05 June 2014

ECB Working Paper: Paying for payments, free payments and optimal interchange fees

This paper examines how interchange fees influence the choice between cash and payment cards. It shows that when consumers do not pay transaction fees to banks - a common feature in bank contracts - card use is declining in interchange fees, and surcharging does not neutralize interchange fees.

When consumers make a card payments, an interchange fees is typically paid from merchants' banks to the consumers' banks. Interchange fees can, in principle, help remedy an imbalances in the distribution of cost and benefits across the consumer's and the merchant's sides of the market. As an example, suppose that consumers derive relatively low benefits from the use of payment cards relative to cash, whereas merchants enjoy substantial cost savings from accepting cards. If, then, the bulk of the cost of producing card payments is borne by consumers' banks, it may not worthwhile for consumers' banks to produce card payments even though the joint benefits of consumers and merchants exceed the joint costs of the banks. An interchange fee is one means of solving this problem.

Much of the earlier literature on interchange fees highlights this potential efficiency. It is argued that interchange fees, if set optimally, increase the use of payment cards. A number of papers also suggests that even though interchange fees are often agreed upon collectively (multilaterally) by banks, they do not function as monopoly prices. This feature is due to the two-sided nature of the market for payment instruments. Furthermore, it is not clear that banks systematically set interchange fees either too high or too low relative to the social optimum. Moreover, even if private interchange fees were biased, it would not be possible to a regulator to set an optimal fee because such a fee would depend on unobservable factors. Taken together, these conclusions make a weak case for regulation of interchange fees.

In this paper, author (Mr Søren Korsgaard) shows that these conclusions are reversed when banks do not charge transaction fees to consumers. A central assumption in the earlier literature is that higher interchange fees translate into lower marginal costs for consumers when paying by card. In practice, however, payment services are often provided either for free or against periodical fees: Consumers' marginal cost is therefore zero, independent of interchange fees. The implication is that as interchange fees increase, merchants' costs of accepting card payments rise, and fewer merchants will accept cards, whereas consumer behavior will remain unchanged. As a result, higher interchange fees are associated with less card usage. Author also finds that banks systematically set interchange fees above the social optimum. Interchange fees do, in fact, function as monopoly prices. The upshot is that there will be fewer card payments than is socially optimal.

A striking result of the model is that the socially optimal fee depends only on the difference in marginal cost of producing card and cash payments. Given adequate data on the cost of producing payments, it can therefore be calculated. The paper therefore not only shows that interchange fees are biased, but provides a potential method for determining the size of the bias. These findings suggest a positive role for the regulation, in particular the lowering of interchange fees as has been proposed by a number of authorities.

Mr Korsgaard also analyses the effects of surcharging. In his model, surcharging does not "neutralise" the effect of interchange fees, as is often argued in the related literature. Intuitively, this is because interchange fees are not passed on to consumers in the first place. In the model, surcharging has two effects. First, surcharging makes it less attractive for consumers to use cards. Second, banks will optimally choose a lower interchange fee, making it more attractive for merchants to accept cards. On balance, though, surcharging will be associated with lower card usage.

Finally, he calibrates the model to data on the cost of producing payments. Analysis of the calibrated model shows that optimal fees are likely to be close to zero, possibly even negative. This is because the optimal interchange fee depends on the difference between the marginal cost of producing card payments and the marginal cost of producing cash payments. Since the marginal cost of producing card payments is lower than that of cash payments, a social planner will therefore want to support to use of the use of cards relative to cash. As the model predicts card usage to be a decreasing function of the interchange fee, the optimal response is to set a low or even negative fee.

Full working paper

© ECB - European Central Bank