Comments Before the Department of Transportation on the Statement of Enforcement Policy Regarding Unfair Exclusionary Conduct in

The Competitive Enterprise Institute (CEI) hereby submits these comments in response to the Department’s April 6 Notice requesting comments on its proposed policy statement regarding airline competition. CEI is a non-profit, 501(c)(3) public policy organization founded in 1984 which conducts research and educates the public and policymakers on the economic consequences of regulation. We also actively participate in regulatory proceedings on key issues and, when appropriate, engage in litigation. Since our founding, we have worked extensively on both antitrust and transportation regulatory issues.

The purpose of DOT’s policy statement is to prevent certain kinds of price cutting in the commercial airline industry. According to DOT’s Notice, the substantial consumer gains from 20 years of airline deregulation are threatened by predatory pricing by major carriers seeking to exclude upstart rivals from competing in the major carrier’s hub markets. DOT states that:

In recent years, when small, new-entrant carriers have instituted new low-fare service in major carriers’ local hub markets, the major carriers have increasingly responded with strategies of price reductions and capacity increases designed not to maximize their own profits but rather to deprive the new entrants of vital traffic and revenues. Once a new entrant has ceased its service, the major carrier will typically retrench its capacity in the market or raise its fares to at least their pre-entry levels, or both.

In response to concerns over the perceived dangers of such actions, the proposed policy statement would limit price cuts and expansion of capacity by major carriers in certain circumstances when such action is in response to competition from a new entrant and results in lower revenue than DOT would see as a “reasonable alternative response.”


We strongly urge DOT not to adopt the proposed policy statement. While the statement is intended to guard against predatory behavior by large airlines toward new entrants, history, economic analysis and legal precedent show that such behavior is rarely successful. However, the potential harm to consumers from new regulations restricting discount pricing is significant. The new guidelines would likely chill competition among airlines, as airlines try to steer clear of penalties for predatory discounting. The current jungle of airline competition could be turned into a virtual petting zoo. The clear losers would be airline consumers, the very group that DOT intends to help.

Our comments are divided into two parts. First, we will review the economics of predation claims in antitrust generally. Second, we will discuss specific concerns with the DOT’s proposed policy statement.

The Economics of Predation

Predatory pricing is the practice of selling output at prices below costs in an attempt to bankrupt rivals. According to predation theory, predatory pricers spend money today on a price war in the hope of achieving a monopoly tomorrow. If the value of the monopoly profits earned as a result of waging a successful price war is greater than the predator’s cost of waging this war, the predation is successful. The predator wins — but consumers lose.

Initially, this theory seems to make much sense. After all, a firm that sells at prices below costs will indeed harm its rivals by taking away their customers. Rivals without customers will eventually go bankrupt. The theory seems to leave little room for doubt that predatory pricing poses a genuine threat to competitive markets.

But not all is as it seems. When examined closely, the initial appeal of predatory pricing as a means of garnering monopoly power disappears. In fact, as we show below, predatory pricing is virtually never a sensible means of monopolizing markets. This point is not controversial in the economic literature. Economists today dispute only whether or not predatory pricing ever occurs; they do not dispute the established fact that predatory pricing, if it does occur, is exceedingly rare.

William Baumol, an economist not known for his hostility to active government regulation, recently noted that there is “general consensus among informed observers that genuine cases of predation are very rare birds.” Similarly, economist Ed Snyder and law professor Thomas Kauper surveyed suits alleging predation and discovered that the overwhelming majority of such suits are attempts by plaintiffs to stymie competition posed by more efficient and dynamic rivals.

These evaluations based upon the factual record fit quite comfortably with a skeptical assessment of predatory pricing: it is a foolish tactic. Of course, just because a tactic is foolish doesn’t mean that its success is impossible. Slashing prices below cost is indeed a possible means of achieving sufficient monopoly profit in the future to make today’s price cutting worthwhile. But the chances of successful predatory pricing are so slim that any accusation of predation deserves acute skepticism — skepticism as intense as would greet, say, a claim that a man who jumped without a parachute from a perch of 5,000 feet survived the fall. It’s possible that someone who falls from such a height, say, lands in a life-saving snow bank. But this remote possibility hardly lends general credence to any claimed sightings of people surviving such falls, and it certainly serves as no recommendation for people to jump without parachutes from such heights.

For the various reasons discussed below, many economists and antitrust scholars now believe that predatory pricing is so highly unlikely to benefit any alleged predator that the only reasonable response to all such allegations is to dismiss them out of hand. Robert Bork, for example, argues that “[i]t seems unwise . . . to construct rules about a phenomenon that probably does not exist or which, should it exist in very rare cases, the courts would have grave difficulty distinguishing from competitive price behavior.”

Not only are a large number of economists intensely skeptical of predatory-pricing claims, the United States Supreme Court has also grown leery of plaintiffs in predatory-pricing suits.

The general reason why cutting prices below cost is an avenue to bankruptcy rather than to monopoly dominance is that each rival who is at least as efficient as the predator has open to it every option exercised by the predator plus other and less costly options that are unavailable to the predator. As economists say, each rival has effective counterstrategies that will thwart the predator’s ambitions. Availability of these counterstrategies means, in turn, that a predator will be unable to charge monopoly prices for a long-enough span to time to recoup the losses it necessarily incurs during the predatory price war.

