On Monday December 12, the Supreme Court of the United States of America denied IBM’s petition for a writ of certiorari in the case of IBM Corp v Allen-Myland Inc, AMI. The decision stems from AMI’s allegation that IBM tied sales of its mainframe reconfiguration services to sales of the components used to reconfigure processors. AMI is an engineering company that refurbished and reconfigured IBM mainframes, generally on behalf of computer leasing companies and used equipment traders. A writ of certiorari the first step in an appeals hearing. Thus, the Supreme Court said that it would not review an appeals decision in the Third Circuit. The Third Circuit had rejected an IBM victory in the lower Federal District Court and remanded the case to the lower court for reconsideration. Although IBM won the initial court dispute on July 21 1988, the lower court’s second decision could well go against it. If the lower court ultimately decides in favour of AMI, the result would be a 20-year setback for IB M and the anti-trust team at the its principal law firm, New York’s Cravath, Swain & Moore. Beginning in the mid-1970s, the Cravath lawyers cut a wide swathe through the courts, toppling every private anti-trust adversary save one, Control Data Corp. IBM reached a settlement with Control Data, and thus did not actually suffer a courtroom defeat; moreover, as part of the settlement Control Data agreed to destroy a computerised index of IBM documents that might have been used against IBM in other cases, including a federal suit filed in 1969. Cravath’s string of courtroom victories culminated in its 1982 defeat of the US Department of Justice after a long and costly battle when the government, in frustration, decided to cease its prosecution.

Dominant products

The cornerstone of Cravath’s legal strategy, formulated by Thomas Barr, who is still with the firm, was a broad definition of the market in which IBM did business. All the anti-trust suits against IBM focused on the mainframe segment, where IBM’s fastest processors and most capable peripheral devices were the dominant products, and where competitors alleged IBM’s practices were abuses of a purported monopoly. In case after case, Barr and his colleagues brought forth witnesses who testified that they had chosen (or nearly chosen) other vendors’ systems over IBM mainframes. These systems included computers of markedly different design from IBM’s flagship systems, such as Tandem systems made of clustered mid-range processors. More recently, Barr’s role as the lead outside lawyer for IBM was taken over by Evan Chesler, another brilliant and aggressive litigator. Under US anti-trust law, companies are prohibited from certain activities when they control a market. These activities, specified in the law, are not illegal in the absence of a monopoly, because customers can go elsewhere, thereby exacting an appropriate punishment for a corporate delinquent. By persuading the courts that all computers, not just mainframes, defined the relevant market, Barr and his legal team were able to show that IBM’s share of the computer business was too small to be considered a monopoly. In the AMI case, the lower court accepted IBM’s definition of the relevant market. But the appeals court saw things differently. In remanding the case, it said that the lower court should carefully examine the case from the perspective that the relevant market was not the whole computer industry, but the mainframe business – and to exclude used IBM mainframes from the relevant market as it reconsidered the case. The findings of the appeals court were based on two key anti-trust cases, Kodak and Alcoa. In 1992, the Supreme Court reviewed the case Eastman Kodak Co v Image Technical Services, Inc. Image Technical claimed that Kodak was driving it out of the copier repair business by refusing to sell it parts. Kodak said that it had a small share of the copier market, so it was entitled to control the related maintenance business if it wished. The Supreme Court disagreed and ruled that the relevant ma

rket was not all copiers or even all copiers of a particular capacity, but the market for maintenance of Kodak equipment. In that case, Kodak had a monopoly and its unwillingness to provide spare parts under reasonable terms was a violation of law. The other major case cited in the IBM v AMI dispute and its appeal is the 1945 appeals court decision in United States v Aluminum Company of America, Alcoa. Alcoa had 90% of the aluminium market in 1945, when it lost the anti-trust suit on appeal.

