Archives For

Steve’s next, perhaps final, installment, responding to Dan’s latest post on the appropriate liability rule for loyalty discounts. Other posts in the series: SteveDanSteveDan, and Thom.

My invitation comes with several hopefully final observations.

(1) Dan says, “There’s neither input foreclose nor output foreclosure if a rival can neutralize a loyalty discount without pricing unprofitably.”  My examples showed several reasons why an equally efficient rival may not be able to overcome a loyalty discount to get distribution, even if the monopolist does not violate the price-cost test.  Dan’s response to my #3(a) example is that “a discount structure that would require the entrant to pay rebates of $71 even while earning revenues of $70, then the rebate system would fail the price-cost screen.” That is not a good answer for three reasons.  First, the monopolist’s profits pre-entry absent the rebate are $200, and Dan uses the pre-entry situation as the base.  Second, the $71 payment would fail a test of comparing the rival’s price and cost, but that it is not the test.  The test compares the monopolist’s price and cost.   Third, if Dan now wants to use as the base the “but-for world” where the former-monopolist would earn only $70 because he would have lowered price in response to the entrant absent the loyalty discount, then he is accepting the “penalty price” scenario, which he elsewhere rejected as unlikely.  So, Dan has to make up his mind on this last point.  And the price-cost test is defective either way.

(2) Dan says, “Steve posits ‘that economic analysis is the same’ for loyalty discounts and contractual commitments not to buy from rivals.  Really?”  The short answer is “Yes.”  As Dan observes himself, “at a high level of generality one would say that in both cases the question is whether the price or contract forecloses competitors.”  By the way, a small rebate certainly will not always exclude.  But, see #3(b) in my previous post, which is a pretty general example of why the rival often would just give up rather than get involved in a bidding war for distribution.  Dan never explain what he thinks is wrong with this example.

(3) Dan says, “Steve presents four examples of how an auction for contractual exclusivity results in a single firm obtaining contractual exclusivity with anticompetitive effects due to asymmetry of incentives, all of which begs the question of how loyalty discounts without contractual exclusivity achieve the same effect.”  I assumed no contract.  Nor is a formal auction required.  The monopolist simply can unilaterally offer a high enough bid that the entrant walks away.  And in the #3(b) example, the unilateral bid does not even need to be very high.

(4) Dan says, “The monopolist is not saved from profit sacrifice, as Steve posits, merely by threatening a price of $105 that it never has to charge.”  Why not?  The monopolist never sells any units at $105 when the threat succeeds.  And, given Dan’s model, it will succeed against an equally efficient competitor.  With adequate information, it is neither expensive nor risky.  If Dan is going to rely on it being “expensive and risky,” then he is changing his story.  More importantly, he also is opining it should be permissible for a well-informed monopolist to exclude with penalty pricing threats because poorly informed monopolists may not follow the same strategy.  That does not make sense.   He also does not provide any evidence about how well informed most monopolists who engage in loyalty discounts typically are.

(5) Dan says, “Finally, let me underline as I did in my last post that I’m not claiming that “penalty pricing” is economically impossible.  Rather, …we should therefore not assume that it will be routinely deployed, and that the starting presumption should be that most loyalty discounts are true reductions from the but-for price.”  I do not have to assume penalties will be “routinely deployed.”   It was only one of my several fatal problems with the price-cost test.  But, in terms of the evidence, what is the evidence that the typical loyalty discount by a monopolist or firm with substantial market power are true reductions from the but-for price.  Reliable presumptions need to be more than simply reflections of one’s ideology.

(6) I want to conclude by observing that modern day courts do not need the defective price-cost test to protect monopolists.  Remember that the rival must prove consumer injury.  That is a high burden to carry and some of the factors Dan mentions might throw light on that bottom line issue.  Neither Josh nor I are claiming that loyalty discounts should be per se illegal.  So, Dan, come on over to the rule of reason.

Guest post by Steve Salop responding to Dan’s latest post on the appropriate liability rule for loyalty discounts. Other posts in the series: SteveDan, and Thom.

(1) Dan says that price-cost test should apply to “customer foreclosure” allegations.   One of my key points was that many loyalty discount claims involve “input foreclosure” or “raising rivals’ costs” effects, not plain-vanilla customer foreclosure.   In addition, loyalty agreements with distributors often involve input foreclosure because “distribution services” are an input and a rebate might be characterized as a reward payment for the (near-) exclusivity.    From his silence on the issue, I am inclined to presume that Dan would agree that the price-cost test should not be applied to such allegations.      Dan, what do you intend?

(2) Dan says that he agrees that the price-cost test should not be required for “partial exclusivity contracts” that involve contractual commitments to limit purchases from rivals.  He says that the price-cost test should apply only where the “claimed exclusionary mechanism is the price term.”  This distinction is peculiar because the economic analysis is the same in both situations.  In addition, even such voluntary exclusivity flowing from a price term can be anticompetitive, and even if the price-cost test is passed.  There are numerous reasons for this, as I explained in my original post. (I also discuss these issues in my contribution to Robert Pitofsky’s volume, “How the Chicago School Overshot the Mark.”  See also articles by Eric Rasmussen et. al., Michael Whinston and others.)

(3) Consider the following numerical examples that concretely illustrate the economic forces at work when there is competition for distribution, even in the absence of contractual commitments.

(a) Suppose that a monopolist is earning profits of $200.  If there is successful entry by an equally efficient entrant, each of the two firms will earn duopoly profits of $70.  (The duopoly profits are less than monopoly profits because of the price competition.)  Suppose that the entrant needs to obtain just non-exclusive distribution from a particular retailer in order to be viable.  In this case, the entrant would be willing to bid up to $70 per period for the non-exclusive distribution.  (In price terms, this would be a payment that led to the entrant’s costs equaling its price.)  But the monopolist would be willing to bid up to $130 for an exclusive (i.e., the difference between its monopoly and duopoly profits), in order to prevent the entrant from surviving.   Thus, the monopolist would win the bidding, say for a price of $71.   The monopolist would easily pass the price-cost test.   Why is the monopolist systematically able to outbid the entrant? This fundamental asymmetry does not arise because the entrant is less efficient.  Instead, the answer is that the monopolist is bidding to maintain its monopoly power, whereas the entrant can only obtain duopoly price.  The monopolist is “purchasing market power” in addition to distribution, whereas the entrant is only purchasing distribution.

