United States: American Airlines: In Search of an Appropriate Measure of Cost for Predation

Predatory pricing, once dismissed as implausible and irrational, now is viewed by some as not only plausible, but profitable.¹ Predation, for example, may be profitable in a multi-market context where predation occurs in one market and recoupment occurs rapidly in other markets.² Recently, the United States Court of Appeals for the Tenth Circuit affirmed a district court’s grant of summary judgment dismissing a multi-market predation/recoupment claim brought by United States Department of Justice ("DOJ") against AMR Corporation, American Airlines, Inc., and American Eagle Holding Corporation ("American").³ In holding that DOJ failed to raise a genuine issue of material fact with respect to the first element of its predatory pricing claim—whether American priced below an appropriate measure of cost—the Tenth Circuit declined to examine the second element of DOJ’s predatory pricing claim—whether American had a dangerous probability of recouping its alleged investment in below-cost prices in markets other than the one where the below-cost pricing allegedly occurred.⁴

DOJ filed suit on May 13, 1999, alleging that American instituted a scheme of predatory pricing in violation of §2 of the Sherman Act.⁵ In DOJ’s view, American’s combined response to low cost carrier ("LCC") competition at American’s Dallas/Fort Worth International Airport ("DFW") hub of lowering prices, increasing capacity and altering yield management (that is, making more seats available at the new, lower prices), constituted an unlawful, anticompetitive response.⁶ By increasing capacity, American purportedly overrode its own internal capacity-planning models for each route, which had previously indicated that such increases would be unprofitable.⁷ In each instance, the results were the same: the competing LCC failed to establish a presence, moved its operations or ceased its separate existence entirely.⁸ After the LCC ceased or moved its operations, American generally resumed its prior marketing strategy, reducing flights and raising prices to levels roughly comparable to those prior to the period of LCC competition.⁹ However, capacity usually remained higher after LCC exit than before the alleged predation.¹⁰ DOJ argued that the "incremental" revenues and costs specifically associated with American’s capacity additions showed a loss.¹¹ Because American spent more to add capacity than the revenues generated by the capacity additions, DOJ claimed that such capacity additions made no economic sense unless American intended to drive the LCCs out of the market.¹² DOJ alleged that by attempting to monopolize the four city-pair routes of DFW-Kansas City, DFW-Wichita, DFW-Colorado Springs, and DFW-Long Beach, American hoped to develop a reputation as an exceedingly aggressive competitor and set an example for all potential competitors.¹³ Thus, fearing American’s predatory response, future potential competitors would decline to enter other DFW market routes and compete.¹⁴ By preventing or at least forestalling the formation of an LCC hub at DFW, American could then charge higher prices on other DFW routes and recoup the losses it incurred from its "capacity dumping" on the four city-pairs.¹⁵

The district court, after reviewing a voluminous record and receiving extensive briefs, granted American’s motion for summary judgment, concluding that DOJ failed to demonstrate the existence of a genuine issue of material fact as to (1) whether American had priced below cost and (2) whether American had a dangerous probability of recouping its alleged investment in below-cost prices.¹⁶ The district court also stated that even if American had priced below an appropriate measure of cost, it nevertheless was entitled to summary judgment because "American’s prices only matched, and never undercut, the fares of the new entrant, low cost carriers on the four core routes."¹⁷ In reaching this conclusion, the district court imported the statutory "meeting competition" defense from the Robinson-Patman Act reasoning that "there is strong inferential support for the idea that the defense may be appropriate in a given case."¹⁸

In Search of An Appropriate Measure of Cost In Brooke Group Ltd. v. Brown & Williamson Tobacco Corp.,¹⁹ the United States Supreme Court established two pre-requisites to recovery in every predatory pricing case. First, every plaintiff must show that the prices complained of are below an appropriate measure of costs.²⁰ Second, every plain-tiff also must show that there is a dangerous probability that the investment in below-cost pricing will be recouped.²¹ Despite a conflict among lower courts over the appropriate measure of cost, the Court declined to definitively state what the appropriate measure of cost is in a predatory pricing case.²² In dicta, however, the Court implied that predatory prices must be below some measure of incremental costs.²³ Since incremental cost is in effect marginal cost (that is, the cost of producing an additional increment of output), as a matter of economic theory prices above a firm’s short-run margin-al cost are presumably not predatory because each additional sale decreases losses or increases profits.²⁴ Because marginal cost is difficult to measure, a commonly accepted proxy for marginal cost in predatory pricing cases is average variable cost ("AVC") (that is, the sum of all variable costs—those that vary with output—divided by output).²⁵ However, sole reliance on AVC to the exclusion of other proxies for margin-al cost may obscure the nature of a particular predatory scheme.²⁶ Whatever the proxy used to measure marginal cost, it must nevertheless be accurate and reliable.²⁷

Despite the United States Supreme Court’s pronouncement in Brooke Group regarding predatory pricing, DOJ argued that one could not simply view America’s capacity additions through the Brooke Group lens of predatory pricing where the challenged conduct is not just price related. Instead, DOJ argued that one must ask whether the challenged conduct makes economic or business sense only because it tends to eliminate or lessen competition.²⁸ If so, then the challenged conduct is predatory. The district court and the Tenth Circuit disagreed: "[w]hile the specific behavior complained of in the instant case is an increase in output or frequency, these actions must be analyzed in terms of their effect on price and costs. Thus, in order to succeed in the present action, the government must meet the standards of proof for predatory pricing cases established in Brooke Group."²⁹

Conceding that AVC is a good proxy for marginal cost in most cases, DOJ nevertheless argued that "there may be times when looking only to a market-wide AVC test will disguise the nature of the predatory conduct at issue."³⁰

