Predatory Pricing : LUSENET : MS-DOJ : One Thread

Predatory Pricing: Laura Steiner

If you have only a limited time, "defensive predation" is the most important section.

THEMES A few themes from Matsushita are important for analysis of Fisher, Schmalansee, and the Microsoft Antitrust case in general.

Firstly, there is a huge OVERDETERRENCE concern with predation. Since predation involves low prices, which generally benefit consumers, if a court mistakenly determines that predation has occurred, the firm will be punished for very efficient competition.

Matsushita also establishes that there is a STRONG PRESUMPTION THAT PREDATIONparticularly conspiring to predationis IRRATIONAL. In Matsushita, the "proof" of below-cost pricing from an economic model was overshadowed by the strong assumption of irrationality. Furthermore, we see that the GOVERNMENT HAS A STRONG BURDEN OF PROOF that the pricing was in fact below cost just to survive summary judgment.

A further illustration of the governments strong burden of proof: if there is no plausible reason to conspire, ambiguous evidence will not be interpreted as a conspiracy to predatory price. Presumably, even without a conspiracy, ambiguous evidence will not be interpreted as predatory pricing, due to the strong presumption that predation is irrational.

Evidence must "ten[d] to exclude the possibility" that petitioners underpriced respondents to compete for business rather than to implement an economically senseless conspiracy." (Matsushita, 597-598) Once again, if there is a competitive story that accounts for the facts, the government will lose a summary judgment motion, at least when the predation story is "economically senseless."

The overdeterrence concern seems equally applicable in our case. If theres a competitive reason to give away browsers free, the government certainly does not want a company to have to charge money for it in order to protect itself from antitrust charges. I think there is a viable argument that the strong presumption of irrationality that exists in most predation cases is not present here (at least with regard to defensive predation), because defensive predation is a new breed of predation.


As Schmalensee pointed out, Fishers definition differs from the courts in Brooke in two ways, 1using "not profitable" instead of "below cost" when describing the pricing AND 2OMITTING THE "REALISTIC PROBABILITY OF RECOUPMENT" REQUIREMENT. This second requirement is important to deal with the concern of courts, as expressed in Matsushita, that competitive pricing is not treated as predation. If a firm couldnt realistically recoup its costs, then it is even more unlikely that it was a trying to predatory price. If there is little chance that a firm could rid itself of its competitors or keep new competitors out of the market, the already unlikely claim of predation becomes almost impossible to believe.

Fisher, by omitting the "realistic probability of recoupment" requirement, removes an important safeguard for competition without justification.

Fisher argues that analysis of predation should be based on the firms knowledge at the time it made the decision to lower its prices, not now. That seems logical. A firm that is trying its best to kill its competitors and does not succeed by some fluke may be guilty of predatory pricing. Schmalensee, however, uses the continued viability of Netscape as evidence that predation did not occur. I think this evidence would be acceptable to a court. In Matsushita, the court considered the fact that the American competitors were still going strong after many years of the conspiracywith even higher market shares than the Japanese competitorsas evidence against predation.


This case is arguably different from a typical predation case because the price charged is zero, rather than a low price. Fisher uses the zero price to argue that he does not have to determine what the cost actually is, since money was spent to develop IE and no money is being charged by MS for it. Fisher says that the zero price avoids our having to determine which measure of cost is appropriate. The problem with Fishers premise to me is that the minute Microsoft can argue that it got some competitive benefit from IE, his zero cost argument falls apart, and he has to evaluate whether the amount of benefit MS got from IE was worth the cost. Also, MS could have gotten a benefit from IE in the past. As long as MS got some benefit from the development of IE at some point, the fact that IE is currently being given away free doesnt help Fisher, since development costs could already have been recouped and the current costs, copying the software, is virtually zero.