(b) Or, consider this interesting variant with sequential bidding for multiple distributors.   Suppose there are two retailers and the entrant needs to get non-exclusive distribution at both in order to be viable.  Suppose that the negotiations at the two stores are sequential.  In this scenario, the entrant would have no incentive even to try to outbid the monopolist.   This is easy to see.   Suppose that the entrant wins the competition to get into the first store by paying the amount $B1.   In bidding for distribution at the second retailer, the monopolist would be willing to bid up to $130, as above.   At this second store, the entrant would not be willing to pay more than $70 (or $70 – $B1, if it is ignores the fact that the $B1 was an already sunk cost).  So the monopolist will win the exclusive at the second retailer and the entry will fail.   Looking back to the negotiations at the first store, the entrant would have had no incentive to throw away money by paying any positive amount $B1 to get distribution at the first store.   This is because it rationally would anticipate that it is inevitable that it will fail to gain distribution at the second retailer.  Thus, the monopolist will be able to gain the exclusive at both stores for next to nothing.   It clearly will pass the price-cost test even as it maintains its monopoly, merely by instituting the competition for distribution.

(c) If the entrant only needs to gain non-exclusive distribution at either one of the two stores, then the situation can be reversed and the entry can succeed.   The monopolist clearly would not be willing to pay $71 each at both stores (equal to a total payment of $142) in order to deter the entry and protect its “incremental” monopoly profits (equal to only $130 in the example).  Therefore, when the entrant bids for distribution at the first store, the monopolist might as well let the entrant win, which means that the entrant can gain access to both stores for next to nothing.   The entry succeeds, but again, the price-cost test would not be relevant to the analysis.

(d) There also can be elements of a “self-fulfilling equilibrium” because of lack of coordination by the distributors.  Suppose that there are 10 retailers and the entrant only needs to get distribution at 5 of them.   Suppose that the entrant offers to pay a $14 rebate for non-exclusive distribution, and it also will offer $14 again in the next period, if its entry succeeds in the first period.   Suppose the monopolist offers a lower rebate for an exclusive that will continue into the second period.   Suppose that each of the 10 retailers anticipates that the other retailers will accept the monopolist’s lower offer out of fear that the entrant will be unable to get 4 other retailers to accept its offer.  In that situation, the entry will fail.  This is not because the entrant is less efficient.  Instead, it is because the entrant faces a classic coordination problem.  If the retailers behave independently, the retailers’ fear of the entrant’s failure can be a self-fulfilling prophecy.   Again, the monopolist will easily pass the price-cost test.

(4) Dan makes the point that the price-cost test does not require adoption of an EEC antitrust standard (i.e., whereby only harm to EECs is relevant to antitrust).  I certainly agree that the price-cost screen does not necessarily rely on the EEC standard.   The price-cost test is better framed as a measure of “profit-sacrifice,” and EEC is simply a misleading way to express the test.  For example, I expect that Dan agrees that predatory pricing law uses the price-cost test as a measure of “profit-sacrifice,” not an assumption that only EECs matter.

(5) But, I was surprised that Dan also says that the EEC theory “has merit.”  In my view, the EEC standard has no merit in rigorous antitrust analysis. The example in my previous post illustrates why that is the case.  Raising the costs and possibly deterring the entry of a less efficient rival harms consumers and reduces output.

(6) Dan says that the “disloyalty penalty” price theory has problems, “including the empirical one that it doesn’t fit the pattern of almost any of the recent loyalty discount cases.”   The validity of Dan’s empirical claim is not obvious clear to me.  To evaluate whether there is a price penalty, you would need to know more than the path of prices over time.  You also would need to know what the price would be in the “but-for world.”   For example, suppose that in the absence of the loyalty discount, the incumbent would have reduced its price to $90.  This observation has two important implications.  First, this is a reason why it is not clear that loyalty discounts are “presumptively beneficial.”  Second, this is another reason why a price-cost test is not a good “screen” in loyalty discount cases.  Implementing the screen involves evaluating what prices would be absent the conduct.  But, after the competitive effects on consumers are known, what is the value of the screen?

(7) As to the question of whether Josh’s speech on loyalty discounts (and this issue of penalty prices) is inconsistent with their joint article on bundled discounts, I will leave that one for Josh and Dan to sort out, at least for the moment.   I certainly will concede the point that Wright is not always right.

(8) Dan began to suggest that the penalty price theory has a “problem of basic economics” in that the penalty price was not short-run profit-maximizing.   Dan subsequently seemed to withdraw this criticism, noticing that one could characterize the loyalty restriction as not profit-maximizing in the same way.   In any event, it is not a “problem” with the theory.  The reason why the firm is willing to sacrifice profits is because it gains the benefit of deterring entry.  By the way, it also may not even end up sacrificing profits.  The threat of the penalty price for non-exclusivity may be sufficient.  If the distributors succumb to the threat and buy exclusively from the incumbent, it never needs to actually charge them the penalty price.

Guest post by Steve Salop, responding to Dan’s post and Thom’s post on the appropriate liability rule for loyalty discounts.

I want to clarify some of the key issues in Commissioner Wright’s analysis of Exclusive Dealing and Loyalty Discounts as part of the raising rivals’ costs (“RRC”) paradigm. I never thought that I would have to defend Wright against Professors Lambert and Crane. But, it appears that rigorous antitrust analysis sometimes makes what some would view as strange bedfellows.

In my view, there should not be a safe harbor price-cost test used for loyalty discounts. Nor should these discounts be treated as conclusively (per se) illegal if the defendant fails the price-cost test. Either way, the test is a formalistic and unreliable screen. To explain these conclusions, and why I think the proponents of the screen are taking too narrow approach to these issues, I want to start with some discussion of the legal and economic frameworks.

In my view, there are two overarching antitrust legal paradigms for exclusionary conduct – predatory pricing and raising rivals’ costs (RRC), and conduct that falls into the RRC paradigm generally raises greater antitrust concerns. (For further details, see my 2006 Antitrust L.J. article, “Exclusionary Conduct, Effect on Consumers, and the Flawed Profit-Sacrifice Standard.”) Commissioner Wright also takes this approach in his speech of identifying and distinguishing the two paradigms.

