Warren Grimes of the American Antitrust Institute (and Southwestern Law School) defends the Korea Fair Trade Commission’s recent Microsoft decision holding that Microsoft abused its dominant position by bundling Instant Messenger and Windows Media Player technologies to its operating system, and its order that Microsoft must sell unbundled versions. Grimes central point is that not we should expect such divergent views in an antitrust environment of “world federalism,” but on that the KFTC was in good company in viewing Microsoft’s bundling practices as anticompetitive.
The Department of Justice and Ronald Cass criticized the KFTC decision, and I piled on a bit while guest blogging at Ideoblog. Grimes thinks they both got it wrong, and goes out of his way to criticize Cass, I think unfairly, accusing Cass of taking the position that “that all bundling of two products is procompetitive.” The problem is that Cass did no such thing. In fact, I criticized Cass for not going far enough in recognizing the pro-competitive aspects of bundling and metering by firms with market power. Specifically, Cass wrote:
In rare cases, bundling doesn’t promote efficiency. Early tying cases probably were examples of firms trying to exploit a monopoly — not by extending it to another product, but by using a complementary product (ink, punch cards) to distinguish among buyers based on how much they value the base product (mimeo machines, computers).
While Cass goes on to note that while “Newer theories, depending on very peculiar assumptions, posit that tying can harm competition . . . in reality, bundling doesn’t hurt competition,” he is not guilty of the assertions Grimes’ alleges. Cass is merely recognizing that the conditions under which bundling can be anticompetitive do not frequently accord with empirical realities. Perhaps I am more worthy of Grimes’ criticisms. I have written the following about bundling as a form of metering demand:
Metering is a form of price discrimination, which despite its nasty label and disfavored position in antitrust, is quite often good for consumers. The static welfare effects of price discrimination are generally ambiguous as some consumers gain and others lose, but likely to be positive if the metering device allows accurate measurement of the intensity of demand since output will increase. When one also accounts for dynamic consumer welfare gains, price discrimination is likely to benefit consumers unless linked to some independent antitrust wrong.
Grimes is not convinced, and points to the fact that he has plenty of company:
“the Justice Department and a majority of antitrust scholars and practitioners who have addressed these issues believe that a dominant firm’s bundling behavior can be strategically employed to stifle competition and innovation.”
Please allow me to register my dissent from this alleged majority….
I agree that it is a well accepted result that bundling can be anticompetitive under a very strict and highly stylized set of conditions created in formal economic models. Such models by top economists fill the pages of economic journals which specify mechanisms under which bundling might be anticompetitive (Whinston, Carlton & Waldman, and Nalebuff are a few prominent examples). These models offer important contributions to the economics literature because they identify the possibility of competitive harm.
However, let us not forget that competition takes place in markets, and not in models. These models typically expressly recognize that they: (1) exclude consideration of any pro-competitive explanations for bundling; (2) do not tell us much about whether bundling is on net, a pro-competitive or anticompetitive practice. Some even explicitly warn against attempting to operationalizing these models to into antitrust rules.
In sum, I’m not sure whether the fact that a majority of antitrust scholars think that bundling might be anticompetitive really matters. What matters is increasing our understanding of this practice in markets, and most importantly, its impact on consumers. As I wrote before:
“it is increasingly important that domestic antitrust jurisprudence sensibly addresses the ubiquitous practices of bundling, metering, and price discrimination more generally. The forthcoming SCOTUS opinion in Independent Ink is a valuable opportunity to clarify the law in this area.”
RSS - Posts
I don’t understand how anyone thinks that customers are better served by a less feature rich product.
Because if they get a less feature rich product, they aren’t forced to pay for the features as a condition for getting the product. Rather, Microsoft and other companies can compete in the market by selling the features separately.
If Microsoft started charging for an IM client, Media Player, or calculator the haters of success would start piling on about how greedy they are and how they are just trying to make a buck.
You’re putting words in people’s mouths.
And even if people did call those things “greedy” there’s a distinction between how greedy an action is, and how likely the greed is to cause harm. If Microsoft charged people $500 for a calculator program, that might be called “greedy”–but it’d be *unsuccessful* greed, since consumers would just refuse to buy it.
I don’t understand how anyone thinks that customers are better served by a less feature rich product.
If Microsoft started charging for an IM client, Media Player, or calculator the haters of success would start piling on about how greedy they are and how they are just trying to make a buck. Microsoft does not prevent PC makers from installing additional software on the base OS, I’m at a loss how Microsoft’s practices are anti-competitive.
Josh. Nice post. It is a pleasure to see an economist put the models that are published in leading journals in perspective, and to help less sophisticated readers realize that they are not (or should not be!) intended to be used to provide policy guidance.