Yet a surprise was in store. Out of the blue and at the 11th hour, Sen. Richard Durbin (D, IL) proposed an amendment to the Dodd-Frank bill that required the Federal Reserve Board to implement a comprehensive system of debit interchange transactions, i.e., the fees that merchants are charged for processing debit transactions through Visa and MasterCard. During his rambling remarks on the Senate floor, Durbin noted that the chief executive officer of Illinois-based Walgreens drugstores (whom Durbin did not name, but who is in fact Gregory D. Wasson) complained that interchange fees (presumably for both credit and debit cards) had become his firm’s fourth largest cost item after wages and salaries, mortgage and rent, and health care.

Durbin took up the cause of his favored constituent to give merchants a “fighting chance against the debit card companies.” Without committee hearings by either chamber of Congress, he proposed (and in all likelihood the merchants had drafted) a price-capping arrangement for debit (but not credit) card interchange fees that had never been proposed or discussed in the extensive academic and industry literature on the subject.

The boldness of his proposal is palpable. Right now, merchants pay three types of fees. The first is to their own bank for organizing their transactions. The second is to the platform operator (e.g., Visa, MasterCard) for its services. And the third is to the card-issuing bank for its services. The Durbin Amendment does nothing to regulate the first two charges, but it radically transforms the debit interchange fee to issuing banks.

Here are the barebones essentials to the scheme. For large banks with assets over $10 billion, the amendment limits interchange fees to an amount that is “reasonable and proportional to the cost incurred by the issuer with respect to the transaction.” Bowing to political pressure, Durbin created a two-tier system that exempted from his amendment small banks, defined as those with under $10 billion in assets. Only about 60 banks (and three large credit unions) are currently large enough to qualify, as compared to about 7,000 banks and 7,500 credit unions that are not. The importance of this exemption becomes evident only after looking at the statutory text of “reasonable and proportional,” which states that “incremental” costs of the “authorization, clearance, or settlement [i.e., ACS fees] of a particular electronic transaction” may be considered, but “other costs” that are “not specific” to a particular electronic transaction shall not be considered. The amendment leaves some discretion in ascertaining how to “consider” the enumerated costs, but it leaves no discretion in determining how not to consider costs. The only sure way to do that is to ignore them.

Given its statutory mandate, the Federal Reserve Board collected extensive survey data to identify the various elements of allowable costs. The included banks ranged from the four huge banks — Bank of America, with $2.3 trillion in assets; J.P. Morgan Chase and Citibank, each with about $2 trillion in assets; and Wells Fargo, with about $1.2 trillion in assets — to about 60 other banks of varying size. My client, TCF Bank (formerly Twin Cities Federal), has about $18 billion in assets, less than 1 percent the size of the three biggest banks. Yet since TCF specializes in retail banking with debit cards, it has no credit card business. The fee restrictions of the Durbin Amendment are far greater as a percentage of TCF’s revenue than they are for most of the major banks — which is why it decided to launch a full-scale constitutional challenge to the amendment. I will discuss that challenge presently.

The Numbers on Debit Interchange

To set the stage for the constitutional question, it is critical to understand how deeply the Durbin Amendment alters established ways of organizing debit transactions. Today, the average debit card transaction is about $35, carrying a 1.35 percent debit interchange fee that Visa or MasterCard (in addition to their own smaller fees) pay to the issuing bank, which works out to about 47 cents per transaction. For TCF, the typical allowable ACS charges would be about 9 cents of that total, for a reduction of over 80 percent in fees.

Many financial institutions pressed ingenious theories on the Fed to give a broader definition of allowable fees, but the Fed stuck to a narrow definition that led it in December 2010 to propose two possible tests: The first was a narrow definition of ACS fees, which were to be approved in the individual case up to 12 cents per transaction. That definition took a restricted view of what it meant to authorize a transaction, such that TCF’s recoverable rate would be 4 cents per transaction. The Fed was undecided on whether to introduce a second payment window by which covered banks could obtain as of right 7 cents per transaction, even if their actual costs were lower.

This payment system cuts out the rich variety of rate structures that are now in use, and blocks in advance others that might become sensible with new technology. But far from worrying about those issues, many retailers howled that this 7‑cent figure promised a 700 percent profit over the penny they claimed was needed to process the debit card transaction.

Perhaps retailers are not clamoring to get into the debit card business because they know that these ACS expenses do not cover all costs. For instance, Durbin’s statutory formula excludes, at a minimum, the fixed costs that are needed to design, build, and maintain an extensive debit card apparatus to deal with the massive flow of debit transactions. The costs also exclude large, non-electronic variable costs, including customer service, billing practices, and advertisement and promotional activities, which are designed either to bring people into the bank or to keep them as clients once they have signed up. Under current practices, all these basic services are provided free to customers whether they use credit, debit, cash, or checks.