By Hesh Wiener

Alcoa took the position that the relevant market included not only trade in new ingot, which it conceded it dominated, but also reprocessed aluminium, which was a more diverse business. The court held that Alcoa’s power in new aluminum gave it the ability to influence strongly if not absolutely control the used aluminium market and therefore used aluminium should be excluded from the relevant market on which the legal decision would be predicated. Alcoa was foiled. IBM had never suffered judgements like those meted out against Alcoa or Kodak… until August 12 1994. On that day, its triumph over AMI turned to ashes. Not only did the United States Court of Appeals for the Third District in Philadelphia, send its lower court victory back to reconsideration, but the Circuit Court also strongly urged the lower court to measure IBM’s power in terms of a different market. The appeals court did not decide the case nor specifically set out the market definition the lower court must use, but its ruling is a very strong one; it would be quite unusual for the lower court to review the appeals ruling and then reject its recommendations. The US Court for the Eastern District of Pennsylvania, also in Philadelphia, was asked to consider a definition of the relevant market including new mainframes made by IBM and others and used mainframes made by vendors other than IBM. Used IBM mainframes, the appeals court suggested, might better be excluded from the market because like Alcoa, IBM’s influence over the fate of its own used equipment is so great. The most likely outcome now is a settlement under which IBM would have to pay AMI a substantial sum of money. The terms of the settlement would almost certainly include an agreement by both sides to keep the details secret and very likely a stipulation, which would be made public, that both parties agree that IBM did no wrong. By settling the case privately, IBM could avoid a likely (but by no means certain) defeat based on the narrow relevant market definition. A settlement would also lessen the value of the AMI case in the event IBM faces another similar anti-trust suit, particularly if AMI states unequivocally that at the end of the affair it decided that IBM had in no way broken the law. There is another aspect of the situation that suggests there will now be a private settlement: IBM’s current management seems predisposed to settle legal conflicts rather than pursue risky and expensive litigation, as the outcome of IBM’s disputes with Comdisco, Seagate and Phoenix show; prior IBM management regimes were greatly inclined to pursue prolonged battles in the courts. However, even if IBM and AMI reach a settlement, a great deal of damage has been done. Should any other party find it necessary to sue IBM over anti-trust violations like those alleged by AMI – actions that violate the Sherman Act – it might learn a great deal from the record of the AMI case. An anti-trust action like AMI’s is based on the notion of tying. Basically, a supplier with control over one market (in this case, mainframe CPU components) uses its power to force customers to buy another product (in this case, the technical services required to install the components). The violation of law involves linking the sale of the tied item to the tying item under terms that make it economically impossible for any other party to compete in the market for the tied product. The normal remedy is a court order that compels the monopolist to price and sell the tied item on a fair basis. Specifically, AMI alleges tha

t IBM’s net pricing of upgrades including installation precluded it from selling its engineering services.

Excessive prices

IBM either did not allow it to buy the basic components – the thermal conduction modules, or TCMs that are used in 308X, 3090 and 9021 mainframes – or to buy them only at excessive prices. IBM, AMI asserts, required buyers of TCMs to return other TCMs removed during an upgrade. This requirement was an economic one: when IBM offered TCMs for outright sale, its prices were sky high compared to net pricing involving the return of other parts. IBM did not separately offer to buy unused TCMs at the same trade-in price. The whole process, AMI says, destroyed its opportunities to compete in the market for engineering services and reduced the availability of removed parts AMI might have used to create its own upgrades in competition with IBM. In deciding the original case in favour of IBM, the lower court did not uphold AMI’s Sherman Act claims because IBM proved to the judge’s satisfaction that it did not have a monopoly position. Assuming IBM and AMI do not reach an out-of-court settlement, the new trial will involve a fresh examination of IBM’s market position that takes into account the Kodak and Alcoa cases. Should the court find that IBM does, on the basis of new criteria, enjoy a monopoly in the relevant market, it could award a significant sum to AMI. If IBM decides to fight on rather than settle with AMI, it could emerge with its anti-trust defences intact. The lower court could follow the suggestions of the appeals court, consider the Kodak and Alcoa cases and in the end determine that the relevant market for AMI’s anti-trust complaint should be a broad range of computer systems. And even if the lower court comes around to the opinion that the correct market is the narrow one, it might find that IBM’s actions were completely legal. But to endure another trial, IBM would have to face all the risks it thought it could withstand during its appeal and, subsequently, when it fruitlessly sought an audience before the Supreme Court.

From the January 1995 edition of Infoperspectives International, publishedby Technology News Ltd, 110 Gloucester Avenue, London NW1 8JA (C) 1995 Technology News Ltd. All rights reserved.