I remain agnostic on Staples. I agree that there was lots of helpful econometric evidence and that there was a relatively strong case made for the narrow market, but I also think, again, that that evidence couldn’t account well for post-merger re-positioning or other market dynamics that may have proved the pre-merger analysis wrong.

Or, more broadly, is there something inherently erroneous about defining a market in part through the distribution channel? Whole Foods surely is at the far end of the reasonableness spectrum, but is Staples? What about the department store merger (Federated/May) that the FTC allowed to proceed? Could they have stated a viable relevant market of departments stores (the “sale of clothing and other personal apparel through department stores”, say).

And at the broadest level, wouldn’t such a limit mean any merger among distributors of goods could not be a relevant market?

But more important, and probably in part to Joe’s point, precisely part of the problem is defining allegedly-relevant markets by their channels of distribution–a definition that tends to require more than six words, as we’ve been discussing.

More generally, unless you can define the retail business in a single word (6-5), no go?

Yes, Posner did say (as the AAG quotes) that “the term ‘Chicago School’ should be retired.” But that’s not because its insights are wrong, but because — according to the very article the AAG cites — “by the nineteen-eighties the basic insights of the Chicago School had been accepted pretty much worldwide.”

On the other hand, if one reads “certain” to mean something like “any inframarginal consumer”, then I think your criticism is appropriately aimed. The use of “certain vulnerable customers” certainly suggests to me that what she has in mind are infra-marginal consumers in unilateral effects cases. As to which of these two, or some alternative, is the intended message I am not sure.

I will note that it is in this very set of cases involving differentiated products mergers that it would be quite difficult to “clearly” identify likely anticompetitive effects. I do not think that our ability to understand and predict re-positioning in these markets is sufficiently empirically grounded at this point that an antitrust agency should be using words like “clearly” unless they really mean to limit themselves in that way. Obviously, more confidence can be had in the case of consummated mergers.

I will note that, along these lines, I thought that Joe Sims (Jones Day) had the single best line of the day during the panel on power buyers when recommending that a desirable method of ensuring that the agencies do not take unilateral effects too far is to require that all market definitions be stated in 6 words or less.