critical loss analysis' to assess the extent to which it corroborates inferences' drawn from other evidence about the profitability of a SSNIP. ³³ 'Critical loss analysis asks whether imposing at least a SSNIP on one or more products in a candidate market would raise or lower the hypothetical monopolist's profits.' ³⁴ When a firm increases it prices, it loses some sales but makes higher profits on the remaining sales. Critical loss analysis considers whether the profit gain on remaining sales outweighs the profit loss on lost sales. If it does, the price increase would be profitable. The premise is that while elasticity estimates may not be available to conduct the SSNIP test, profit margins may be available and could be used to infer whether a price increase would be profitable. ³⁵

In a monopolisation case, critical loss analysis suffers from the same issues as the hypothetical monopolist test. In particular, a decline in the firm's profits in response to SSNIP could demonstrate either that the firm has no market power, or that the firm has a monopoly power and has already increased the price to a profit-maximising level. Further, as noted above, economic profits are rarely available and thus the 'profits' used in a critical loss analysis will likely not be meaningful.

Supply side substitution in market definition

The PHMGs, like the HMGs, do not treat supply-side and demand-side substitution symmetrically when defining relevant markets. Rather, they define a market based on demand substitution alone and then separately consider the possibility of supply-side substitutability and entry after defining the market. ³⁶ This creates an asymmetry where market participants may be identified based on a supply response, but supply-side responses are not considered when drawing the boundaries of the product market in which these market participants compete. ³⁷

To an economist, a customer's ability and willingness to substitute away from one product to another necessarily involves an analysis of what products are available (or will be available at the time the customer makes a purchase) for substitution - or what products are (or will be) supplied. This calls for analysing the supply-side of the market. From the economist's point of view, one ought to treat market definition symmetrically and consider both supply-side and demand-side substitution when defining a relevant market. As described by a popular economic textbook, 'A proper definition of the product dimension of a market should include all those products that are close demand or supply substitutes.' ³⁸ The HMGs' approach has been, and with the PHMGs will apparently continue to be, problematic when applied to monopolisation cases because monopolisation claims arise from restrictions on the supply side of the market.

Summary

The PHMGs will likely have effects beyond horizontal mergers. In particular, the PHMGs will likely affect economic analysis in monopolisation cases as well.

We discuss how the PHMGs may affect the analysis of market definition and market power issues in monopolisation cases. We find that the changes in the PHMGs may lead to narrower markets than analysis conducted using the methodology from HMGs would suggest, thereby increasing the instances in which market power and the ability to monopolise will be found.

Notes

1

. The Department of Justice and the Federal Trade Commission issued the final revised Horizontal Merger Guidelines on 19 August 2010. Although this was too late to update the article, our comments remain relevant because the final version of the guidelines is largely similar to the proposed version discussed herein. The one substantial change in the final guidelines that impacts our discussion is the removal of any mention of using a 10 per cent rather than a 5 per cent SSNIP when applying the SSNIP to the value added by the merging firms. This change reinforces our broader point that the emphasis the new guidelines place on using value added for the SSNIP rather than the price of goods sold is likely to lead to narrower markets.

2

. US Department of Justice and Federal Trade Commission, 'Horizontal Merger Guidelines For Public Comment' April 20, 2010, ('PHMGs').

3

. Public comments on the proposed update of the Horizontal Merger Guidelines are available at www.ftc.gov/os/comments/hmgrevisedguides/.

4

. See eg, Kentucky Speedway LLC v NASCAR Inc., 531 F.2d 196 (plaintiff's expert performed what the district court termed his 'own version' of SSNIP test); United States v Visa USA, Inc. 344 F.3d 229 (2d Cir. 2003) (the government's expert testified that in the face of a significant increase in card fees customers would not switch to another form of payment); United States v Archer-Daniels-Midland Co. and Nabisco Brands, Inc., 866 F.2d 242 at 244 (1988) (the government argued that under the Merger Guidelines 'the evidence that HFCS [high fructose corn syrup] purchasers will not switch back to sugar as long as HFCS remains cheaper is sufficient to establish that sugar and HFCS are not in the same product market').

5

. US Department of Justice and Federal Trade Commission, 'Horizontal Merger Guidelines,' Revised 8 April 1997, ('HMGs'), section 1.11.

6

. Essentially, the SSNIP test attempts to establish whether an increase in prices by the merged firm will be thwarted by consumers substituting towards products produced by other firms. It does this by finding the set of products to which consumers would turn in response to the merged firm increasing the prices of its products. So fundamentally, the question asked by the SSNIP test is about the degree of substitution between the merging firms' products and other products. When there is little substitution, the SSNIP test will determine a narrow market where the products of the merged firm will have a high market share, implying the merger will likely lead to an increase in market power.

7

. 15 U.S.C. section 18 (1988).

8

. HMGs section 0.1.

9

. United States v Grinnell Corp., 384 US 563, 570-571 (1966).

10

. See footnote 3.