This raises the question of which framework is better suited for addressing exclusive dealing and loyalty discounts (that is, where the conduct is not pled in the complaint as predatory pricing). Commissioner Wright’s speech articulates the view that theories of harm alleging RRC/foreclosure should be analyzed under exclusive dealing law, which is more consistent with the raising rivals’ costs approach, not under predatory pricing law (i.e., with its safe harbor for prices above cost). (Incidentally, I don’t read his speech as saying that he has abandoned Brooke Group for predatory pricing allegations. For example, it seems clear that he would support a price-cost test in a case alleging that a loyalty discount harmed competition via predatory pricing rather than RRC/foreclosure.)

To understand which legal framework – raising rivals’ costs/exclusive dealing versus predatory pricing/price-cost test – is most relevant for analyzing the relevant competitive issues, I want to begin with a primer on RRC theories of foreclosure. This will also hopefully bring everyone closer on the economics.

Input Foreclosure and Customer Foreclosure

There are two types of foreclosure theories within the RRC paradigm — “input foreclosure” and “customer foreclosure.” Both are relevant for evaluating exclusive dealing and loyalty discounts. The input foreclosure theory says that the ED literally “raises rivals’ costs” by foreclosing a rival’s access to a critical input subject to ED. The customer foreclosure theory says that ED literally “reduces rivals’ revenues” by foreclosing a rival’s access to a sufficient customer base and thereby drives the rival out of business or marginalizes it as a competitor (i.e., where it lacks the ability or incentive to move effectively beyond a niche position or to invest to grow).

Commissioner Wright’s speech tended to merge the two variants. But, it is useful to distinguish between them. (I think that this is one source of Professor Lambert being “baffled” by the speech, and more generally, is a source of confusion among commentators that leads to unnecessary disagreements.)

In the simplest presentation, one might say that customer foreclosure concerns are raised primarily by exclusive dealing with customers, while input foreclosure concerns are raised primarily by exclusive dealing with input suppliers. But, as noted below, both concerns may arise in the same case, and especially so where the “customers” are distributors rather than final consumers, and the “input” is distribution services.

Analysis of exclusive dealing (ED) often invokes the customer foreclosure theory. For example, Lorain Journal may be analyzed as customer foreclosure. However, input foreclosure is also highly relevant for analyzing ED because exclusive dealing often involves inputs. For example, Judge Posner’s famous JTC Petroleum cartel opinion can be interpreted in this way, if there were solely vertical agreements.

Cases where manufacturers have ED arrangements with wholesale or retail distributors might be thought to fall into the customer foreclosure theory because the distributors can be seen as customers of the manufacturer. However, distributors also can be seen as providing an input to the manufacturer, “distribution services.” For example, a supermarket or drug store provides shelf space to a manufacturer. If the manufacturer (say, unilaterally) sets resale prices, then the difference between this resale price and the wholesale price is the effective input price.

One reason why the input foreclosure/customer foreclosure distinction is important involves the proper roles of minimum viable scale (MVS) and minimum efficient scale (MES). The customer foreclosure theory may involve a claim that the rival likely will be driven below MVS and exit Or it may involve a claim that the rival will be driven below MES, where its costs will be so much higher or its demand so much lower that it will be marginalized as a competitor.

By contrast, and this is the key point, input foreclosure does not focus on whether the rival likely will be driven below MVS. Even if the rival remains viable, if its costs are higher, it will be led to raise the prices charged to consumers, which will cause consumer harm. And prices will not be raised only in the future. The recoupment can be simultaneous.

Another reason for the importance of the distinction is the role of the “foreclosure rate,” which often is the focus in customer foreclosure analysis. For input foreclosure, the key foreclosure issue is not the fraction of distribution input suppliers or capacity that is foreclosed, but rather whether the foreclosure will raise the rival’s distribution costs. That can occur even if a single distributor is foreclosed, if the exclusivity changes the market structure in the input market or if that distributor was otherwise critical. (For example, see Krattenmaker and Salop, “Anticompetitive Exclusion.”)

At the same time, it is important to note that the input/customer foreclosure distinction is not a totally bright line difference in many real world cases. A given case can raise both concerns. In addition, customer foreclosure sometimes can raise rivals costs, and input foreclosure sometimes (but not always) can cause exit.
While input foreclosure can succeed even if the rival remains viable in the market, in more extreme scenarios, significantly higher costs inflicted on the rival could drive the rival to fall below minimum viable scale, and thereby cause it to exit. I think that this is one way in which unnecessary disagreements have occurred. Commentators might erroneously focus only this more extreme scenario and overlook the impact of the exclusives or near-exclusives on the rival’s distribution costs.

Note also that customer foreclosure can raise a rival’s costs when there are economies of scale in variable costs. For this reason, even if the rival does not exit or is not marginalized, it nonetheless may become a weaker competitor as a result of the exclusivity or loyalty discount.

These points also help to explain why neither a price-cost test nor the foreclosure rate will provide sufficient reliable evidence for either customer foreclosure or input foreclosure, which I turn to next.

(For further discussion of the distinction between input foreclosure and customer foreclosure, see Riordan and Salop, Evaluating Vertical Mergers: A Post-Chicago Approach, 63 ANTITRUST L.R. 513(1995). See also the note on O’Neill v Coca Cola in Andrew Gavil, William Kovacic and Jonathan Baker, Antitrust Law in Perspective: Cases, Concepts and Problems in Competition Policy (2d ed.) at 868-69. For analysis of Lorain Journal as customer foreclosure, see Gavil et. al at 593-97.)

The Inappropriateness of a Dispositive Price-Cost Test

A price-cost test obviously is not relevant for evaluating input foreclosure concerns, even where the input is distribution services. Even if the foreclosure involves bidding up the price of the input, it can succeed in permitting the firm to achieve or maintain market power, despite the fact that the firm does not bid to the point that its costs exceed its price. (In this regard, Weyerhaeuser was a case of “predatory overbuying,” not “raising rivals’ cost overbuying.” The allegation was that Weyerhaeuser would gain market power in the timber input market, not the lumber output market.)

Nor is a price-cost test the critical focus for assessing customer foreclosure theories of competitive harm. (By the way, I think we all agree that the relevant price-cost test involves a comparison of the incremental revenue and incremental cost of the “contestable volume” at issue for the loyalty discount. So I will not delve into that issue.)