Fisher also includes opportunity costs as costs that MS paid to distribute IE. This is a tough thing to prove, I think. The government has to prove MS could have made a certain amount of money and it wanted to. (If MS gives customers a free mug, what is the opportunity cost? Whos to judge whether or not the goodwill makes up for the "cost?") Although particular agreements that Fisher discusses seem to be trading IE distribution or prioritization for something else, like desktop real estate, it is not clear how much that trade was worth to MS. Schmalensee, of course, claims opportunities MS gave up werent worth much.

Schmalensee suggests benefits that MS got from the development of IE. Some of them seem a bit far-fetched, but if even one is possibly true, it seems to me the government has to engage in the analysis of whether or not IE was, overall, distributed below cost. Schmalensee said IE increased the value of Win98. (May be truebut distributing it free decreased the incentive of Win95 people to upgrade.) He claimed that IE allowed MS to sell complementary goods. Fisher denied this and the proof of this was the evidence that MSs goal was to crush Netscape. Logically, both could be true. If MS could in fact sell complementary goods if it distributes IE for free, should MS be stopped from doing so merely because it would also like to crush Netscape? No. MS is acting competitively and gets the extra benefit of hurting Netscape.

Schmalensee also claimed that more Web sites would be written for IE and ISVs would write more applications for IE if IE is more widely distributed, which makes Windows more attractive to buyers. Also, he claimed consumers will want to buy a version of Windows that has IE more fully integrated. He claimed free distribution of IE increased consumer goodwill.

Schmalensee also suggested that the costs of the development of IE were recouped in the sales of Win98, as are the costs of developing all new features of Win98.

Some (most?) of Schmalensees claims arent too believable, but I think he only needs one semi-believable one, and the government has to prove that IE was distributed below cost and not just asserting so from the free distribution of IE.

Also, I think courts may be wary about condemning free distribution, since it encourages all firms (competitive and predatory) to charge a little to cover themselves.


This case introduces a whole new type of predation: defensive predation. Fisher suggests that since MSs OS monopoly was threatened by Netscape, MS reduced its price on IE to hurt Netscape in order to preserve its OS monopoly. MS, Fisher claims, is recouping the costs of IE now or soon by maintaining its OS monopoly; MS never has to charge for IE.

In my opinion, MS would have plenty of incentive to predatory price even if it couldnt maintain its OS monopoly forever. Holding off OS competition even for a year or two would be worth a lot of money to MS. Therefore, Schmalensees pointing out that Fisher didnt show that there werent possible threats in the future or other threats isnt persuasive to me. Fisher claims the recoupment is certain; Im not sure about that. As Schmalensee pointed out, Netscape could be a risk to the OS monopoly with a smaller browser market share. Once Netscape, Java-based applications are written, MSs OS monopoly is threatened. However, recoupment doesnt have to be certain for a predatory pricing claim.

MS would presumably pay a lot of money to maintain its monopoly. But if giving away IE free has some benefit to MS aside from killing Netscape, is the fact that it also kills Netscape enough? Presumably not. Unless the price is below cost, the fact that MS hurts a competitor is not problematic. Fishers evidence that MS knew it was threatened suggests some sort of intent element, but antitrust generally doesnt care about intent, it cares about anti-competitive behavior. (And most firms want to destroy their competitors.)

The big question to me is: what should the test for defensive predatory pricing be? Below-cost pricing and a feeling of being threatenedand maybe also the realistic probability of reducing the threat? With offensive predatory pricing, the "realistic probability of success" is, to me, an important safeguard against overdeterrence. I certainly dont want a test that mainly relies on below-cost pricing (a very difficult thing to determine) and evidence of being threatened (which would, I imagine, be very easy to get generally, since competition threatens firms). Furthermore, feeling threatened and reducing the threat sound like normal aspects of competition. Another firm enters the market with a cheaper product, I feel my market share is being threatened, I lower my prices, and there is the realistic probability of reducing the threat.