The competitive decision not to charge customer fees means necessarily that these interchange fees are critical for a bank like TCF, which had, even before the Durbin Amendment, a very high turnover rate in its customer population, reaching between 30 and 35 percent per annum. Low- and moderate- income customers are highly price sensitive. In order to obtain from them the revenue lost through the interchange system, either monthly fees or swipe fees would cost just under $100 per year for an average account, which could easily send many of the bank’s customers scurrying off to an exempt bank that can continue to offer free debit cards.

The narrow definition of “reasonable and proportional” has also spawned serious concern in both the House and Senate, including a recent public pronouncement from Rep. Barney Frank (D, MA) stating that he was prepared to work with Republicans to “replace or repeal” a law that has battered the price of Visa and MasterCard stock, just as his more recent January remarks attacking the amendment gave a boost to their value. In time, some prospects for legislative amendment may yet emerge, although getting a change of heart in the Senate could prove difficult. In the meantime, TCF has filed its constitutional suit in the District Court of South Dakota before Judge Lawrence L. Piersol, a Clinton appointee to the federal bench.

The Economics of Two-Sided Markets

The legal challenges to the Durbin Amendment require understanding that debit (and credit) card systems are organized as part of a two-sided market. These markets operate on centralized platforms, which are commonly but not exclusively supplied by Visa and MasterCard. In one sense, the term “two-sided market” sounds otiose. After all, all markets have buyers and sellers, thereby constituting two sides of the market. So why should debit cards be different?

The key answer to this question was supplied by the late William Baxter, writing in the Journal of Law and Economics in 1983. In an ordinary market, each buyer and seller goes about his business without caring whether or not other traders succeed or fail. Thus the seller of apples wants high prices from customers who seek low prices. So too for buyers and sellers of bread. The market for each product reaches equilibrium when the price equalizes the quantities both demanded and supplied.

Two-sided markets do not have this independence between separate products. In virtually all cases, the ability to satisfy one side of the market depends on the continued participation of the other. Thus a traditional singles bar that only attracts male patrons would quickly go out of business. As an empirical matter, it is far easier to get men to attend singles bars than women, implying that a uniform price for drinks for all customers results in an excess of males, at which point the market can collapse. What is needed is to find a cost-effective way for the men to subsidize the purchase of women’s drinks, which meets the following constraint: the total prices paid by the men and women together must exceed the total costs of supplying drinks to both groups — including the transaction costs of putting the deal together.

The challenge is determining how to arrange for the needed cross payments. Assume that the uniform price of a drink is $10, when men on average are willing to pay $15 to attract women, while the women will only pay $8. The total benefits are $23 and the total costs are $20. The shortage of women will lead the men to stay away as well. But charging women $8 and men $15 more than covers costs, while also equalizing the number of men and women. Quite simply, some portion of the men’s surplus is paid to the women in the form of differential prices. There is, of course, no direct payment from any particular man to any particular woman. Instead, the proprietor sets different prices to bring the market into equilibrium, making both sides better off.

The distinctive feature of the two-sided market is the transfer payment between the two groups, which exploits a previously untapped source of gain from trade. The singles bar illustrates the mechanics, while the two following examples show the variety of these two-sided markets. Newspapers must serve both advertisers and readers, and in some cases the entire costs are borne by advertisers to lure in readers with free handouts. Shopping malls are two-sided markets with merchants and customers: sometimes customers are charged parking fees at below cost; in other cases they park for fee. The dynamics of each of these markets differ in subtle ways, but properly executed, all of these complex pricing systems should meet the gold standard for all voluntary transactions: leaving everyone better off and no one worse off than before.

The same basic dynamic works with payment systems, because the merchant’s demand for customers is less price-sensitive than the customer’s demand for merchants. The debit interchange system thus works the needed transfer payments to bring the market into equilibrium, only now on a far greater scale. In this setting, it would be a transactional nightmare for each retailer to negotiate a separate interchange fee with each individual cardholder; the transactions costs would consume the entire gains from the system. So now the debit card platform takes the place of the proprietor of the singles bar, the newspaper owner, or the shopping mall manager. Issuing banks receive payments from merchants to help support their efforts to lure consumers. The platform operator sets the rates that link the two sides together, subject to the constraint that the total benefits from the system must exceed the total costs. Baxter was the first to articulate how the transfer payment from merchants to consumers could equalize demand on both sides of stable, two-sided markets.