First, and most fundamentally, the price-cost test is premised on the erroneous idea that only equally efficient competitors are worth protecting. In other words, the price-cost requires the premise that the antitrust laws only protect consumers against competitive harm arising from conduct that could have excluded an equally efficient competitor. This premise makes absolutely no economic sense. One simple illustrative example is a monopolist raising the costs of a less efficient potential competitor to destroy its entry into the market. Suppose that monopolist has marginal cost of $50 and a monopoly price of $100. Suppose that there is the potential entrant has costs of $75. If the entry were to occur, the market price would fall. Entry of the less efficient rival imposes a competitive constraint on the monopolist. Thus, the entry clearly would benefit consumers. (And, it clearly often would raise total welfare as well.) It is hard to see why antitrust should permit this type of exclusionary conduct.

It is also unlikely that antitrust law would allow this conduct. For example, Lorain Journal is probably pretty close to this hypothetical. WEOL likely was not equally efficient. The hypothetical probably also fits Microsoft pretty well.

Second, the price-cost test does not make economic sense in the case of the equally efficient rival either. Even if the competitor is equally efficient, bidding for exclusives or near-exclusives through loyalty discounts often does not take place on a level playing field. There are several reasons for this. One reason is that the dominant firm may tie up customers or input providers before the competitors even arrive on the scene or are in a position to counterbid. A second reason is that the exclusive may be worth more to the dominant firm because it will allow it to maintain market power, whereas the entrant would only be able to obtain more competitive profits. In this sense, the dominant firm is “purchasing market power” as well as purchasing distribution. (This point is straightforward to explain with an example. Suppose that the dominant firm is earning monopoly profits of $200, which would be maintained if it deters the entry of the new competitor. Suppose that successful entry by the equally efficient competitor would lead to the dominant firm and the entrant both earning profits of $70. In this example, the entrant would be unwilling to bid more than $70 for the distribution. But, the dominant firm would be willing to bid up to $130, the difference between its monopoly profits of $200 and the duopoly profits of $70.) A third reason is that customers may not be willing to take the risk that the entry will fail, where failure can occur not because the entrant’s product is inferior but simply because other customers take the exclusive deal from the dominant firm. In this case, a fear that the entrant would fail could become a self-fulfilling prophecy because the customers cannot coordinate their responses to the dominant firms’ offer. Lorain Journal may provide an illustrative example of this self-fulfilling prophecy phenomenon. This last point highlights a more general point Commissioner Wright made in his speech — that successful and harmful RRC does not require a below-cost price (net of discounts). When distributors cannot coordinate their responses to the dominant firm’s offer, a relatively small discount might be all that is required to purchase exclusion. Thus, while large discounts might accompany RRC conduct, that need not be the case. These latter reasons also explain why there can be successful foreclosure even when contracts have short duration.

Third, as noted above, customer foreclosure may raise rivals’ costs when there are economies of scale. The higher costs of the foreclosed rivals are not well accounted for by the price-cost test.

Fourth, as stressed by Joe Farrell, the price-cost test ignores the fact that loyalty discounts triggered by market share may deter a customer’s purchases from a rival that do not even come at the expense of the dominant firm. (For example, suppose in light of the discounts, the customer is purchasing 90 units from the dominant firm and 10 from the rival in order to achieve a “reward” that comes from purchasing 90% from the dominant firm. Now suppose that entrant offers a new product that would lead the customer to wish to continue to purchase 90 units from the dominant firm but now purchase 15 units from the rival. The purchase of these additional 5 units from the rival does not come at the expense of the dominant firm. Yet, even if the entrant were to offer the 5 units at cost, these purchases would be deterred because the customer would fall below the 90% trigger for the reward.) In this way, the market share discount can directly reduce output.

Fifth, the price-cost test assumes that the price decreases will be passed on to final consumers. This may be the clear where the exclusives or loyalty discounts are true discounts given to final consumers. But, it may not be the case where the dominant firm is acquiring the loyalty from input suppliers, including distributors who then resell to final consumers. The loyalty discounts often involve lump sum payments, which raises questions about pass-on, at least in the short-run.

Finally, it is important to stress that the price-cost test for loyalty discounts assumes that price actually represents a true discount. I expect that this assumption is the starting point for commentators who give priority to the price-cost test. However, the price may not represent a true discount in fact, or the size of the discount may turn out to be smaller than it appears after the “but-for world” is evaluated. That is, the proponents of a price-cost test have the following type of scenario in mind. The dominant firm is initially charging the monopoly price of $100. In the face of competition, the dominant firm offers a lower price of (say) $95 to customers that will accept exclusivity, and the customers accept the exclusivity in order to obtain the $5 discount. (To illustrate, suppose that absent the exclusive, the customers would purchase 90 units from the dominant firm at $100 for total revenue of $9000. With the exclusive, they purchase 100 units at a price of $95 for total revenue of $9500.
Thus, the dominant firm earns incremental revenue of $500 on the 10 incremental units, or $50 per unit. If the dominant firm’s costs are $50 or less, it will pass the price-cost test.) But, consider next the following alternative scenario. The dominant firm offers the original $100 price to those customers that will accept exclusivity, and sets a higher “penalty” price of $105 to customers that purchase non-exclusively from the competitor. In this latter scenario, the $5 discount similarly may drive customers to accept the exclusive. These prices would lead to a similar outcome of the price-cost test. (To illustrate, suppose that absent the exclusive, the customers would purchase 90 units from the dominant firm at $105 for total revenue of $9450. With the exclusive, they purchase 100 units at a price of $100 for total revenue of $10,000. Thus, the dominant firm earns revenue of $550 on the 10 incremental units, or $55 per unit. Here, the dominant firm will pass the price-cost test, if its costs are $55 or less.) However, in this latter scenario, it is noteworthy that the use of the “penalty” price eliminates any benefits to consumers. This issue seems to be overlooked by Crane and Lambert. (For further details of the role of the penalty price in the context of bundled discounts, see Barry Nalebuff’s articles on Exclusionary Bundling and the articles of Greenlee, Reitman and Sibley.)

* * *

For all these reasons, treating loyalty discounts as analogous to predatory pricing and thereby placing over-reliance on a price-cost test represents a formalistic and unreliable antitrust approach. (It is ironic that Commissioner Wright was criticized by Professor Lambert for being formalistic, when the facts are the opposite.)