Forcing the predator to lose more money. One such counterstrategy is simply to sell less at the below-cost price than the predator sells — therefore forcing the predator lose more money. It is important to remember that a predator’s low prices as such do no harm to its prey. Rather, a predator harms its prey only by driving consumer demand for the preys’ outputs down so low that the prey must sell their outputs at a loss. Lowering prices is only a means of taking customers away from the prey. To strip its prey of valuable customers, the predator has no choice but to increase the amount it sells below cost during the price war.

Each prey, in contrast, can (and will) respond to the reduction of demand by selling a lower quantity of outputs at the predatorially low price. The predator does not have the preys' luxury of reducing the amount it sells below cost. Consequently, the predator must sell much more than is sold by its prey at below-cost prices — which, in turn, means that the predator’s self-inflicted losses are larger than are the losses that the predator inflicts on each prey.

Access to capital markets. Recognizing the fact that the predator necessarily suffers losses larger than those inflicted on the prey immediately casts doubt upon any claim of predatory pricing. The only way to salvage any such claim is to argue that the predator has a greater capacity than the prey to absorb losses. Such greater capacity enjoyed by the predator is absolutely necessary for the success of any predatory pricing campaign.

Common sense seems to dictate that a wealthier and more established firm can better afford to absorb predatory-pricing losses than can a less-wealthy and less established rival. Common sense, however, is here misleading. No firm, no matter how large, profitable or “dominant,” necessarily possesses greater power to absorb losses than do any of its equally efficient prey.

The reason is that a firm targeted for extinction by a predator can borrow what it needs to withstand attacks by a predator. Those who doubt this fact forget that financial credit supplied to survive a predatory-price attack differs in no fundamental respect from financial credit supplied to build or expand factories, to acquire supplies or inventories, or to do any of the thousand-and-one other things that firms do with credit. In all these cases, including surviving price wars, lenders lend monies to less-liquid borrowers who can profitably use these funds but do not presently have sufficient liquidity of their own to finance necessary current expenditures.

The same argument that says that prey who own less liquidity than does a predator will be unable to finance a price war also says that any firm that needs credit today for any reason, but who cannot repay until later, will be unable to acquire credit. Everyday experience contradicts this argument.

Capital market failure v. regulatory failure. There are, of course, cases where an equally-efficient firm might find it impossible to secure a credit line to fund its price-war struggles against a rival. Capital markets are imperfect. Lacking omniscience, lenders can and surely do fail to identify some worthwhile lending prospects.

However, such failure does not justify active government policing against price cutting. The very same incompleteness of information and understanding that renders capital markets imperfect renders government decisions imperfect. In fact, government is more likely to err in assessing claims of predation. As UCLA economist Harold Demsetz has written:

One of the most important reasons for trusting in competition is that there is little basis for faith that outsiders — be they government planners, judges, lawyers, or academics — possess the knowledge and the motivation required to fine-tune on behalf of consumers. A complex economy made up of many specialists is productive precisely because each need only deal with a relatively small part of the world; the actions of these specialists, disciplined by competition and motivated by the prospects of gain, linked together and made consistent by market prices, offer a superior method for utilizing knowledge on behalf of the consumer. Denial of this premise opens the door to centralized regulation of business activity.

Not only are government regulators more likely to err than are private investors, but the consequences of government errors are far more difficult to reverse than are those of private investors. For instance, when private lenders mistakenly find that low prices are predatory (and, thus, extend credit to rivals of the price cutter) the more-efficient firm is not prevented from charging the low prices made possible by its greater efficiency. Consequently, mistaken capital-market findings of predation are eventually exposed for the errors that they are. In contrast, a mistaken government finding of predation prevents the more-efficient firm from charging the low prices warranted by its superior efficiency. Unlike errors made by private lenders, errors made by government regulation are protected from exposure by the very remedy government prescribes — namely, mandated higher prices.

The Assets of Bankrupt Firms Are Not Destroyed. Even in cases where a firm does not obtain adequate financing and fails, competition is not necessarily reduced in the long-run. The assets of bankrupt firms aren’t destroyed; they remain available after bankruptcy to be used again. While this may be of no solace for an individual competitor, it is of great importance to consumers.

Assuming that the rivals are just as efficient as is the predator, they will likely be sold by the bankruptcy trustee to new owners intact. Indeed, if predatory-pricing theory is correct that the predator’s insistence on charging below-cost prices reduces the market value of its rivals, then the predator’s predatory-pricing strategy makes its rival firms attractive investment opportunities for other entrepreneurs. (If, as is more likely, a predatory-pricing campaign does not reduce the value of the predator’s rivals, then this fact means that the predation isn’t working in the first place — the alleged “prey” are never under a real threat.) And, in the unlikely event that a bankruptcy trustee is unable to sell the rival’s assets intact, these assets would likely still be used in competition with the predator.

For example, suppose predation by Delta Airlines bankrupts AirTran. AirTran’s bankruptcy trustee then has the choice of selling AirTran’s assets piecemeal or selling off the firm intact. If AirTran is just as efficient as is Delta, buyers would be eager to purchase AirTran and to put it back in competition with Delta. The market value of AirTran sold intact will be higher than the market value of AirTran’s assets sold off piecemeal. Because a bankruptcy trustee is obliged (and has incentives) to sell the firm at as high a price as possible, it will likely be sold intact. Moreover, if Delta has a reputation as an especially stubborn price cutter, all the better for potential buyers of bankrupted AirTran: the price that AirTran will sell for out of bankruptcy will be discounted by the threat of predation by Delta. AirTran may well come out of bankruptcy with lower costs than Delta!

