Bundling of private goods is fundamentally anti-competitive and, accordingly, reduces innovation.I'm sure he doesn't really mean this. Wal-Mart bundles groceries and soft goods. Such bundling can have deleterious effects on less clever competitors, but that is not the same thing as lessening competition. Amusement parks bundle all sorts of things. Power companies bundle power by day and power by night. There has to be some qualification to this claim.
Prohibiting Microsoft from bundling, say, media players and search engines into the Windows operating system is critical to preserving competition and promoting innovation.That's not the qualificatin I'm looking for. Perhaps I'm showing my age, but there's something about contemporary high-powered operating systems with graphic user interfaces and embedded programs that makes me skeptical about defining bundling in software in any non-trivial way. As the operating system has to be able to work with the applications, whether an application counts as bundled or not strikes me as part art. Programmers of a certain age will remember punch cards and subroutines. Quibbles over whether a program is part of the operating system or free-standing are often as foolish as arguing whether the cards should be placed as a subroutine behind the  ' END card or simply linked to as required, say by a computed  ' GO TO within the deck.
The Barry Nalebuff paper Professor Bainbridge cites promises less than he claims.
In this paper, we look at the case for bundling in an oligopolistic environment. We show that bundling is a particularly effective entry-deterrent strategy. A company that has market power in two goods, A and B, can, by bundling them together, make it harder for a rival with only one of these goods to enter the market. Bundling allows an incumbent to defend both products without having to price low in each. The traditional explanation for bundling that economists have given is that it serves as an effective tool of price discrimination by a monopolist. Although price discrimination provides a reason to bundle, the gains are small compared to the gains from the entry-deterrent effect.Any such oligopolist has to face one further problem: the prices of A and of B must be such that no rival will be tempted to produce either A or B, because the incumbent firm's prices would allow such a rival to cover the stand-alone costs of one product.