This analysis is not to say that the court should be indifferent to the lower prices, where there is a true discount. To the contrary, lower prices passed-on would represent procompetitive efficiency benefits. But, the potential for lower prices passed-on does not provide a sufficient basis for adopting a price-cost safe harbor test for loyalty discount allegations, even ones that can be confidently characterized as purely plain vanilla customer foreclosure with no effects on rivals’ costs.

Thus, the price-cost test should be one relevant evidentiary factor. But, it should not be the primary factor or a trump for either side. That is, above-cost pricing (measured in terms of incremental revenue less than incremental cost) should not be sufficient by itself for the defendant to escape liability. Nor should below-cost pricing (again, measured in terms of incremental revenue less than incremental cost) should not be a sufficient by itself for a finding of liability.

Such “Creeping Brookism” does not led to either rigorous or accurate antitrust analysis. It is a path to higher error rates, not a lower ones.

Nor should courts rely on simple-minded foreclosure rates. Gilbarco shows how a mechanical approach to measuring foreclosure leads to confusion. Microsoft makes it clear that a “total foreclosure” test also is deficient. Instead, a better approach is to require the plaintiff to prove under the Rule of Reason standard that the conduct harms the rival by reducing its ability to compete and also that it harms consumers.

I should add one other point for completeness. Some (but not Commissioner Wright or Professor Crane) might suggest that the price-cost test has administrability benefits relative to a full rule of reason analysis under the RRC paradigm. While courts are capable are evaluating prices and costs, that comparison may be more difficult than measuring the increase in the rivals’ distribution costs engendered by the conduct. Moreover, the price-cost comparison becomes an order of magnitude more complex in loyalty discount cases, relative to plain vanilla predatory pricing cases. This is because it also is necessary to determine a reasonable measure of the contestable volume to use to compare incremental revenue and incremental cost. For first-dollar discounts, there will always be some small region where incremental revenue is below incremental cost. Even aside from this situation, the two sides often will disagree about the magnitude of the volume that was at issue.

In summary, I think that Professor Wright’s speech forms the basis of moving the discussion forward into analysis of the actual evidence of benefits and harms, rather than continuing to fight the battles over whether the legal analysis used in the 1950s and 1960s failed to satisfy modern standards and thereby needed to be reined in with unreliable safe harbors.

One of the most controversial merger policy decisions during the Bush administration was the DOJ’s failure to bring a complaint against the Whirlpool/Maytag merger.  Indeed, the decision was even criticized by Carl Shapiro, the economic expert retained by the DOJ on the case.   Jonathan Baker and Carl Shapiro summarize this conclusion as follows:  “The March 2006 decision by the DOJ not to challenge Whirlpool’s acquisition of Maytag was a highly visible instance of underenforcement.”   Orley Ashenfelter, Daniel Hosken and Matthew Weinberg have now posted a working paper that estimates the price effects of the merger.  Using scanner data, the authors compare the prices of Whirlpool and Maytag appliances to price changes in the appliance markets most affected by the merger to other markets less or not affected.  They find large significant price increases for clothes dryers and dishwashers, but not for refrigerators and washing machine.

A Washington Post editorial last week reached the surprising conclusion that a series of vertical and horizontal acquisitions that led to a firm owning about 40% of the gas stations in the District of Columbia was procompetitive.  The editorial apparently concluded that the vertical integration efficiencies were more important than the adverse horizontal effects.  The editorial cited to an FTC report on the efficiencies of vertical integration.  This is a very counterintuitive conclusion for an allegedly liberal newspaper.   Even more counterintuitive, however, is the fact that the editorial also reports the results of a natural experiment that concluded that prices have risen as a result of the acquisitions.  The gap between prices in DC and Maryland/Virginia rose from 10 cents to 17 cents.  According to the numbers in the editorial, tax increases account for about half of the gap.  Does this mean that WAPO has bought on to the Aggregate Welfare Standard over the Consumer Welfare Standard?  Or, is this just one more example of skillful advocacy by the gas station owner combined with poor understanding of economics by the editorial board?  I vote for the latter.  Do others disagree?

Merger Retrospective

Steve Salop —  4 April 2011

Several years ago, the DOJ cleared a merger between Whirlpool and Maytag.   The primary defense was that post-merger prices could not rise because of intense competition from foreign competitors like LG and Samsung. Apparently the actual competition was more than Whirlpool wanted to bear.  Guess What?  Mr. Laissez-Faire Antitrust, meet Dr. Public Choice.  The Wall Street Journal has reported that Whirlpool has filed a dumping complaint against LG and Samsung.   Whirlpool’s dumping complaint involves refrigerators and the merger concerns involved washers and dryers more than refrigerators.  But, the complaint sends a signal to LG and Samsung.  The comlaint also certainly does raise a caution about relying on foreign competition, and suggests a potential remedial provision.

Smoothing Demand Kinks

Steve Salop —  4 April 2011

One criticism of the unilateral effects analysis in the 2010 Merger Guidelines is that demand curves are kinked at the current price.  A small increase in price will dramatically reduce the quantity demanded.  One rationale for the kink is that people over-react to small price changes and dramatically reduce demand.  As a result of this behavioral economics deviation from standard rational behavior, it is claimed, merging firms will not raise prices when the merger increases the opportunity cost of increasing output.  (The opportunity cost increases because some of the increased output now comes from the new merger partner.)  It has been argued that such kinks are ubiquitous, whatever the current price is.  For some recent views on this issue, see the recent anti-kink article by Werden and the pro-kink reply by Scheffman and Simons.

A story in today’s New York Times nicely illustrates one of the problems with the kinked demand story.  Instead of raising prices, consumer products firms can and commonly do raise per unit prices by reducing package sizes.  Changes in package sizes do not create a disproportionate reaction, perhaps because they are less visible to busy shoppers.   Whatever the reason, this smaller package size raises the effective price per unit while avoiding the behavioral economics kink.  Of course, this is not to say that firms never raise prices; they do.  Moreover, even a kink did exist for reasons grounded in behavioral economics or menu costs, any kink likely is just temporary.  In contrast, a merger is permanent.

It is for these reasons that this kinked economics has gotten much traction in the current debate.  But, these presumptions do not mean that kinked economics arguments can never be raised in a merger.  If there were evidence of a low pass-through rate of variable cost into higher prices over a significant period of time, that evidence would be relevant to a more refined analysis of upward pricing pressure.