According to DOJ, the market-wide AVC test is inappropriate where there is a challenge to "well-defined incremental conduct, and where incremental costs may be directly and confidently measured utilizing alternative proxies to AVC."³¹ DOJ objected to the use of a market-wide AVC test comparing price to average cost over an entire range of sales. In DOJ’s view, a market-wide price-cost comparison "risks masking the reality of predation . . . [because] the market-wide approach focuses on the profitability of the defendant’s overall operations in the market, and not on the profitability of the challenged exclusionary conduct."³² Thus, according to DOJ, "reliance on a market-wide AVC test is likely to disguise the unprofitability of challenged conduct, while an incremental test is likely to reveal it."³³

Considering American’s conduct to be just that—"well-defined incremental conduct"—DOJ proffered four tests to "measure reliably incremental costs—the precise costs associated with the capacity additions at issue."³⁴ Each test relied on cost measures used by American in its internal decisional-accounting system (that is, accounting measures that are used for internal decision making, not financial reporting).³⁵ Due to similarities among the four tests, the district court grouped them as Tests Two and Three, and Tests One and Four for purposes of its analysis, and the Tenth Circuit did the same.³⁶

DOJ’s Tests Two and Three were grouped together by the district court because they "purport to measure incremental cost by looking to whether certain of American’s internal cost-accounting measures became negative following the allegedly predatory capacity additions."³⁷ The Tenth Circuit rejected Tests Two and Three, as did the district court, because Tests Two and Three were based on a "fully allocated earning measure, meaning that general operating expenses are arbitrarily allocated by American’s decision accounting system to the flight or route level, and do not necessarily represent the exact costs associated with a particular flight or route."³⁸ While Tests Two and Three included some costs directly associated with a particular flight or operations of a particular route, they also included costs that were not related to the operation of a particular flight or route.³⁹ Since DOJ’s Tests Two and Three included fixed costs—costs that do not vary with the level of output - the Tenth Circuit upheld the district court's finding that "utilizing these cost measures would be the equivalent of appying an average total cost test, implicitly ruled out by Brooke Group’s mention of incremental costs only."⁴⁰ Thus, since Tests Two and Three relied on cost measures that were not variable or avoidable with respect to the capacity increases at issue, the Tenth Circuit held Tests Two and Three "invalid as a matter of law as a measure of allegedly predatory capacity increases."⁴¹

The district court grouped DOJ’s Tests One and Four together "labeling them as short-run profit-maximization tests."⁴² Test One examined changes in America’s profitability by employing the fully allocated earning measure discussed above, and an internal measure of American’s variable costs.⁴³ If these measures declined following a capacity addition, Test One allegedly demonstrated that adding capacity forced American to forego better profit performance elsewhere.⁴⁴ Test Four supposedly compared the revenue from incremental passengers with the average avoidable cost of adding capacity.⁴⁵ In other words, Test Four attempted to "reveal American’s predatory conduct by measuring and comparing the incremental costs incurred by American when it added capacity to the city-pair routes in question to the incremental revenue it received from the additional capacity."⁴⁶ Under Test Four, if incremental revenues were below incremental costs, there was "evidence of sacrifice."⁴⁷

Rejecting Test One, the Tenth Circuit observed that "rather than determining whether the added capacity itself was priced below an appropriate measure of cost, Test One effectively treat[ed] forgone or ‘sacrificed’ profits as costs, and condemns activity that may have been profitable as predatory."⁴⁸ Because Test One simply performed a "before-and-after" comparison of a route as a whole looking to whether profits on the route as a whole declined after capacity was added, not whether the challenged capacity additions were done below costs, the Tenth Circuit concluded that "Test One indicates only that a company has failed to maximize short-run profits."⁴⁹ Because "[s]uch a pricing standard could lead to a strangling of competition, as it would condemn nearly all output expansions" the Tenth Circuit held Test One "invalid as a matter of law."⁵⁰ 

Finally, the Tenth Circuit rejected Test Four finding that Test Four failed to identify the actual costs associated with the capacity additions.⁵¹ American argued that some of the variable costs included in Test Four did not vary proportionately with the level of flight activity, but instead were allocated arbitrarily to a flight or route by its internal decisional-accounting system.⁵² Therefore, according to American, as an allocated variable cost measure, Test Four cannot be used to calculate the avoidable cost of the added capacity.⁵³ The Tenth Circuit agreed: "[b]ecause the cost component of Test Four includes arbitrarily allocated variable costs, it does not compare incremental revenue to average avoidable costs. Instead, it compares incremental revenue to a measure of both average variable cost and average avoidable costs. Therefore, Test Four does not measure only the avoidable or incremental cost of the capacity additions and cannot be used to satisfy the government’s burden in this case."⁵⁴

The Tenth Circuit concluded that the four proxies used by DOJ to show pricing below an appropriate measure of cost were "fatally flawed in their application, and fundamentally unreliable."⁵⁵ Because of DOJ’s failure to establish pricing below an appropriate measure of cost, the Tenth Circuit Act’s "meeting competition" defense in the context of a §2 Sherman Act predatory pricing case.⁵⁷

Despite approaching the matter with "caution" but not "with the incredulity that once prevailed," the Tenth Circuit rejected several tests put forth by DOJ for measuring costs in evaluating predatory behavior. In doing so, the Tenth Circuit held that strict adherence to Brooke Group is required, including when non-price predation is at issue. At a time when theories of predation are finding increased acceptance, the Tenth Circuit’s opinion thus provides important guidance for measuring costs in connection with a broad variety of predatory conduct. Jonathan L. Lewis (Chicago)