Mixed bundling is one of the hottest topics in antitrust, not just since LePage’s. There are essentially three categories of mixed bundling cases, only one of which should be controversial.

1. X makes two products, A and B. The marginal costs of making A and B are $1/unit. In the market for product A, X enjoys significant market power. In the market for product B it doesn’t. Among others, X competes with Y, who only sells product B. Y’s costs are identical to those of X. If purchased separately, X sells product A for $2 and B for $1.50. If purchased in a bundle, X sells product A for $1.50 and B for $0.90. Assuming that recoupment is possible, that’s a rather straightforward case of predatory pricing with respect to product B as $0.90 - $1 < 0.
2. Now suppose that X sells A for $1.50 and B for $1.10 in a bundle. In both markets, X’s prices are above cost and a straightforward predatory pricing claim by Y fails. That’s where the first rule specific to mixed bundling kicks in: All discounts for all products in the bundle must be applied to the product in the market in which X and Y compete, here the market for B. X’s discount in market A is $.50 and in B is $0.40 for a total discount of $0.90. Applying the total discount to B, we arrive at a price of $0.60 ($1.50 - $0.90), which is below the cost of making B ($1). On the basis of the “aggregated discount” rule, Y may bring a predatory pricing claim against X.
3. What if X sells A for $1.80 and B for $1.30 in a bundle? No straightforward predatory pricing as $1.80 > $1 (market for product A) and $1.30 > $1 (market for product B). Also no predatory pricing on the basis of the aggregated discount rule: $1.50 - ($0.20 + $0.20) = $1.10. And, evidently, $1.10 - $1 > 0. This case, and only this case, should be at the heart of the recent controversy. The question is: Can these above cost bundles violate §2?

Here is a table, summarizing the three categories:

Case	Price-Cost	Outcome
(1)	A<0 and/or B<0	Predatory pricing
(2)	A>0 and B<0, after aggregation	Predatory pricing, after aggregation. SmithKline, Ortho.
(3)	A>0 and B>0, after aggregation	Unclear. LePage’s

Leaving aside questions of great practical importance, such as how to determine marginal costs, or more specifically, what properly constitutes variable and fixed costs in calculating average variable costs, the commonly accepted proxy for marginal costs, category (3) cases present us with a value judgment that so far has not been resolved: Should we require a multi-market firm to behave in every market as if it was a single-market firm? Or should a multi-market firm be allowed to cross-subsidize lower (above cost) prices in one market by gains from another market? There are very strong arguments for a per se legality rule, as proposed by our co-editor Dan Crane. Any counter-argument, I suppose, would have to proceed along the following lines:

• Expand the time horizon. Short term, customers will benefit from mixed bundles, because mixed bundles result in lower prices for all three relevant customer groups: those who only buy B, those who only buy A, and those who buy both A and B. But long term they will suffer from a lack of options and resulting higher prices. Given our inability to make reasonable long term predictions, this argument faces the same problems as other recoupment stories, and on its own is probably rather weak.
• Include variety into the definition of consumer welfare. What makes above-cost bundling seem innocuous from a consumer welfare perspective is the almost exclusive focus on price. But retarding the growth of Y in market B by A’s cross-subsidies may well lead to diminished consumer choice not just in the long run. For some homogenous products, that probably doesn’t matter much. But I can think of markets in which I value the availability of choices more than price.

Comment feed: RSS 2.0. You can leave a response, or trackback from your own site.