The proposed Horizontal Merger Guidelines (HMGs) have been treated by some as a major shift in enforcement approach away from a tight structure that begins with market definition to a more flexible and open-ended competitive effects approach.  Some of the specific concerns that have been raised are that the proposed HMGs dramatically change enforcement policy by (i) downplaying or ignoring the role of market definition in favor of a more holistic approach to competitive effects; (ii) defining overly narrow relevant markets; (iii) defining whatever market “works” to show anticompetitive effect; (iv) ignoring market definition and concentration in unilateral effects analysis; (v) creating a presumption of anticompetitive unilateral effects in every merger with differentiated products, and failing to have a safe harbor; (vi) fearing unilateral effects among firms that are not closest substitutes; (vii) overemphasizing diversion ratios among the merging parties and ignoring competition from non-merging products; and (viii) ignoring the competitive constraints from the potential for repositioning.

As Joe Sims has observed, “elections matter,” so a change in enforcement policy would not be all that surprising.  It obviously is too soon to tell whether and by how much enforcement will change in practice.  But, to check on the extent to which the language of the proposed HMGs represents a shift, I went back to the 2006 Merger Commentary to see what the previous administration said about these issues.  I assign the Merger Commentary to my advanced antitrust students.  But, this report is often ignored.  I have reproduced some quotations from the 2006 Merger Commentary below.  Readers can judge for themselves the extent to which elections matter for how the Agencies frame their analysis.

*          *          *          *          *

  • Downplaying or ignoring the role of market definition in favor of a more holistic approach to competitive effects

“At the center of the Agencies’ application of the Guidelines, therefore, is competitive effects analysis. That inquiry directly addresses the key question that the Agencies must answer: Is the merger under review likely substantially to lessen competition?” (Page 2)

“In some investigations, before having determined the relevant market boundaries, the Agencies may have evidence that more directly answers the ‘ultimate inquiry in merger analysis,’ i.e., ‘whether the merger is likely to create or enhance market power or facilitate its exercise.’ … Evidence pointing directly toward competitive effects may arise from statistical analysis of price and quantity data related to, among other things, incumbent responses to prior events (sometimes called ‘natural experiments’) such as entry or exit by rivals.” (Page 10)

“Evidence pertaining more directly to a merger’s actual or likely competitive effects also may be useful in determining the relevant market in which effects are likely.” (Page 10)

“Application of the Guidelines as an integrated whole to case-specific facts–not undue emphasis on market share and concentration statistics–determines whether the Agency will challenge a particular merger.” (Pages 15-16)

  • Defining overly narrow relevant markets

“Defining markets under the Guidelines’ method does not necessarily result in markets that include the full range of functional substitutes from which customers choose.”  (Page 6)

“The Agencies frequently conclude that a relatively narrow range of products or geographic space within a larger group describes the competitive arena within which significant anticompetitive effects are possible.” (Page 6)

“The description of an ‘antitrust market’ sometimes requires several qualifying words and as such does not reflect common business usage of the word ‘market.’ Antitrust markets are entirely appropriate to the extent that they realistically describe the range of products and geographic areas within which a hypothetical monopolist would raise price significantly and in which a merger’s likely competitive effects would be felt.”  (Page 12)

“The boundaries of a market are less clear-cut in merger cases that involve products or geographic areas for which substitutes exist along a continuum. …. Even when no readily apparent gap exists in the chain of substitutes, drawing a market boundary within the chain may be entirely appropriate when a hypothetical monopolist over just a segment of the chain of substitutes would raise prices significantly.”  (Page 15)

  • Defining whatever market “works” to show anticompetitive effect

“This process [of defining markets] could lead to different conclusions about the relevant markets likely to experience competitive harm for two similar mergers within the same industry.”  (Page 12)

  • Ignoring market definition and concentration in unilateral effects analysis

“Indeed, market concentration may be unimportant under a unilateral effects theory of competitive harm. … The concentration of the remainder of the market often has little impact on the answer to that question.”  (Page 16)

  • Creating a presumption of anticompetitive unilateral effects in every merger with differentiated products, and failing to have a safe harbor

“Section 2.2 of the Guidelines does not establish a special safe harbor applicable to the Agencies’ consideration of possible unilateral effects. Section 2.2.1 provides that significant unilateral effects are likely with differentiated products when the combined market share of the merging firms exceeds 35% and other market characteristics indicate that market share is a reasonable proxy for the relative appeal of the merging products as second choices as well as first choices.” (Page 26)

  • Overemphasizing diversion ratios among the merging parties and ignoring competition from non-merging products

“For mergers involving differentiated products, the “diversion ratios” between products combined by the merger are of particular importance. … In general, for any two products brought under common control by a transaction, the higher the diversion ratios, the more likely is significant harm to competition.”  (Page 27)

The unilateral effects of a merger of differentiated consumer products are largely determined by the diversion ratios between pairs of products combined by the merger, and the diversion ratios between those products and the products of non-merging firms have at most a secondary effect.” (Page 28)

“A merger may produce significant unilateral effects even though a large majority of the substitution away from each merging product goes to non-merging products. … A merger may produce significant unilateral effects even though a non-merging product is the “closest” substitute for every merging product.”  (Pages 27-28)

  • Ignoring the competitive constraints from the potential for repositioning

“The Agencies rarely find evidence that repositioning would be sufficient to prevent or reverse what otherwise would be significant anticompetitive unilateral effects from a differentiated products merger.” (Page 31)

Remember: These quotations are taken from the 2006 Merger Commentaries, not the proposed HMGs!

Of course, the Merger Guidelines need to be updated.  Except for efficiencies, they haven’t been updated in 17 years.   Lawyers and economists with a regular antitrust practice may not require an update in light of their knowledge of the 2006 Commentary, speeches and agency experience.  But, the rest of the antitrust world does.  The most obvious audience is the courts, who should know what the agencies believe is best practice.  Moreover, as FTC Commissioner Kovacic has stressed, the world marketplace for antitrust ideas needs to have the guidance of the US enforcement agencies.  They should not have to ferret it out from commentaries and speeches.  The Merger Guidelines have been the most emulated feature of US antitrust enforcement worldwide.  It would be shame to squander the leadership role.

The “most important” revision is harder to identify.  I am caught here between market definition and unilateral effects – and, clearly, revisions to either issue will have important implications for how the agencies and practitioners approach the other issue.  Of course, coordinated effects and entry could use some work too.  But, let me stick just with the first two and just a sampling of issues there.