If AirTran’s bankruptcy trustee can’t sell AirTran intact, AirTran’s assets remain a competitive threat to Delta. Bankruptcy doesn’t literally annihilate AirTran’s airplanes, computers, pilots, mechanics, landing slots, and other assets. These assets will find ready deployment where they promise the highest rate of return — which, because Delta is charging monopoly prices on some routes, is in the commercial aviation industry in competition with Delta.

The Recoupment Problem. There is another important reason why charging prices below costs is highly unlikely to generate monopoly power. In brief, the predator’s prospects of recouping in the form of monopoly profits the losses it incurs during the price war are scant — the principal reason being that monopoly profits attract new competition which keeps the predator from charging monopoly prices. Even if a predator’s current rivals are wholly unable to defend themselves from predatory prices, predation will not pay if entry by new rivals is sufficiently swift (relative to the time it takes to oust its prey) to keep the predator’s prospects for monopoly profits low. The most intense yearning for monopoly profits will not prompt a firm to incur $100 of losses today in return for only $90 of monopoly profits tomorrow.

Firms pursuing monopoly power through predatory prices face what is for them an unfortunate symmetry in the speed with which their rivals will exit and enter their industries. The ideal situation for a predatory pricer is to have its current rivals leave soon after it cuts its prices and to have any new entrants wait quite a long time before entering the industry in response to the predator’s monopoly prices. This way, the predator incurs losses only briefly but earns monopoly profits for quite a long time.

But reality doesn’t cooperate with predators. The same conditions that make an old rival’s exit swift also ensure that new rivals’ entry can be swift. Likewise, the same conditions that make entry slow make exit slow.

Why would a firm quickly fold its tent upon learning that a better-funded rival is pricing predatorially? Answer: if the cost of redeploying its assets elsewhere is low. And the cost of redeploying its assets in other markets (or industries) is low if the firm’s assets are not highly specialized to the market sporting the predator. Being useful in other markets in which no predation is occurring, the collection of assets that comprise the targeted firms can easily be redeployed into other markets until the predator jacks prices up to monopoly levels. The assets quickly driven by predation into other markets can then quickly return.

Unfortunately for predators, the only conditions consistent with slow entry in response to monopoly prices also guarantee slow exit (that is, that predators’ up-front losses will be huge). The reason is that slowness of entry in the face of monopoly prices is attributable to the necessity of making investments in that industry that are highly specific to that industry. If entry requires that a firm purchase a $1,000,000 machine that is useable only in the industry in question, investors will hesitate before making such an investment. Or, at any rate, investors will hesitate far longer before sinking $1,000,000 into a machine that is useful in no other industry than they will hesitate when contemplating an investment of $10,000,000 in a machine that can be used profitably in a number of different markets and industries.

In brief, predatory pricing in industries in which exit and entry are slow is more akin to financial Russian roulette than to rational profit seeking. And predatory pricing in industries in which exit and entry are quick is futile because even “victorious” predators can rarely earn monopoly profits for any substantial-enough length of time.

Evidence of Predatory Pricing? As discussed further below, there is nothing about the airline industry that exempts it from this general analysis of predatory pricing. There is no reason why any upstart airline that can efficiently compete with more-established carriers should be generally unable to acquire sufficient financing to become established in whatever routes it can efficiently serve.

This is true despite the fact that there are cases where price cuts by larger firms cause upstart rivals to abandon particular markets. New entry is supposed to reduce prices whether or not the upstart rival is efficient enough to compete successfully over the long run against the incumbent firm. Any upstart may or may not be sufficiently efficient to become a long-run rival of the incumbent. No upstart can really know prior to actually entering the market if it has what it takes to compete successfully with incumbents. Some upstarts will have what it takes; others won’t. Therefore, some upstarts whose entry prompts incumbent firms to lower their prices more closely to costs will find themselves unable to compete profitably with incumbents. These upstarts will leave the industry so that their resources are more usefully employed elsewhere. This is as it should be.

We mention this issue because it is common for those who fear predatory pricing to identify as real-world instances of predation cases in which upstarts’ entry causes incumbents to lower their prices, thereby driving the upstarts from the market and allowing these incumbents to again raise prices to their pre-entry levels. For example, in response to an argument published in the Wall Street Journal by an author of the present filing, noted airline deregulator Alfred Kahn identified a real-world instance of price cutting as supposed proof that predation does in fact occur.

Professor Kahn cites the fact that Northwest Airlines cut its fares between Detroit and Boston in response to entry on this route by Spirit Airlines. Discovering that it could not profitably serve this route with Northwest charging these lower fares, Spirit eventually abandoned this route. Not surprisingly, Northwest raised its fares after Spirit left. Professor Kahn mistakenly interprets this price-cut by Northwest as proof that predatory pricing occurs.

