Interchange Fees: The Economics and Regulations of What

16 INTERCHANGE FEESneed to balance the supply and demand of card transactions. This seems somewhat artificial, sinceas long as some merchants have agreed to accept a card, the volume of transactions is determinedunilaterally by card-carrying consumers’ deciding whether or not to use the card for particulartransactions. Nonetheless, the proposition that at some interchange fee levels the system wouldnot be viable is likely to hold in more general models. The important—and robust—insight fromBaxter’s analysis, although it is not framed precisely this way, is that the interchange fee helpsinternalize an externality between the two customer groups and, in so doing, has the potential ofmaking both customer groups better off.C. Imperfect Competition in Issuing and AcquiringThe assumption of perfect competition between homogeneous issuers and acquirers is notrealistic and leads to most interesting questions having indeterminate answers—since there areno profits to be had, there is no motivation for doing anything. The next logical step was takenby Richard Schmalensee. 40 He made the standard assumption that, except for the payment systemunder study, the economy was perfectly competitive. 41 This implies that the demand systemfacing the system, which Schmalensee took as given and did not derive from first principles,could be used for standard welfare analysis. Schmalensee allowed for imperfect competitionamong issuers and/or among acquirers and made a particular assumption about the functionalform of the demand system and, thus, about the structure of indirect network effects. As in theBaxter analysis, the system itself was assumed to operate like the Visa and MasterCard systemsoperate in fact, on a break-even basis, and fixed costs and access prices were again assumed away.Thus, as in the numerical example above, the interchange fee simply shifts costs between issuersand acquirers, raising costs on one side of the market by exactly as much as it lowers costs onthe other side.In this model, the level of the interchange fee can affect the profits earned by issuing andacquiring banks whenever competition among them is imperfect, making it possible to compareprofit-maximizing and welfare-maximizing fees. Under some special assumptions, the comparisonis simple: when there is a single issuer and a single acquirer and demand curves are linear,for instance, Schmalensee shows that the profit-maximizing interchange fee also maximizes systemoutput and economic welfare. In this case, regulation could only reduce overall performance.In general, however, these three fees may be different even under the particular demandstructure Schmalensee assumed, and even when the further assumption of linear demand isimposed. Thus, even under strong assumptions about demands and costs, the relations among40Schmalensee (2002).41Much of the subsequent literature has persuasively called that standard assumption into question in this context, aswe discuss in the next section. Nonetheless, it is instructive to see its implications before considering how it might bealtered in the direction of greater realism.