11

. 351 US 377. DuPont, which sold 75 per cent of the cellophane in the United States, was alleged to have monopolised the US cellophane market. The Supreme Court concluded that DuPont did not possess monopoly power because the price that it could charge was constrained by other flexible packaging materials and thus the relevant market was the market for 'flexible packing materials,' not the market for cellophane. The decision has been criticised because other evidence, such as the profits DuPont earned on cellophane and lack of changes in cellophane prices when prices of other wrapping materials were changing, suggested that had cellophane been priced more competitively it would not have competed with other 'flexible packing materials.' The most important criticism of that decision comes from the recognition that at the monopoly price (ie, the price that maximises the monopolist's profits) the demand faced by the monopolist is elastic and the monopolist cannot further increase price without a loss in profits. Landes, William M., and Richard A. Posner, 'Market Power in Antitrust Cases,' 94 Harv. L. Rev. 937 (1980-1981), pp.960-961.

12

. There may be exceptions to this general rule. For example, if there is already collusion in the market or one firm already has close to monopoly power in a market, prices even in a merger case may start at or near a monopoly level and the SSNIP test may lead one to define a market that is too broad. These circumstances are likely to be relatively rare in merger cases in contrast to monopolisation cases.

13

. 'First, the Agency assesses whether the merger would significantly increase concentration and result in a concentrated market, properly defined and measured.' HMGs section 0.2.

14

. 'Market definition is not an end in itself: it is one of the tools the Agencies use to assess whether a merger is likely to lessen competition. Market definition identifies an arena of competition and enables the identification of market participants and the measurement of market shares and market concentration. . . The Agencies' analysis need not start with market definition. Some of the analytical tools used by the Agencies to assess competitive effects do not rely on market definition, although evaluation of competitive alternatives available to customers is always necessary at some point in the analysis.' PHMGs section 4.

15

. See, eg, 'A reduction in the role of market definition in merger analysis would be wrong as a matter of policy, as a matter of law, and as a matter of international competition norms that the Agencies and other have espoused.' Comments of the ABA Section of Antitrust Law, HMGs Revision Project – Comment, Project No. P092900, 4 June 2010, p.6. Also, 'to the extent that the Agencies may seek to reduce the role of market definition in some investigations… the risk is not only that the Agencies may fail to include a market definition 'reality check' in their own investigations that would help screen out potentially weak cases, but also that they may inadvertently encourage less rigorous analysis abroad.' 'Comments of General Electric Company, United Technologies Corporation and Honeywell International Inc.,' 4 June 2010, p. 11.

16

. PHMGs section 4.

17

. PHMGs section 6.1.

18

. 'Relevant markets need not have precise metes and bounds.' 'The hypothetical monopolist test ensures that markets are not defined too narrowly, but it does not lead to a single relevant market.' PHMGs section 4.

19

. In particular, the PHMGs state that '[d]efining a market broadly to include relatively distant product or geographic substitutes can lead to misleading market shares' and that '[m]arket shares of different products in narrowly defined markets are more likely to capture the relative competitive significance of these products, and often more accurately reflect competition between close substitutes.' The PHMGs also encourage defining 'a group of products as a relevant market even if customers would substitute significantly to products outside that group in response to a price increase.' PHMGs sections 4, 4.1.1.

20

. PHMGs section 4.1.2.

21

. HMGs section 1.11 (4).

22

. PHMGs section 4.1.2.

23

. PHMGs section 4.1.2.

24

. The PHMGs offer two examples of how using the 'value they contribute' might apply in practice. The first example is an oil pipeline that buys oil at one end and sells it at another. The firm's specific contribution is then the difference in these two prices. Similarly, for a company that installs computers, 'the implicit installation fee is equal to the package charge to customers less the price of the computers.' PHMGs section 4.1.2. The pipeline example was present in the HMGs as well, but the emphasis on the value added is new in the PHMGs.

25

. For an intermediate product or a service that would be the price of the firm's product or a service, not the price of a final product.

26

. If, alternatively, the gas stations are owned by a fully vertically integrated company that does oil exploration and extraction, then the SSNIP would be based on the price customers pay for gas and a 5 per cent SSNIP would be used.

27

. Note that a 5 per cent change in the product price would exceed a change in the product price due to a 10 per cent change in the value added if the value added accounts for less than 50 per cent of the product price, or if inputs account for more than 50 per cent of the final product price.

28

. PHMGs section 4.1.2.

29

. PHMGs section 4.1.3.

30

. PHMGs section 4.1.3.

31

. PHMGs section 4.1.3.

32

. See Tirole, Jean, The Theory of Industrial Organisation, The MIT Press, 2003, pp. 218-221.

33

. PHMGs section 4.1.3.

34

. PHMGs section 4.1.3.

35

. However, this premise is only true because profit margins and elasticities are directly related. For example, for a monopolist, a profit margin is equal to the inverse of the elasticity of demand. Hence, if demand elasticities are unavailable, economic profit margins also are unlikely to be available, limiting the usefulness of critical loss analysis.

36

. PHMGs sections 4, 9.

37

. The PHMGs specify that 'Market definition focuses solely on demand substitution factors, ie, on customers' ability and willingness to substitute away from one product to another in response to a price increase or a corresponding non-price change such as a reduction in product quality or service.' PHMGs section 4.

38

. Carlton and Perloff, Modern Industrial Organisation, 4th ed. 2005, p. 646.