The hypothetical monopolist SSNIP test is an elegant but complicated and imperfect methodology.   I think it should be explicitly conceded in the Guidelines that market definition is a very imperfect exercise in practice in which the “appropriate” market definition is often controversial and there may not convincing evidence for a single “most appropriate” market.  That concession would continue the process of downgrading the perceived importance of market definition in antitrust and focusing the courts more on the bottom line issue of competitive effects.  After all, market definition is mainly valued for helping to analyze competitive effects.  The imperfections of market definition and the implications for merger analysis have been made by Katz & Shelanski and should be made explicit in the update.   In fact, I’d like the Guidelines to start with a “first principles” overview of competitive effects and fit market definition into that framework, not the reverse.

The SSNIP test definitely needs some renovation after all these years.  There is considerable controversy today over the proper way to implement the test. For example, standard critical loss analysis often used by economic experts in testimony generally focuses just on the profitability of a SSNIP, whereas the Guidelines focus on whether a SSNIP is profit-maximizing.  It also may ignore relevant information.  Articles by Katz & Shapiro and O’Brien & Wickelgren (“KSOW”) have proposed a methodology that takes into account the information about demand elasticity contained in the price/cost margins of profit-maximizing Bertrand competitors.  This is a step in the right direction.

This KSOW methodology does tend to lead to narrower markets than often are found now.  So, the agencies need to decide whether this is the desired outcome or whether they should raise the SSNIP level, or downgrade the importance of market shares and concentration, to compensate.

The “smallest market principle” should be erased.  First, the SMP erroneously suggests that a careful application of the ssnip test will identify one and only one relevant market.   That outcome is not the case in practice.  Depending on the starting point (and there is no unique starting point), the algorithm does not lead to a unique “smallest” market.  And, once price discrimination is taken into account (in a negotiated price market, for example) the smallest market may well be a single customer.

The principle also makes no sense as a matter of policy.  Suppose there were a narrow market for premium baby food comprised of just Gerber, Beechnut, and organic brands.  That finding should not imply that a merger between Gerber and Heinz necessarily would be free of significant competitive issues.

Moving on to unilateral effects, it is clear that the Unilateral Effects section also needs work.  One need go no further than the garbled description of the impact of a 35% combined market share.  After this part is revised and clarified, I will mourn the lost opportunity of reading it to my students.  What a fabulous example of drafting by a committee in conflict!  But, the offsetting benefit is that my students, the courts and others may actually understand the enforcement intentions of the agencies, including the issue of whether or not there is a safe harbor.

The Unilateral Effects section also needs to clarify the role of the repositioning.  The 2006 Merger Commentary suggests that repositioning is seldom, if ever, sufficient to eliminate concerns.  Coupled with tight agency standards towards cognizable efficiency benefits, this has led to a longstanding concern among commentators that virtually every differentiated products merger could be said to raise a significant danger of adverse unilateral effects.  I think that this is mainly an evidence issue.   If there has been repositioning in the pre-merger market, claims of a post-merger repositioning constraint are more credible.   Otherwise it needs to be explained why repositioning would not been observed as the market has been evolving over time.

More attention also must be paid to gauging the magnitude of the predicted unilateral effects, both at the screening stage and in the ultimate inquiry.  Greg Werden, O’Brien & Salop, and Farrell & Shapiro each have proposed variants of “upward price pressure indices” (“UPPIs”).  These UPPIs rely on information about diversion ratios, margins and efficiency benefits.  They could be made an explicit part of the analysis.  They also could form the basis for an initial screen that would make more economic sense for unilateral effects than the HHI, which at best is related to coordinated effects.  I have heard criticism that the UPPIs cannot be derived rigorously from a general oligopoly model without making certain additional limiting assumptions.  True enough, but neither can the HHI!

Formulating an initial screen or safe harbor UPPI will take some work.  Should there be an explicit “efficiency credit?” Or, should the agencies simply set a critical level for the UPPI, based on expertise, experience and policy preferences?  (In this regard, the HHI thresholds of 1000 and 1800, and the Delta-HHI threshold of 100 were never really justified.)  If there is a formal credit, how should it be formulated – as a fraction of cost or price or both?  Should the efficiency credit be a “standard deduction” or a “personal exemption?” That is, should cognizable efficiencies proven by the parties be added to the credit (as with a personal exemption), or replace the credit (as itemized deductions replace the standard deduction)?

And, how should the efficiency credit take into account optimal deterrence, including the goal of allowing well-functioning markets for the sale of property and corporate control, two general efficiencies noted in the previous versions of the Guidelines?  In addition, what does “incipiency” mean in today’s world — does it mean that the agencies and courts should set standards tilted more towards a concern with false negatives?  In my view, optimal deterrence is the most important issue of all, but clearly one of the hardest and most controversial.

The UPPIs also need to be connected up to market definition.  The UPPI is a close cousin of the KSOW form of critical loss analysis.  Indeed, this connection serves as a reminder that market definition generally has reduced importance in unilateral effects.  This does not mean that market definition is irrelevant.  Market definition analysis provides a focus on closeness of substitutes, which is central to unilateral effects analysis.  The importance of the cross-elasticity of demand was even noted by the Court in its notorious DuPont decision.  Market definition also is relevant for coordinated effects – to define a possible coordinating group of firms.  And, of course, Section 7 makes it clear that a market must be defined.  So, it is clear that market definition cannot be dispensed with, but only downgraded in importance.

In response to my first post on joint monopsony conduct to countervail monopoly power, Mike Ward raises the issue of justifying a merger among sellers on the basis that it will countervail alleged monopsony power.  Labor unions have an antitrust exemption for just that purpose.   In terms of merger policy, Tom Campbell has written an article proposing just such a defense for merger to monopoly (74 Antitrust Law Journal 521 (2007)).  His article was strongly criticized in a Comment by Jon Baker, Joe Farrell and Carl Shapiro.

I’ll stick to my knitting and take a pass on Geoff’s and Josh’s questions about when antitrust exemptions are appropriate.   Josh framed the question in an interesting way.  We usually think that courts do a better job than the legislature on antitrust issues.   So, when does it make sense to support antitrust exemptions or (for that matter) laws such as a Leegin-repealer.  I’m looking forward to hearing others’ comments on this issue.