Professor Kahn’s error is caused by his unstated assumption that Spirit Airlines is just as efficient as Northwest at serving the Detroit-Boston route. But no evidence supports this assumption. While it’s possible that Northwest’s reduced fares reflected predatory intent, it’s at least as possible that these reduced fares reflected Northwest’s superior efficiency. As Northwestern University economist Robert Gordon writes, “It’s true, some new carriers have failed, but that’s hardly due to predatory pricing. Some start-ups underpriced their product, some lacked a sound business strategy and a number were managed poorly.”

Proposed Tests for Combatting Predatory Price Cutting by Major Airlines

To combat the alleged anticompetitive price cutting by major carriers, the D.O.T. plans to:

institute enforcement proceedings to determine whether the carrier has engaged in unfair exclusionary practices when one or more of the following occurs:

(1) the major carrier adds capacity and sells such a large number of seats at very low fares that the ensuing self-diversion of revenue results in lower local revenue than would a reasonable alternative response,

(2) the number of local passengers that the major carrier carries at the new entrant’s low fares (or at similar fares that are substantially below the major carrier’s previous fares) exceeds the new entrant’s total seat capacity, resulting, through self-diversion, in lower local revenue than would a reasonable alternative response, or

(3) the number of local passengers that a major carrier carries at the new entrant’s low fares (or at similar fares that are substantially below the major carrier’s previous fares) exceeds the number of low-fare passengers carried by the new entrant, resulting, through self-diversion, in lower local revenue than would a reasonable alternative response.

None of these tests make sense. We examine each in turn.

Test #1. Test #1 essentially says that a major carrier that cuts fares and adds capacity — seats — to any of its routes served by an upstart carrier is presumed to be practicing unlawful price cutting if the action causes the major carrier to receive less revenue on the route. This conclusion doesn’t follow.

Suppose that prior to entry by a new carrier, Delta alone serves the Pensacola/Atlanta market. (Atlanta is Delta’s major hub.) Let’s say that each week Delta supplies 2,000 seats on the Pensacola-to-Atlanta route, and 2,000 seats on the Atlanta-to-Pensacola route. Now an upstart — say, AirTran — enters this market and charges fares substantially lower than those charged by Delta. How should Delta respond?

The first fact to note is that AirTran’s entry and low prices leaves Delta with empty seats on its Pensacola/Atlanta route. Because the marginal cost to Delta of filling these seats is near-zero, even if Delta lowered some of its fares on this route to $1, it would not be pricing predatorially because $1 covers the marginal cost of filling a seat on a half-empty aircraft. Under test #1, DOT is careful not to condemn fare-cutting per se; instead, DOT condemns fare cutting combined with added capacity to this route.

Requiring a increased capacity along with price cuts makes sense. Lower prices on this route will attract new customers. If, before AirTran’s entry, Delta filled most of its seats on this route at the higher fare, then fare cuts by Delta can harm AirTran only if Delta itself satisfies much of the increased demand prompted by the lower fares. Otherwise, if Delta slashes its fares but doesn’t increase its capacity on this route, it leaves available to AirTran a group of customers willing to pay fares likely to yield revenues sufficient to cover AirTran’s costs of serving this route. As is true of predatory pricing in general, a predatory airline must increase the number of customers it serves at its below-cost fares.

Nevertheless, test #1 is motivated by the DOT’s mistaken view that when a firm’s total revenue falls after lowering its price in response to new competition, this fall in revenue strongly indicates predatory pricing. But why? Entry by AirTran reduces demand for Delta’s seats on the Pensacola/Atlanta route; it also obliges Delta to price more closely to its marginal cost. The result could easily be that Delta’s output rises and total revenues fall.

A familiar supply-and-demand graph shows this point nicely. In Figure 1 below, D1 shows the demand for Delta flights between Pensacola and Atlanta before AirTran’s entry. Being the exclusive supplier on this route, Delta maximizes its profits on this route by equating its marginal revenue with its marginal cost. QM is the resulting pre-entry level of output, and PM is the fare Delta charges. Delta’s total revenue before AirTran enters is shown by the area PMAQMO. When AirTran enters it reduces the demand for Delta’s services on this route. This demand reduction is depicted in the graph by an inward movement and flattening of the demand curve confronting Delta. The exact distance of the inward shift and extent of the flattening of the demand curve depends on a variety of factors, including the kinds of services offered by AirTran. Nevertheless, it is not only possible, but likely, that a shift as depicted in Figure 1 — from D1 to D2 — is what will happen to the demand for Delta’s services on a particular route when AirTran enters in competition with Delta.

D2 depicts a lower as well as much more elastic demand confronting Delta. D2 is more elastic because, with another competitor in the market, any fare increase by Delta will cause its quantity demanded to fall by more than it would fall in the absence of AirTran. D2 also crosses Delta’s marginal-cost curve at some quantity greater than QM — suggesting that when Delta prices competitively it will (because it is the better-known carrier) supply a larger share of the market than will AirTran .

Note two features of Figure 1. First, Delta’s profit-maximizing output on this route increases (from QM to QC) after AirTran enters. This explains why Delta might rationally increase capacity on this route. Second, Delta’s total revenue after AirTran’s entry (PCBQCO) is lower than before AirTran’s entry (PMAQMO).

Of course, the graph can be drawn differently. But the general features depicted in Figure 1 are certainly neither contorted nor rare. Therefore, DOT’s test #1 emphatically does not identify monopolistic threats while leaving legitimate competitive responses unassaulted — a fact that highlights the more general problem with test #1, namely, its failure to distinguish legitimate responses to competition from responses intended to protect monopoly power.

