SIEPR Discussion Paper No. 00-51 Network Effects and Microsoft Timothy F. Bresnahan Stanford University
Stanford Institute for Economic Policy Research Stanford University Stanford, CA 94305 (650) 725-1874
The Stanford Institute for Economic Policy Research at Stanford University supports research bearing on economic and public policy issues. The SIEPR Discussion Paper Series reports on research and policy analysis conducted by researchers affiliated with the Institute. Working papers in this series reflect the views of the authors and not necessarily those of the Stanford Institute for Economic Policy Research or Stanford University. ____________
Network Effects and Microsoft
Timothy F. Bresnahan* Abstract
Economic theories of network effects have a number of testable implications. Testing of the theory by econometric methods, however, has proved difficult. I turn to documentary methods in order to test the theory, drawing on the body of documents used as evidence in U.S. vs. Microsoft. Analysis strikingly similar to the theory, including not only the main implications but also key analytical distinctions, guides business decision making. The same analysis suggests several interesting areas in which the theory is incomplete. Document based research raises a number of novel methodological issues.
Professor of Economics, Stanford University, and Gordon and Betty Moore Senior Fellow, Stanford Institute for Economic Policy Research. Email address: email@example.com. I am grateful to several people for comments, but especially to Jacques CrJmer. While I served in the Antitrust Division of the United States Department of Justice and continue, at this writing, to consult to the Division, this paper does not represent the opinion or policy position of the Division but only my own view.
Network Effects and Microsoft
The economic theory of network effects has received a great deal of sustained attention, as it appears to capture some of the most important features of modern high tech industries.1 The positive implications of the theory are important for understanding the structure of those industries, especially over time. They include positive feedback in the decision rules of individual actors, indeterminacy of equilibrium, lock-in to particular network standards, first-mover advantages or barriers to entry, high inertia for established standards but high volatility for nascent ones, and strategic competition that is intense in the period of establishing a network standard, then largely absent after lock-in. This is not the simplest body of implications to test, for two reasons. As the theory involves strong elements of positive feedback and the resulting coordination, econometric testing faces severe difficulties associated with distinguishing the behavior of different actors.2 Another critical implication of the theory, multiple equilibria and the resulting indeterminacy, poses very difficult problems for the empirical scholar of attempting to observe what didn’t happen.3 This paper tests the theory by looking at business documents from the Microsoft antitrust case. While the documents are public because attorneys thought they would be useful in the policy context of a trial, my use is entirely positive, not normative. The unique perspective offered by internal documents gives us an opportunity to examine the relationship between the theory and the marketplace in several ways. Microsoft is a very analytical firm, and thinks of itself as involved in complex strategic games involving many outside agents. This means, first, that there is a great deal of internal discussion of the theory of behavior of other agents in the marketplace, including customers, competitors (where they exist) and complementors.4 Further, many of the documents take as their...