Presumption of irrationality? Matsushita and other cases teach us that predation is generally irrational. But is defensive predation irrational? Thats much less clear. It doesnt depend on future domination of a market and the ability to keep other firms from entering; it depends on maintaining a monopoly for the present. It is much more realistic for a monopolist to sacrifice some profit now to maintain his very profitable monopoly now. Furthermore, a monopolist who lowers his prices below cost when a competitor tries to enter the market will discourage new entry (as Fisher indicates) in general. Firms dont want to enter a market with a monopolist who viciously destroys anyone who enters; theyd rather go to a more competitive market. This type of behavior is something the courts would be eager to punish, though there is still the difficulty of determining whether the monopolist lowered his price to his cost or below his cost. This behavior is analogous to our case, in which a monopolist in one market takes a loss in a second market to preserve his monopoly in the first market


The typical predation story is that a firm lowers its prices to force its competitors to leave the market at which point the firm charges monopoly prices, while somehow keeping out new entrants. Generally, this scenario is very unlikelysee Matsushita (although Matsushita involves an even more unlikely conspiracy to predatory price)and it seems very unlikely in this case.

Fisher remarks that MS will make supranormal profits in the second market, which seems to contradict his assertions elsewhere that MS had a plan of giving away IE free forever. And if Fisher believes MS would get profits through complementary services, he has destroyed his argument that MS is not making a profit from giving away IE for free.

Fisher asserts that MS will be able to maintain a monopoly in the browser market because of network effects, but more proof of this seems necessary in order to sustain the "realistic probability of success" requirement. Fisher notes that MS has created a barrier to entry by stopping up the distribution channels for browsers. Fisher also points out that fear of zero pricing can keep competitors out in the future, which I find somewhat convincing. Its difficult to get good data on that.

As Schmalensee points out, it does not seem to be reasonable to assume that MS could completely destroy Netscape. The fact that Netscape is alive and kicking now helps the point (though see discussion about needing evidence from time of decision to lower prices). He also points out the MSs prospects of success were very iffy; MS was dependent on its enemy, AOL, and a contract which AOL could end at will. Schmalensees assertion that the software market is competitive and unpredictable ring true, though his assertion that the OS market is competitive doesnt work for me. Of course, Schmalensee loses a bit of his credibility when he expresses his surprise at the merger of AOL, Sun, and Netscape. I guess he knows the market is unpredictable, but the crazy turns it takes still surprise him.

In a couple of places, Schmalensee seems to run in the wrong direction. Schmalensee, in trying to show the inadequacy of Fishers definition of predation, asserts that Netscape could have had the monopoly on browsers and would have been caught by Fishers overly broad definition. It seems hes assuming that it is possible for one browser to monopolize the market. He also discusses winner-take-all or winner-take-most markets, and browsers seem to be included. If browsers are a winner-take-all market, then MSs aggressive marketing could realistically be for the purpose of taking the whole market. MS could have a reasonable possibility of success in dominating the market.

I am still skeptical about the strength of network effects and the realistic possibility of a monopoly over the browser market.

Altogether, I think the offensive predation story is hard to buy. Furthermore, the government has a high burden of proof, as established by Matsushita.

-- Anonymous, January 21, 1999



Two points (1) The excluding the possibility of independent action language in Matsushita is the test for concerted action under section 1 of the Sherman Act. It is not a necessary element in a predatory pricing claim. DOJ's predatory pricing claim is based on section 2 of the Sherman Act, not section 1.

(2) Recall that Colburn (from AOL) testified that Microsoft paid AOL 25 cents per user switched to IE. In Barksdale's words, this is a "better than free" deal. When price is set below 0, MS has a hard time arguing that it is not below marginal cost. The exclusive dealings contracts with AOL and other ISPs (which only exclude the other top 2 browsers by market share) combined with Microsoft's pricing strategy make a pretty convincing case that Microsoft was engaging in predatory conduct -- a blatant attempt to kill off a young, start-up Netscape.

I have a hard time believing any procompetitive reason for this behavior. Why would Microsoft undergo so much trouble and expense to promote a free product?

-- Anonymous, January 21, 1999

Moderation questions? read the FAQ