Thinking about the European Microsoft Case: F.M. Scherer Discusses the Relevance of the IBM Case.

Reproduced with permission from Antitrust & Trade Regulation Report, Vol. 84, No. 2090, pp. 65-66 (Jan. 24, 2003).

Copyright 2003 by The Bureau of National Affairs, Inc. (800-372-1033)

F. M. Scherer

January 2003

There are remarkable parallels between the Microsoft case currently pending before the competition policy authorities of the European Union and the action pursued by the European Commission during the early 1980s. There are also valuable lessons for responding to Microsoft conduct that almost surely will be found to violate EU competition law.

Unlike the U.S. Department of Justice complaint against Microsoft, which targeted Microsoft's conduct during the mid- to late 1990s, the European investigation appears to be focusing upon two still-contemporary issues: Microsoft strategies that, by keeping certain interface information and programming tools secret, give its workgroup server operating systems a significant advantage over independent server software writers attempting to achieve inter-operability with ubiquitous Microsoft desktop computer operating systems; and the bundling of Microsoft's audio and video playing software with its desktop operating systems.

The European Commission's investigation of IBM during the early 1980s also emphasized bundling and non-disclosure of interface information. Beginning in the late 1960s, IBM experienced increasing competition from "plug-compatible" manufacturers (PCMs). Some like Telex, Memorex, and StorageTek built peripheral equipment such as tape storage drives, disc drives, and add-on memory units that plugged into the standard interfaces used on IBM System 360 and then System 370 mainframes. Others such as Amdahl (founded by the principal designer of IBM's System 360), Fujitsu, and IPL Systems, produced main frames (i.e., central processing units) that could be used interchangeably with IBM computers, but offered higher performance and/or lower cost. To combat these threats, IBM among other things began bundling peripheral equipment control functions into mainframe hardware. It also departed from what had previously been a full disclosure policy, keeping operating system software source code secret, reducing the amount of information disclosed on interface specifications, and delaying interface disclosures until they came too late to be of use by PCMs trying to design compatible equipment.

In response to complaints by several American computer manufacturers, the European Commission's competition policy directorate commenced an investigation of IBM's bundling and disclosure policies. After the U.S. Department of Justice abandoned its 13-year-old antitrust suit against IBM in January 1982, William F. Baxter, then head of the U.S. Antitrust Division, contacted the director of the European Commission's competition unit and "expressed concern" over continuation of the European action. The Europeans persisted, however, and were on the verge of issuing a formal complaint when IBM agreed in August 1984 to an "undertaking" accepted tentatively by the EC in lieu of an official mandate.

In the undertaking, IBM agreed to disclose within four months of the announcement of a new System 370 product, or at the time a product became generally available, if earlier than the four-month deadline, interface information sufficient to allow both hardware and software manufacturers to attach their offerings to IBM main frames and to its Systems Network Architecture. It also agreed to sell within the European Common Market unbundled System 370 computers with no more main memory than what was required for testing purposes. Violation of the undertaking, an EC official announced at the time, could lead to reinstatement of formal case proceedings. The undertaking was to remain in force until at least January 1990. In fact, it was discontinued in July 1994.

The undertaking reflected a compromise between IBM and EC authorities. Originally, the EC staff sought disclosure of interface specifications within 30 days of product announcement; the compromise settled on 120 days. The staff also sought disclosure of underlying operating system source code, but this provision was not covered by the undertaking. No guidelines were provided for compensation to be paid by PCMs receiving the IBM information, and so IBM in fact was able to charge what one PCM veteran said to the author was "quite a bit of money."

One moral of the history is that the European Commission sought to remedy perceived competition policy problems even though the U.S. Department of Justice chose to abandon its antitrust case and urged the Europeans similarly to withdraw.

In interviews, two former U.S. plug-compatible manufacturer executives expressed to the author differing views on the impact of the European undertaking. One important rival of IBM stated that, given the secrecy policies IBM had previously implemented, the undertaking "was absolutely essential to our ability to compete." Another former PCM executive said that the settlement had little effect, mainly because by the mid-1980s IBM had such a huge base of computers in use by customers that it was forced to standardize its interfaces and keep their specifications relatively constant. He distinguished the situation in the 1980s from current Microsoft disputes, observing that software operating system interfaces can be changed much more easily, and with more debilitating consequences for rivals attempting to maintain compatibility, than the hard-wired interfaces embodied in IBM main frames.

From the European IBM case experience, several lessons to help inform impending European decisions on Microsoft can be extracted. First, unbundling remedies are not unprecedented in the information technology realm. Second, timely disclosure of interface specifications is crucial to the attainment of inter-operability. Prompt reporting of post-announcement software modifications that affect interfaces is equally essential. An IBM competitor reported that following the 1984 undertaking "IBM was pretty good in disclosing bug fixes." Third, interviewees made it clear that although they were able to function without access to underlying IBM source code, they could have achieved even better plug compatibility with relevant source code information. Fourth, the question of compensation for timely disclosure of interface and source code information cannot be overlooked. It would not be inappropriate for the European Community authorities to reduce the fines they might otherwise levy and correspondingly require that Microsoft disclose information at zero monetary cost to firms developing Microsoft-compatible software. Finally, in view of Microsoft's perennially uncooperative stance toward antitrust inquiries and remedies, the European Union would be well advised to issue a formal order mandating desired behavioral changes rather than accepting, as it did in the IBM case, an undertaking enforceable only in cases of non-compliance with a substantial and deleterious lag.


"IBM Shocks the Plug-Compatibles," Business Week, June 18, 1979, p. 107.

"I.B.M. Faces Europe on Antitrust Charge," New York Times, January 19, 1981.

"IBM's Mimics Struggle To Keep Pace," Business Week, September 20, 1982, p. 93.

"Europeans Aim Action at I.B.M.," New York Times, April 27, 1984.

"Under the Gun: Common Market Officials Are Still Watching IBM Closely, Although Perhaps Not Close Enough," Datamation, September 1, 1984, pp. 42-48.

Organisation for Economic Co-operation and Development, Competition Policy in OECD Countries: 1984-1985 (Paris: 1987), pp. 253-254.