Figure 1 is drawn under the assumption that, prior to AirTran’s entry, Delta behaved like a monopolist. It’s possible, however, that even if Delta is the sole carrier on a particular route, it behaves more like a competitor than a monopolist because it knows that new entry into this route is a looming possibility. The attraction to a new entrant of the Pensacola/Atlanta route is obvious if Delta behaves like a monopolist on this route. But there are three reasons why AirTran might enter this market even if Delta isn’t charging monopoly prices: (1) AirTran believes that a larger number of passengers can profitably be served on the Pensacola/Atlanta route; (2) AirTran believes that, while the number of passengers Delta serves on this route is the correct number that can be optimally served, AirTran hopes to share these passengers with Delta — a hope that, if correct, suggests that Delta operating alone does not serve these 2,000 weekly passengers as efficiently as would two airlines flying between Pensacola and Atlanta; and (3) AirTran hopes to replace Delta as the exclusive carrier on this route serving the 2,000 passengers weekly.

Consider first the possibility that AirTran believes that more than 2,000 passengers weekly can profitably be served on the Pensacola/Atlanta route. Perhaps Delta underestimated the profitability of serving more than 2,000 passengers weekly on the Pensacola/Atlanta route. Perhaps consumer demand to fly between Pensacola and Atlanta has increased (because of, say, economic growth in Pensacola or Atlanta). Delta is expected, and should be permitted, to compete for these additional passengers. Price reductions are emphasized when people discuss the benefits of competition, but economics makes plain that quantity increases are the flip-side of the coin of price reductions. Competition reduces prices only because competition increases output. To criticize Delta for increasing output and reducing its fares under such circumstances is to criticize Delta for competing.

Consider the second possible reason why AirTran enters this route, namely, to share the 2,000-passengers-weekly with Delta. If under these circumstances Delta increases its capacity on this route beyond 2,000 — and if Delta is certain that 2,000 passengers per week is the optimal total number of passengers weekly to be served on this route — a sensible conclusion is that Delta is indeed trying to protect its exclusive position on this route. But it is a futile attempt. By assumption here, AirTran can profitably serve some of the passengers on this route. Moreover, it’s possible that if Delta now tries to serve more than 2,000 passengers weekly, its costs per passenger will rise — thus increasing AirTran’s chances of success in entering this market. Regardless, though, of what happens to Delta’s per-passenger cost as it expands service on the Pensacola/Atlanta route, AirTran here is assumed to have a profitable niche in this market.

Delta’s only hope for success is that the price war causes AirTran to abandon this route. But for the reasons reviewed above, AirTran will probably not be repelled by Delta’s below-cost pricing because AirTran can (again, by assumption here) profitably serve this route. Knowing that AirTran will probably survive a price war, Delta is unlikely to increase its capacity on this route as part of a predatory-pricing campaign. Any observed increase in capacity by Delta is, therefore, best interpreted as indicating Delta’s sincere (though perhaps mistaken) belief that optimal service on the Pensacola/Atlanta route has increased beyond 2,000 passengers weekly.

Consider now the third and final reason that AirTran might enter the Pensacola/Delta route — namely, to try to take this entire market away from Delta. It’s possible that 2,000 is the optimal number passengers weekly served on this route and that costs are minimized when only one airline supplies these 2,000 seats each week. Economists call this possibility a “natural monopoly” — which occurs whenever a single supplier can serve a market at a lower unit cost than can two or more suppliers.

When AirTran enters the route (here assumed to be a natural monopoly) and Delta expands capacity and lowers fares, the reason for Delta’s expanded capacity and lower fares indeed is that it seeks to protect its ability to serve this route exclusively. But so what? Nothing stops AirTran from matching each of Delta’s efforts. More importantly, Delta’s fare-cutting and capacity-expanding activities under these circumstances cannot generate higher prices in the long-run than would otherwise exist. If competition for the market (rather than competition within the market) is what’s going on, then — assuming that both airlines are equally efficient at serving this market — it makes no difference which airline wins the competition. One will eventually oust the other, and the winner will then charge profit-maximizing monopoly fares on this route. If indeed the Pensacola/Atlanta route can efficiently support only one carrier, preventing airlines from lowering fares as they compete to be the single supplier of this market denies passengers the low fares and greater service they would enjoy during the price war without keeping prices lower than they would otherwise be in the long run.

Also keep in mind that Delta might wish to serve as many passengers as possible on the Pensacola/Atlanta route because it wants to keep as low as possible its costs of serving customers at its Atlanta hub. As DOT itself admits, more than one-half of all passengers flying from non-hub airports to hub airports have as their final destinations other non-hub airports. Imagine a passenger wishing to fly round-trip between Pensacola and Charleston. If this passenger flies Delta from Pensacola to Atlanta, he is more likely to fly Delta on the leg from Atlanta to Charleston.

Suppose that AirTran enters the Pensacola/Atlanta market. Some of these passengers who, had they flown Delta from Pensacola to Atlanta, would fly Delta from Atlanta to Charleston, will now fly some other airline from Atlanta to Charleston. Delta thus suffers unfilled seats not only on its Pensacola/Atlanta flights, but also on its Atlanta/Charleston flights — and, indeed, on many of its flights from and to its Atlanta hub. To keep load factors on its Atlanta flights high, it may well be worthwhile for Delta to serve as much of the Atlanta/Pensacola market as it can even if maintaining its “dominance” in the Atlanta/Pensacola route is a net loser when considered independently.

