“If, as a number of other commentators have suggested, competition is not hampered because other venues for coffee shops are readily available for Starbucks’ rivals, is Starbucks paying hard cash for an illusory benefit?”

Nope. It simply does not logically follow from the proposition competition (in the coffee market, to be clear) is not hampered by the exclusive leases that Starbucks must be paying for an illusory benefit. There is, as Harry knows, a voluminous literature on the efficiency-enhancing properties of exclusivity (and de facto exclusivity) that have nothing to do w. anticompetitive exclusion. Firms with market power and firms with little or zero market power in highly competitive markets adopt exclusive arrangements throughout the economy for a variety of reasons.

While that literature is too voluminous to do justice to in a post, and I hope that Harry does not accuse me (again) of “blithely” ingoring important disputes. I do not think it is difficult to think of pro-competitive explanations for Starbucks’ desire to not have rival coffee shops in the immediate vicinity. The bottom line is that exclusive arrangements can increase the joint surplus earned by the landlord/ SB combination and that these premiums represent a sharing of the premium — not Starbucks’ ignorance.

But as your comment recognizes, it makes no difference how powerful Starbucks is in coffee if it doesn’t have market power in real estate. That is the sense in which I mean that the extent of Starbucks’ market power in real estate is the relevant inquiry here. Sure, if one could show market power in real estate (which, as you recognize, is quite a hurdle) it then might be relevant under Section 2 whether Starbucks is a monopolist in coffee (itself a dicey proposition) or at least in a properly defined market in “high end coffee” (which is also unlikely).

I like your second point about the rent premium. Yes, either there is one or there isn’t one. I guess David’s point was that it’s doubtful whether there’s a premium at all. But even if there is a premium, Geoff and Josh have explained that it’s the landlord that has the market power in the building lobby “micro market” (if I may coin such a term), not Starbucks, and if the landlord selects Starbucks rather than some other coffee store as sole coffee shop tenant, that discretionary amenity choice is itself the product of a fiercely competitive real estate leasing market.

I find one point particularly interesting. Landlords seem to be demanding a premium for exclusivity. This makes sense because the landlords are narrowing their potential universe of leasees. Allegedly, Starbucks is paying the premium. What does Starbucks get out of the deal? If, as a number of other commentators have suggested, competition is not hampered because other venues for coffee shops are readily available for Starbucks’ rivals, is Starbucks paying hard cash for an illusory benefit?

The question is whether this distinction makes any difference for the competitive analysis, which is I addressed in my comment (as did Geoff’s analysis of why a landlord might desire such a term, which might be read as making the case that this distinction cuts in favor of a pro-competitive interpretation of SB’s conduct).

Pro-competitive justifications aside, the question is whether Starbucks’ exclusive dealing contracts can harm competition by foreclosing rivals from access to real estate. The point I made above, which I will now reiterate, is that the necessary conditions for such harm are highly unlikely to be satisfied here. The alternative argument, that any given building is an essential facility and therefore the exclusive should be a violation of AT law whether or not there is substantial foreclosure, is absurd (there is a lot of real estate that can potentially be used as a coffee shop should SB’s would-be supra-competitive prices attract entry…). Perhaps these lease terms are of such a scope as to actually prevent rival coffee shops from entering into agreements to obtain retail space in sufficient magnitude to prevent entry throughout a well-defined market. But I doubt it.

This is a distinction with at least some relevance to this discussion.