Payment cards | Unlike bars, however, the payment system is more complex, for reasons that Baxter did not discuss in his 1983 article.

First, each side still has to organize its own efforts. Thus, on the merchant side, merchants must develop their own processing systems. They may do this with the assistance of what are called acquiring banks, which are paid negotiated fees that do not operate through the platform. On the other side of the market, the issuing banks have to figure out how best to price their services to their own customers. In general, the dominant pattern follows a TCF innovation made when debit cards were first introduced in 1995. The bank issues debit cards for free and then charges customers for defaults to help defray the cost of the system. This payment pattern rewards good behavior and imposes sanctions on defaulting customers, thus reducing the number of system breakdowns. It follows, therefore, that the transfer payment does not fund the full cost that the issuing bank incurs in operating the system.

The steep limitations on debit fees for all issuing banks could easily lead to a smaller bundle of services that issuing banks supply to merchants.

Second, the issuing banks assume more functions than they did in the older system. In the early checking systems, the merchant bore the risk of loss from bad checks because the customer’s bank had little or no information about the account balance when the check was drawn. For large transactions, cashier’s checks eliminated that risk with a built-in prepayment mechanism. But for routine transactions, those costs were so high that the merchant ate the loss. The debit card changed the risk-bearing capabilities of the two sides. No merchant can assemble reliable electronic data on thousands of customers who do business at different banks around the country. Each bank, however, has that information about its own customers. Quick communications across the network platform let the issuing bank harness that information to monitor both credit and debit transactions. With debit cards, the banks take the credit risk by authorizing payment, say, on an overdrawn account, into which a paycheck or Social Security monthly installment is scheduled to be deposited. The charge for taking that risk away from the merchant is then bundled into the interchange fee.

Third, the asymmetry in positions becomes more dramatic because the promotional activities by banks for debit cards lure more people into retail outlets. When banks advertise various promotions to their customers, these rewards are not tied to their purchases at any specific firm. They are made to build up the overall system in ways that let each retailer gain in rough proportion to its contributions. Retailers have no way to expand the base of general card users by their own efforts, because to promote the card lets other retailers free ride off its efforts. In effect, this payment system overcomes a collective action problem that would otherwise stymie any effort by individual merchants to improve the network performance separately.

These new functions depend on new technology. Checks and debit cards differ in the same way that faxes differ from emails and typewriters from computers. The first element in each of these three pairs is mute. The second is dynamic and allows for the constructive use of the acquired data. The combined effect of the multiple business functions of the expanded debit card transforms and enriches the original Baxter analysis. The Durbin Amendment, however, blocks any collective merchant payments for services across the debit card platform, thereby introducing a serious regulatory inefficiency for debit cards. The steep limitations on debit fees for all issuing banks could easily lead to a smaller bundle of services that issuing banks supply to merchants, who in turn might, for example, be asked to assume the debit card losses from forgery, fraud, and overdrawn accounts. That inefficient funding structure will also lead to cutbacks in system-wide advertisements, the growth of debit card use, and the introduction of new technologies. No legislative scheme with these retrograde features could be regarded as pro-consumer when measured against the achievements of any active and expanding system. The retailers have no reason to think that a combination of higher debit card fees and reduced services will outperform the status quo. Already newspaper and magazine articles are casting a jaundiced eye on “greedy” banks for raising fees — without mentioning the impact of the Durbin Amendment.

The Constitutional Challenges

The massive financial dislocations threatened by the Durbin Amendment have, to date, provoked only one constitutional challenge — that of my client, TCF. Its willingness to take the lead with this challenge rests upon its distinctive market position. Under the Durbin Amendment, a “large” bank is one with about $10 billion in assets. Yet TCF is heavily committed to the retail banking sector, with its largely blue-collar, low- to middle-income customer base. TCF has no credit card business, nor does it have any of the diversified set of activities of the mega-banks with which it is grouped. Simply put, TCF is the 47th largest bank by asset count, but the 12th largest debit card issuer. At the same time, it faces intense competition from the exempt small banks. TCF cannot shift customers to credit cards and it cannot do without the debit card. It has no effective business strategy to mitigate the revenue losses from the Durbin Amendment, such as by altering the mix of fees it charges to its customer base.