Also, AirTran’s entry into even one market — for example, Pensacola/Atlanta — releases Delta’s capacity in many of its other markets (because, again, passengers flying to Atlanta from Pensacola on carriers other than Delta are less likely to fly Delta from Atlanta to other destinations). This fact means that any diversion by Delta of capacity from other routes into the Pensacola/Atlanta route is not as suspicious or as costly as DOT suggests. It may well be that losing only a handful of passengers on, say, it’s Atlanta-to-Charleston flights makes it profitable for Delta to shift aircraft from serving Charleston/Atlanta to serving Pensacola/Atlanta.

In brief, there are any number of sound and socially beneficial reasons why major carriers lower fares and expand capacity in response to new competition. Establishing fare cutting and capacity expansion as prima facie evidence of predation in the face of new entry is to declare normal means of competition as prima facie suspect. Airlines will naturally be more reluctant to compete on the basis of price. Far from protecting consumers, the proposed guidelines would severely harm them.

Perhaps in recognition of this, the proposed test does contain a saving clause making lower fares and expanded capacity illegal only if they would result in lower local revenue than a “reasonable alternative response,” thus allowing some or all points made above to be raised as a defense. This clause is of little comfort, however, raising as many troubling questions as it solves. Who, for instance, determines whether an alternative response is “reasonable”? Will DOT be required to judge the relative efficacy of alternative business practices? How is “reasonableness” to be defined? In the end, we fear that the federal government will simply be put in the role of second-guessing the business strategies of major airlines – a role it has not played in 20 years, and one it has never played successfully.

Tests #2 and 3. Many of the weaknesses and inconsistencies that we identify above as plaguing DOT’s proposed test #1 apply with equal force to proposed tests #2 and #3. So we need not spend much time on these latter tests. We do point out, however, that tests #2 and #3 propose a standard of comparison that, even apart from the many other weaknesses making these tests invalid, render them especially suspect.

Both of these tests condemn an incumbent carrier that cuts its fares and, as a result, carries on the route in question a greater number of passengers than is carried by the new entrant (or a greater number of passengers than the new-entrant’s capacity permits it to carry) and that, as a result of this fare cut, the incumbent carrier earns less revenue than is available from any “reasonable” alternative responses to the heightened competition.

Such comparisons of different carriers’ actual passenger loads or capacities are irrelevant for purposes of antitrust scrutiny. The elasticity of demand for Delta on a certain route might be sufficiently high to cause the increase in the profit-maximizing number of passengers carried by Delta at the lower rate to exceed AirTran’s capacity on that route. Nothing in economic theory remotely suggests that a monopolist compelled by new entry to reduce its fares to competitive levels will not serve a larger number of consumers than are served by the new entrant and, at the same time, suffer revenue losses. Figure 1 is again helpful. It shows clearly why the incumbent might simultaneously increase its capacity and suffer lower revenues in response to entry by an upstart.

Airlines’ Difficulty of Sustaining Monopoly Pricing. Recall our earlier discussion of a successful predator’s need to keep new rivals from entering after the prey are ousted. For predation to be profitable, the predator must be able to charge monopoly prices after the price war in order to recoup its price-war losses. Unfortunately for any predatory airline, rivals can quickly enter routes temporarily plagued by monopoly pricing. Here is DOT's own admission that aircraft and other resources are easily transferred from route to route:

Compared to firms in other industries, a major air carrier can . . . shift resources between markets much more readily. Through booking and other data generated by computer reservation systems and other sources, air carriers have access to comprehensive, “real time” information on their competitors' activities and can thus respond to competitive initiatives more precisely and swiftly than firms in other industries.

Precisely. Confusedly, however, DOT uses this fact as a reason to fear predation rather than as a reason to dismiss claims of predation.

It bears repeating that predatory pricing, if it is to be worthwhile for the predator, must promise a substantial-enough stream of monopoly profits in the future. But monopoly profits cannot be earned for any length of time if soon after a firm begins charging monopoly prices other firms enter that market. DOT here admits that aircraft can be shifted “swiftly” from route to route. If such swiftness is indeed a fact, then it is a reason sufficient in itself to dismiss any allegations of predation in this industry.

Quite frankly, we are befuddled — although pleased — by DOT’s explicit admission of just how easy it is for airlines to swiftly shift their aircraft and other resources from route to route. This admission alone destroys DOT’s case for predation.

The “Demonstration Effect”

Advocates of stronger restrictions on predatory pricing often try to buttress their arguments by alleging what they call “the demonstration effect.” The demonstration effect supposedly occurs when a firm that operates in several markets charges below-cost prices in one or a few markets and keeps these prices below cost until all rivals in these markets leave (or cooperate sheepishly in keeping prices and outputs at monopoly levels). The idea is that a larger, multi-market firm has a long-term incentive to suffer even very large losses in some markets in order to demonstrate to its rivals in all of its markets that it is serious about using price cutting to oust rivals who compete aggressively.

