Wednesday, September 19, 2007
Posted by Bill Page and John Lopatka
I’ve had something to say about the CFI decision, but I should post something more directly about our new book, The Microsoft Case: Antitrust, High Technology, and Consumer Welfare. Much of the analysis in the book is relevant to the issues raised in the European case. Chapter 4 of our book, for example, addresses the problem of integration in the U.S. case, particularly how the D.C. Circuit approached the question of integration of the browser and the Windows operating system (the OS). We first examine the court’s analysis of the short-run effects of three levels of integration on consumers. We then examine its treatment of the long-term effects of integration—that is, whether Microsoft’s integration of the browser and the OS actually caused a reduction in competition by preventing Netscape’s browser (with or without Java) from developing into a rival platform that might have reduced the network effects (the “applications barrier to entry”) that protect the Windows monopoly.
1. Short-Run Effects on Consumers
The three levels of integration we discuss are (1) simple bundling of the browser and the OS; (2) preventing computer manufacturers (OEMs) or users from removing visible means of access to browsing functionality (icons, etc.); and (3) commingling browser and OS code. We suggest that none of these forms of integration harms consumers so long as all browsers are free and OEMs and end users are able to install rival browsers. (Both of those conditions are necessary for the analysis.) Microsoft may gain an advantage over rivals by bundling its browser with the OS, but only to the extent consumers benefit by receiving a free browser pre-installed. We see no reason to think allowing OEMs to delete access to the pre-installed browser would benefit consumers significantly; the greater consumer benefit would be for the OEM to pre-install a rival browser and allow the user to choose between them. (The notion that consumers would be confused by the sight of two browser icons is not plausible.) Indeed, OEMs might delete Microsoft’s browser as part of an exclusive deal with a rival browser producer, not to benefit consumers. One might argue that consumers will benefit if the rival outbids Microsoft for exclusivity on the desktop; the rival can pay more than Microsoft only if consumers prefer the rival’s product. But Microsoft was not allowed to participate in such an auction – it could not obtain exclusive placement. All of that said, we so no real harm in the eventual remedy of allowing end users flexibility to delete visible means of access the browser or to set their own preferred defaults, unless there are good technical reasons to override those choices.
Commingling browser and OS code does not harm consumers in a
relevant way for all the same reasons, and with the same qualifications. Requiring deletion of browser-only code, on
the other had, could potentially harm consumers by fragmenting the Windows API
set, as the courts found in the
2. Long-Run Effects on Consumers
The court of appeals held that the government did not have to prove Netscape’s browser would really have become a rival platform and eroded the applications barrier to entry. It was enough that the browser was a “nascent” rival and was viewed as such by Microsoft. We argue, however, that the government should have at least been required to offer a coherent theory showing how Microsoft’s actions in the browser market could have reduced competition in the OS market, where Netscape did not compete. We further argue that the government’s failure to prove there was a browser market doomed the most plausible scenarios under which the browser might have become a rival platform. Finally, we argue that the fact (and it was a fact) that Microsoft viewed Netscape as nascent competitor should not have been sufficient to condemn the actions it took against Netscape. Just as the plaintiff in a predatory pricing case must prove that the defendant will likely recoup its losses after the predatory campaign (that is, that competition will really be reduced) the government in Microsoft should have been required to show that the middleware threat was more than a mirage.