Economic hardship, standing alone, does not substantiate a constitutional grievance. In the current legal environment, moreover, a generalized “substantive due process” claim is hard to mount in the teeth of a highly deferential “rational basis” test under which courts give only cursory scrutiny to the challenged legislation. Accordingly, any successful constitutional attack must come from another quarter — in this instance, the law of rate regulation. That body of law first developed in the late 19th century in response to the emergent natural monopolies in railroads, telecommunications, electricity, and gas. Transportation and telecommunications grids cannot operate as perfectly competitive markets so long as rival carriers must make unique interconnections with each other, giving rise to a holdout risk by carriers that refuse to accept traffic originating in other networks. In contrast, the natural monopolies for electricity and gas need not, over the relevant range of output, fear new entry, given that any discrete geographical market is best supplied by a single provider. To be sure, rate regulation may not be the best strategy for dealing with natural monopolies, for that expensive and error-prone system often shields the incumbent supplier from entry by technologically superior firms. It may just be best to let technological improvements erode the once–natural monopoly base.

Because a generalized “substantive due process” claim is hard to mount, a successful constitutional attack must come from another quarter.

However strong that critique, no one thinks that it renders standard forms of rate regulation unconstitutional. It gives some hint, though, of the serious risks of undertaking this venture. Set the rates too high, and the public utility earns a monopoly profit. Set the rates too low, and government can confiscate the original capital investment that is already in the ground, denying both cost recovery and a reasonable profit over the useful life of the investment. To protect against the expropriation of sunk costs, courts have allowed the regulated parties to challenge insufficient rates on constitutional grounds, usually under the Due Process Clauses of the Fifth and the Fourteenth Amendments.

How, then, should courts set these rates? In general, courts have devised two standard methods. The earlier method requires the firm to bear the risk that its capital investments will not prove useful in the business; wasteful expenditures are thus excised from the rate base. After all, competitive firms do not receive any return on foolish investments either. The quid pro quo for the smaller rate base is a higher rate of return, just as in a competitive market. The alternative methodology does not ask whether the initial investments were prudently made; instead, all the invested capital is added into the rate base. In exchange, the utility receives a lower rate of return. The added simplicity of this system is, however, necessarily offset by the inferior incentives it creates for prudent investment. Which approach is best is far from clear, so that the current law, as reflected in the 1989 opinion of the late Chief Justice William Rehnquist in Duquesne Light Co. v. Barasch, leaves that choice to the regulator — but only if the “end result” or the “bottom line” lets the firm earn a sufficient return to attract and maintain capital. This test was propounded by Justice William O. Douglas’s 1944 decision in Federal Power Commission v. Hope Natural Gas Co. This bottom-line requirement set up a fixed and inflexible standard.

The key question is how this approach to rate regulation carries over to the payments industry, with its distinctive two-sided market structure. On the consumer side, the payment system is a purely competitive industry in which legions of issuing banks seek to attract debit card customers. On its other side, the issuing banks take as given the interchange rates set by Visa and MasterCard. The common, but overblown, fear is that these two companies will exert some monopoly power akin to that possessed by the traditional public utility. To be sure, the traditional antitrust law makes illegal any collusive arrangement between Visa and MasterCard to set rates, of which there is now no evidence. Quite the opposite, these two companies compete with each other vigorously in every market in which they operate, and they face constant inroads from American Express and Discover, which work on different business models, not to mention a large number of small companies that are poised in the wings. There is nothing remotely comparable to the territorial monopolies for gas and electric companies. These risks of supracompetitive prices are, in my judgment, small relative to the enormous transactional simplifications that debit card networks create. The alternative antitrust claim posits a vertical conspiracy between either Visa or MasterCard and the issuing banks. But this argument quickly founders in the absence of a clear explanation of how their cooperation could restrict output or raise prices. Why, if Visa and MasterCard had any such power, would they use it to benefit issuing banks instead of themselves?

A second feature separates debit card systems from traditional public utilities. The public utilities have only one source of revenue: their customer base. Banks that issue debit cards have two: the current interchange fees collected from merchants and fees from debit card customers. The Durbin Amendment does not regulate the fees that issuing banks can charge their own customers. The government and the merchants insist that those fees allow issuing banks to recoup, dollar for dollar, any lost interchange revenues, so that no economic loss stems from the regulation. Thus, the Fed’s rulemaking proposal of December 16, 2010 stated: “[U]nlike in the case of public utilities where the utility’s only source of revenue is the fees charged for the service or commodity, issuers have other sources, besides interchange fees, from which they can receive revenue to cover their costs of operations and earn a profit.” But this argument does not work. The only possible sources of revenue are fees to debit card holders, but at no point does the Fed ask whether market conditions, including the small bank exemption, allow them to earn the reasonable rate of return that this pronouncement implicitly concedes must be provided.