This “demonstration effect” argument is aimed at showing that a firm can have the incentive to price below cost and expand output, because even if it can never recoup what it lost in that market, it will often be worthwhile (so goes the allegation) for this firm to instigate such predation in order to frighten rivals in other markets into not competing aggressively. The argument says, in effect, that a firm can profit from money-losing behavior in a particular market if it has similar opportunities for losing money in other markets. Moreover, the demonstration-effect argument proceeds as if the predator is the only firm that can benefit from demonstrating such perverse stick-to-it-tiveness.

Suppose Delta operates in only a single route — say, Atlanta/Miami. If Delta prices predatorially on this route, and if its rivals meet its price cuts, then Delta must either abandon its predation (because it loses more money than do any of its rivals), or mulishly keep its fares below cost until it goes bankrupt. For the reasons we reviewed above, Delta is extremely unlikely to pursue a policy of below-cost pricing as a means of capturing monopoly power on this route.

Now suppose that Delta serves several routes in addition to Atlanta/Miami. The idea of the demonstration effect is that, whereas Delta might not stick very long with below-cost pricing on the Atlanta/Miami route if that is its only route, Delta will persevere with below-cost pricing on this route when it has something to prove to rivals on the other routes that it serves. Rival carriers on other routes will notice that Delta willingly suffered large and non-recoupable losses on its Atlanta/Miami route. The reputation that Delta earns as a result — the reputation of being a firm unafraid to suffer non-recoupable losses — will frighten Delta's rivals on other routes into avoiding any price-war challenges against Delta.

But what's so frightening? Having a reputation of stubbornly sticking with a policy of suffering non-recoupable losses is, at bottom, having a reputation for being unafraid of bankruptcy. Delta's stubborn insistence on charging below-cost prices is really Delta's stubborn insistence on going bankrupt. If Delta earns such a reputation, its rivals will have an easier time acquiring sufficient financing to keep operating during the price war.

We agree with Judge Frank Easterbrook’s assessment of the demonstration effect (also called “strategic predation”):

Arguments about strategic predation are overblown. The idea . . . is that a given firm operates in a cascade of markets, and that if it only puts on enough of a predatory show in the first market it will be able to prevent people from competing in any market.

It seems to me that is the strategic equivalent of Napoleon’s promise to fall on his sword if the Russians don’t surrender. The promise to lose a great deal of money in Market 1 — probably much more money than your new entrant is losing in Market 1 — in order to convey the message that you will be willing to do the same thing in Market 2 and Market 3 and Market 4, is incredible. The message you convey is, you have a strategy that can succeed if and only if you never again have to carry it out. Once you have conveyed the message, and rivals know that the strategy can succeed only if not executed, you have also conveyed the message to somebody else that now is that right time to enter Market 2. Everybody knows that your predatory threat is profitable only if it is not carried out. That is the perfect time to enter.

More generally, the demonstration-effect argument (as with all predation arguments) rests implicitly on the assumption that the predator has access to greater funds with which to fight a price war than do the prey. How else could such a demonstration work? Any equally efficient rival with access to the same amount of financing as the predator could counter the predator's demonstration of willingness to lose money with the rival's own demonstration of willingness to lose money. And because the predator would lose more money than the rival — meaning that the predator would go bankrupt first — the rival's threat would always be more believable!

So, as with all arguments supporting government prevention of predatory pricing, those built upon the demonstration effect sneak in the implicit assumptions that capital markets consistently will fail to channel funds to efficient rivals of predators and that government decision makers are not hampered by the same information imperfections that mar capital markets. Both assumptions are unwarranted.


For the reasons outlined above, we believe the proposed guidelines are unnecessary, and present a significant risk to competition and to airline consumers. We urge DOT to reject them.

Respectfully submitted,

James L. GattusoVice President for Policy and ManagementCompetitive Enterprise InstituteDonald J. BoudreauxPresident, Foundation for Economic EducationAdjunct Scholar, CEI1001 Connecticut Ave. N.W., S. 1250Washington, D.C. 20034(202) 331-1010

Figure 1Demand for a major-carrier’s seats before (D1) and after (D2) entry of new carrier