Regarding the appropriate welfare standard for antitrust, I’ve been an advocate of the consumer welfare standard.  I also think the Courts have generally favored consumer welfare by placing their main focus on price and output effects.

Regarding the merchants and the Visa/MasterCard issuers, I have never thought about the issue in Coasian property rights terms, but I’ll give it an initial try, with the caveat that a thorough analysis clearly requires more work.  I think that there are market failures here – transaction costs in Coasian lingo – that imply that one cannot count on a laissez faire market for the efficient outcome.

First, Visa and MasterCard arguably are bank card issuer “cartels” that jointly fix the price that merchants pay for acceptance of their cards and have a collectively dominant combined market share.  The membership overlap in the two networks is very high.  Network effects would seem to create barriers to entry.   Regarding this type of “strategic behavior” market failure, Visa and MasterCard also were found liable for anticompetitive exclusionary conduct towards competing credit card companies, American Express and Discover Card.  They also agreed to a multi-billion dollar settlement of allegations that they illegally tied the acceptance credit cards and debit cards to merchants.  (For the record, I consulted for the plaintiffs in this matter.)

Second, the high interchange fees are not “simply a transfer” from merchants to the banks.  They likely raise the price of the merchandise purchased by consumers.  Few merchants surcharge the use of credit cards.  There are transactions costs to inform consumers of the extra set of prices, and I understand that Visa and MasterCard rules may well prohibit such surcharges.  These higher merchandise prices likely are partially — but not nearly fully — offset by the banks redistributing the cartel profits back to the cardholders through lower annual fees, rewards programs or other features.   It appears that a possibly significant fraction of the profits likely have been dissipated in marketing — classic rent-seeking expenditures.

Third, the increase in the price of merchandise likely also inflicts harm on non-users of credit cards, the customers who pay by cash and check.  This external effect would be seen in Coasian terms as a classic third-party effect, and the injured non-users would face coordination costs in attempting to undo the effect.

Finally, the fact that merchants are willing to pay the high interchange fees does not make it efficient.  Paying high fees for credit card acceptance would seem to be a prisoner’s dilemma for merchants.  I assume that Coasian theory applauds the formation of coalitions to coordinate an escape from the prisoner’s dilemma.

Geoff and Josh raise an interesting issue about collective market conduct by buyers.  Suppose that a group of final consumers face a monopolist.  Should the consumers be permitted to band together into an “association” to jointly negotiate a lower price from the monopolist?  Some would say that such buyer “cooperatives” are permitted, whereas others would say that such buyers “cartels” are antitrust violations.  Putting aside the terminology (which simply reflects the commentator’s conclusion rather than any real analysis), this hypothetical example raises a number of interesting antitrust questions.

First, if this buyer association involves final consumers and they move the market from the monopoly output to the competitive output, is that a “cognizable” efficiency under the antitrust laws, even if it involves what appears to be “naked” price fixing (i.e. price fixing not accompanied by any “integration” of production facilities or incentives)?  What should the law conclude?  What would Ronald Coase and John Nash say?

Second, should the antitrust outcome depend on how the incremental “total surplus” is split?  In particular, suppose that the association actually reduces the monopolist’s surplus, aside from whatever effects it has on total output.  Is that transfer of surplus from the monopolist to the consumers to be applauded (“antitrust is a consumer welfare prescription”), or attacked as counterproductive (“monopoly prices is what attracts business acumen”), or treated neutrally (“merely a transfer”)?

Third, suppose that the association moves the market from an inefficient monopoly output to an inefficient monopsony output, but the new outcome involves higher consumer surplus.  Should the increase in consumer surplus be a complete defense under the Sherman Act?  If the total surplus falls, should the conduct be illegal under the rule of reason? Or, should the law simply apply the per se rule to this concerted action.

Fourth, should the analysis be different if the seller is not a monopolist but rather a joint venture with monopoly power, or a group of (legally) tacitly coordinating sellers?  For example, in BMI, the Supreme Court placed considerable weight on the fact that the composers granted non-exclusive licenses to ASCAP and BMI, and the Court relied on the assumption that customers were able to constrain prices by negotiating directly with the composers.  However, suppose that it were assumed instead that such direct negotiation would be an ineffective constraint on prices.  In that scenario, should the music users be permitted to negotiate jointly with ASCAP and BMI over the terms of the blanket license?  Would permitting such a buyer association have been a better remedy than the price regulation remedy constrained in the consent decree with the Justice Department?

While he was an Appeals Court judge, Justice Breyer dealt with some of these issues in the Kartell case.  (If you don’t know the case, I promise that I have not made up the name.)  In Kartell, a group of dentists (including Dr. Kartell) wanted to block the arguably monopsony conduct of Blue Cross of Massachusetts.  As the case name might have pre-ordained, Blue Cross was exonerated.  Then-Judge Breyer’s reasoning was a bit peculiar, however,  He characterized Blue Cross as a single buyer, not as a buyer association. But, hedid not characterize Blue Cross as a reseller with market power, despite its dominant market share.  Instead, he dodged that issue by suggesting that Massachusetts insurance regulations somehow would force Blue Cross to pass on the fee reductions to consumers in the form of lower insurance prices.  Of course, this is the same result that would be achieved if Blue Cross were the cartel ringmaster for an insurance buyer cartel.

So, what do you make of this opinion?  In writing my article on “anticompetitive overbuying,” I raised the question of whether Judge Breyer was adopting the consumer welfare standard over a total welfare standard.  Or, was he simply relying on the kindness of regulation to replace market forces?  Or, did he make the classic error of assuming that monopsony is inherently good because it leads to a lower input price without noticing that it also reduces output and leads to a higher output price?

Now, to return to Josh and Geoff’s posts, what are the implications of Judge Breyer’s reasoning for the merchants who accept credit cards and want to join together to countervail the joint monopoly power of the Visa and MasterCard banks?  Would it be permissible under the Sherman Act for all the merchants unilaterally or through and agreement hire a single agent to provide them with buyer-side “interchange transfer services,” and then rely on that agent to negotiate lower interchange fees with Visa and MasterCard?

Finally, regarding the specific question raised by Josh — if you were counseling the merchants , would you be confident that you would be exonerated under the antitrust laws?  Or would you try to obtain an antitrust exemption as a little extra insurance?