Indeed, given the major inefficiencies in the regulation, the Fed’s statement is unlikely to be true under ideal circumstances, even if the debit interchange regulation hits all banks equally, which of course it does not because of the $10 billion small-bank exemption. Just recently, Visa announced that it will introduce a two-tier interchange system to implement the Durbin Amendment; MasterCard will likely follow suit. In this skewed market, banks like TCF cannot raise fees in the face of direct competition from small banks that can continue to offer free debit cards. Consumer demand in this market is elastic, and customers will flee TCF in droves if it raises its rates. TCF must fight the Durbin Amendment because it receives no new revenue source from the passage of the legislation, for its preexisting right to charge customers is worth nothing under the current tilted playing field. What matters for its constitutional claim is not its abstract right to raise fees, as the Fed claims. Instead, what matters is the value of that right, which is in fact worth very little.

The Fed has written that the public utility cases do not apply to the Durbin Amendment because “[i]ssuers are unlike public utilities, which, in general, are required to make their services regularly available to the public.” In its view, the power to restrict rates is unlimited because competitive firms always have the option of leaving the market if they do not like the regulations in question. For TCF, these words are a death sentence, for there is quite simply no “checking” account without the debit card. The bank that does not offer debit service along with cash and checks must close its doors. Indeed, by the Fed’s logic, there is no reason to stop with a slash in present revenues. The government could tell any firm to give away all its goods for free so long as it has the option to exit the business, causing it to lose any return on the business infrastructure that it created.

The key reason for the constitutional investments is to encourage investment in long-term assets today by eliminating the fear of expropriation tomorrow, and that rationale applies with equal force to both competitive firms and monopoly industries. In fact, the exit rights issue receives appropriate protection for public utilities with monopoly power. In practice, the rate system has to meet two standards. First, under the so-called Brooks-Scanlon doctrine, it cannot require the regulated industry to subsidize its regulated businesses with revenues that it derives from non-regulated businesses. To let it do so is to confiscate the revenues of a second business to prop up the first. For banks, this proposition means that the government cannot cut revenues on the debit card business for any bank that has made profits on its lending business. Second, the government can only mandate cross subsidies among various groups of customers of the regulated business so long as the regulated party keeps its overall rate of return on the entire portfolio. It is therefore wholly unacceptable to allow the regulated entity to withdraw selectively from losing lines of business while keeping the profitable ones alive, so long as the regulator’s prescribed overall rate of return meets competitive levels. But as a traditional common law matter, public utilities could withdraw from the entire market so long as they give reasonable notice of their intention to allow some other party to pick up the slack.

It is now possible to put together all the pieces of the puzzle. The salient points are these: First, the constitutional protections against confiscatory rate regulation, originally developed in connection with natural monopolies, apply with equal or greater force to firms that operate in competitive markets. They cannot be given the Hobson’s choice of leaving a market in which they have made extensive capital investments or entering into a huge number of transactions all at a loss. Staying out of bankruptcy does not confer a competitive rate of return. Second, any possible justification for rate regulation is far weaker for debit cards than for electricity or gas because no bank enjoys any monopoly power. The current system of debit interchange is efficient in the sense that no changes in the current institutional arrangements by contract or regulation can make matters better off from a social point of view. There is close to perfect competition on the customer side of the market, leaving no choice but to accept take-it-or-leave-it offers from platform operators like Visa and MasterCard, who have no incentive to dispense freebies to the issuing banks with whom they deal at arm’s length. The power of these assumptions is made clear by the impact that the Durbin Amendment will have on TCF, whose rate of return will drop (in the current bad market) from 10 to 5 percent, by cutting out close to 40 percent of the profit on the overall business and 90 percent of the profit on the debit card business if the rates are limited to actual ACS costs under the Fed’s proposed regulations.

Conclusion

The Durbin Amendment’s regulation of debit interchange represents a radical effort to extend price regulation to areas in which it has never been attempted before. That effort was undertaken in response to strong factional industry pressures, but without any serious examination of how payment systems operate.

Those errors are evident at every stage in the argument. At a descriptive level, the amendment rests on the false assumption that debit cards do not offer advantages of convenience and certainty of payment that are not now found in checks. On an economic level, it rests on an underestimation of the major efficiencies associated with the use of debit interchange in two-sided markets. Finally, on the constitutional level, it imposes confiscatory regulation by the deep and dramatic cuts that it makes into the rate of return of the firms that have invested in debit card technology on the assumption that they could have a fair opportunity to recoup their original investment over time.

The entire episode has yet to play itself out. However, based on what has happened thus far, it shows the true dangers of using major legislative initiatives to interfere with settled practices that are far more equitable and efficient than the legal regime that Congress wishes to impose by brute force.