1United States Department of Transportation, Office of the Secretary, “Statement of Enforcement Policy Regarding Unfair Exclusionary Conduct,” Docket No. OST-98-3713, April 6, 1998. (Available online at All page references throughout to this DOT Statement are to the online version.2See, i.e., Competitive Enterprise Institute and Consumer Alert v. NHTSA, 956 F.2d 321 (D.C. Cir. 1992).3Id., p. 2.4William J. Baumol, “Predation and the Logic of the Average Variable Cost Test,” 39 Journal of Law and Economics 49-72, April 1996.5Edward A. Snyder & Thomas E. Kauper, “Misuse of the Antitrust Laws: The Competitor Plaintiff,” 90 Mich. L. Rev. 551 (1991).6Robert H. Bork, The Antitrust Paradox (New York: Basic Books, 1978), p. 154.7The Court, however, has so far refused to dismiss altogether the possibility of successful predatory pricing. See Matsushita Electric Industrial Co. v. Zenith Radio Corp., 475 U.S. 574 (1986) and Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 113 S. Ct. 2578 (1993). We are mystified at the Court’s failure to declare as per se legal all price cutting by firms not subject to explicit government price regulation. The Court uses per se rules of illegality to condemn outright those practices that it regards as harmful (e.g., explicit price fixing) even though the Court concedes the possibility that in rare circumstances these per se illegal practices might prove socially beneficial. Considerations of judicial economy justify the use of per se rules: if, after extensive investigations, courts will almost always find a practice to be illegal. Courts could then justifiably avoid the investigations by declaring that the practice in question is always prohibited.The same considerations of judicial economy, however, counsel the creation of explicit rules of per se legality. And given what economists and courts know about the overwhelming benefits of price competition and the extreme unlikelihood of successful predatory pricing, a judicial rule declaring per se legal all price cuts (by firms not subject to government price regulation) would be a good place to begin in crafting rules of per se legality. See Donald J. Boudreaux, Kenneth G. Elzinga, & David E. Mills, “The Supreme Court’s Predation Odyssey: From Fruit Pies to Cigarettes,” 4 Supreme Court Economic Review 57 (1995).8″[P]rice cutting, though conventionally viewed with grave suspicion, does not provide a likely means of predation because it requires the predator to bear losses that are much larger, both absolutely and proportionally, than those inflicted on the intended victim.” Robert H. Bork, The Antitrust Paradox (New York: Basic Books, 1978), p. 148. See also William F. Shughart II, The Organization of Industry (Houston: Dame Publications, 1997), p. 462.9We assume here (along with most popular commentors on predatory pricing) that large firms tend to have more ready liquid assets on hand than do its smaller rivals. In reality, this assumption in any particular case may or may not be true.10A large or “dominant” firm may well be more efficient at serving the market than is a smaller or less experienced firm — but in this case the larger firm need not price predatorially in order to win market share from its less-efficient rivals.11Harold Demsetz, “The Trust Behind Antitrust,” in Eleanor M. Fox & James T. Halverson, eds. Industrial Concentration and the Market System (Washington, DC: American Bar Association, 1979), pp. 47-48. See also Frank H. Easterbrook, “The Limits of Antitrust,” 63 Texas Law Review 1, 24-25 (1984).12See Frank H. Easterbrook, “Predatory Strategies and Counterstrategies, 48 University of Chicago Law Review 263 (1981).13See Bork, The Antitrust Paradox, p. 153: “Indeed, by depressing the value of the victim’s business, the predator makes that business an attractive investment for any purchaser with adequate reserves.”14We deal below with the so-called “demonstration effect” argument that says that many incumbent firms have incentives to stubbornly keep prices below cost in order to develop a reputation for price cutting that frightens rivals never to challenge these stubborn price cutters. 15The pioneering work on the economics of recoupment is Kenneth G. Elzinga & David E. Mills, “Testing for Predation: Is Recoupment Feasible?” 34 Antitrust Bulletin 869 (1989).16See James L. Gattuso, “Don’t Outlaw Cheap Airfares,” Wall Street Journal, April 8, 1998. Professor Kahn’s response is “Top Airlines Gobble the New Guys,” Wall Street Journal, Letters to the Editor, April 20, 1998. Further exchanges in this series of letters were “Competitors, Not Predators,” Letters to the Editor, May 18, 1998; and “Successful Predation in Unfriendly Skies,” Letters to the Editor, June 12, 1998.17Robert J. Gordon, “Flights of Fancy,” Chicago Tribune, April 7, 1998.18Id., p. 5.19Another reason the DOT may be requiring capacity increases as evidence of predation might be that it is trying to be cautious — that it wants evidence of predation in addition to price cutting. Additional capacity is that other evidence. Because capacity added to this route enables Delta to offer improved service (such as more scheduled flights between Pensacola and Atlanta), Delta harms AirTran not only by undercutting AirTran’s prices but also by outdoing AirTran in service. Predation occurs both on price and nonprice fronts. This does not change our basic analysis, however. All that we say in this paper against legal action aimed at predatory pricing applies to actions against so-called “non-price predation” as well actions against price predation.20An important difference between neoclassical and Chicago-school conceptions of competition and the Austrian-school conception is that Austrians emphasize the dispersion of knowledge among market participants and the process of learning. In the Austrian view, the fact that Delta profitably served (say) 2,000 passengers weekly on the Pensacola/Atlanta route does not imply that Delta accurately assessed all the possibilities for profitably serving this market. When AirTran enters this market with lower fares, this entry might well teach Delta something that in an ideal world it would already know, but which in reality it does not know until AirTran teaches it about this possibility.21See, e.g., Robert B. Ekelund, Jr. & Richard Ault, Intermediate Microeconomics (Lexington, MA: D.C. Heath, 1995), pp. 326-332.22Some might argue that government regulation is necessary to prevent carriers from charging monopoly prices on those routes for which each of these carriers enjoy a “natural monopoly.” Although we don’t share the optimism of those who endorse price regulation, whether or not price regulation is necessary to avoid monopoly pricing on such routes is a separate question from that of predation.23″Flow traffic, or the passengers that the major carrier is transporting from their origins to their destinations by way of its hub, typically accounts for more than half of the traffic in local hub markets.” DOT, Statement of Enforcement Policy, p. 2.24DOT Statement of Enforcement Policy, pp. 4-5; emphasis added.25Remarks of Frank H. Easterbrook in Steven C. Salop, ed., Strategy, Predation and Antitrust Analysis, “Roundtable Discussion,” p. 665.