Chairman Crapo, Ranking Member Brown, and distinguished members of the Committee on Banking, Housing, and Urban Affairs, my name is George Selgin, and I am the Director of the Cato Institute’s Center for Monetary and Financial Alternatives. I am also an adjunct professor of economics at George Mason University, and Professor Emeritus of Economics at the University of Georgia.
To assist Congress in that endeavor, I wish to draw your Committee's attention to some dangers posed by the Federal Reserve decision to proceed with FedNow — a real-time retail payments service that will compete directly with private-sector retail payments services. Specifically, I wish to discuss four ways in which the Fed's plan might hinder rather than facilitate the achievement of an equitable, efficient, and safe U.S. fast payments system, and to suggest steps Congress should take to guard against this outcome.
The Federal Reserve as a Payment Service Competitor
As a rule, competition is an effective — if not the most effective — means for encouraging providers of services to price those services equitably, to produce them efficiently, and to improve their quality over time. However, these outcomes depend on the presence of a level playing field on which all providers compete — that is, they depend on the various providers having roughly equal legal privileges and obligations. In the absence of a level playing field, the presence of multiple providers alone does not guarantee good outcomes. Instead, special care must be taken to guard against bad ones.
The Federal Reserve banks enjoy many legal advantages over private suppliers of payment services. They command a monopoly of bank reserves that serve as means of final payment; they are empowered to regulate commercial banks and some other private-sector payment service providers; and they are exempt from antitrust laws. Finally, although the 1980 Monetary Control Act requires that the Fed charge prices for its services that recover those services' capital and operating expenses, it only needs to do so over a "long run" of unspecified length, and then only according to accounting methods of its own choosing that are not subject to external review.
These and other Fed privileges mean that, when it enters into direct competition with private-sector payment service providers, it does so on a playing field that it can easily slant in its favor. It is owing to this that the Fed itself has established strict criteria it must meet before offering any new payment service, including the requirement that the service in question "be one that other providers alone cannot be expected to provide with reasonable effectiveness, scope, and equity."2
In responding to the Fed's request for comment regarding "Potential Federal Reserve Actions to Support Interbank Settlement of Faster Payments," I argued against the Fed's then-proposed retail RTGS ("Real Time Gross Settlement") payment service partly on the grounds that it did not meet the Fed's own criteria for providing new payments services.3 I also argued that the new service would delay progress toward a ubiquitous U.S. fast payments system. I continue to hold these views.
I also fear that, instead of preventing private-sector payment service providers from engaging in anticompetitive behavior, the Fed will itself engage in such behavior. In my testimony today, I wish to draw attention to four particular anticompetitive dangers that the Fed's entry into the fast payments business poses, and to recommend steps Congress should take to guard against each.
Postponed Fed Settlement System Reform
The first danger is that the Fed will treat FedNow as a substitute for a 24x7x365 expansion of the operating hours of Fedwire, its wholesale RTGS service, and NSS (the National Settlement Service), a separate multilateral settlement service that is also owned and operated by the Federal Reserve banks.4 The availability of 24x7x365 Fed settlements is essential to achieving faster (though not necessarily real-time) and safer payments on other payment services. But instead of hastening to offer that service, the Fed may delay doing so to limit private payment services' ability to compete with it.
The danger here stems from the Fed's monopoly of final means of payment, including bank reserves. Because of that monopoly, most private noncash payments, including most check, card, and ACH ("Automated Clearing House") payments, can only be completed with the help of either Fedwire or the NSS or both. Only once settlement takes place can recipient banks credit funds to a payee's account without assuming some credit risk. Because Fedwire and the NSS operate only on weekdays, excluding holidays, and then with limited hours, retail payment services that rely on them are correspondingly limited in their ability to complete payments both quickly and safely at all times.5
Although it would also enhance the efficiency of private real-time payments services, the main benefit of 24x7x365 Fed settlement services would consist of a substantial reduction in delays on "legacy" payment networks.6 For example, today's Fedwire and NSS operating hours currently stand in the way of National Automated Clearing House Association's (NACHA) long-standing effort to enhance ACH's same day payment services by providing for a third ACH "processing window." Although NACHA had hoped to make this third window available by September 2020, and the change required only a minor extension of Fedwire and NSS operating hours, the Fed failed to prepare for the change on time, forcing NACHA to postpone its planned reform until March 2021.7
When the Fed requested public comment on whether it should establish its own fast payments network, it also asked whether it should either arrange to have Fedwire and the NSS operate 24x7x365 or establish a new "Liquidity Management Tool" for the purpose of allowing 24x7x365 transfers among commercial banks' Federal Reserve accounts. Almost every response to this question favored having the Fed pursue one of these proposed reforms (most respondents did not care which), making the proposal much less controversial than the Fed's plan to establish its own retail RTGS service. Yet despite this, and the relative easiness and great potential benefits of the asked-for reform, the Fed ultimately chose to do no more than continue to "explore" the possibility of offering 24x7x365 settlement services, and to perhaps seek comment upon the proposal yet again!8
Why is the Fed dragging its feet on an almost universally favored reform that could alone suffice to eliminate most of the more notorious payment delays in this country?9 The Fed's actions seem at odds with its overarching public mission. But they are what one would expect from a firm endeavoring to compete successfully with rival payment service providers. For example, when NACHA was first endeavoring to make same-day ACH payments possible, its efforts were opposed by several large banks. It was widely suspected, according to a contemporary report, that this opposition stemmed from those banks' intent "to build their own proprietary electronic payment systems, which could give them a leg up on smaller banks."10 The Fed's hesitation to make 24x7x365 Fed settlements available to private payment service providers may likewise reflect its own desire to give FedNow "a leg up" on other payment networks.11
Whatever the Fed's motives, Congress should not allow it to delay a badly-needed enhancement of its settlement services any longer. Instead, it should give the Fed two years within which to either place its Fedwire and NSS services on a 24x7x365 operating basis, or establish an alternative 24x7x365 Liquidity Management Tool. If Congress does not do this, I fear that Congress will overlook the most important of all steps it might take to dramatically and rapidly enhance the speed of U.S. retail payments.
Volume-Based Pricing Favoring Large Banks
A second danger the Fed's entry into the fast payments business poses is that, by resorting to volume-based pricing, the Fed will ultimately put small banks that wish to offer fast payment services to their customers at a disadvantage.
Because many are counting on the Fed to guard against rather than introduce volume-based fast payment fees, some background is required to understand why that expectation exists, and why just the opposite might happen.
The only potentially ubiquitous real-time payments service that exists at present, the RTP system established by TCH (The Clearing House) in 2017, presently operates on a contractually-binding flat-rate basis, with no minimum volume requirements.12 But TCH's flat-fee commitment isn't absolute: instead, it allows RTP to alter its pricing policy in the event that the Fed enters into competition with it. Noting this, Fed officials and others have argued that RTP cannot be trusted to make certain that small banks continue to receive equitable treatment, instead of finding themselves placed at a disadvantage relative to their large competitors. That TCH is itself owned by 25 of the nation's largest banks makes the risk to smaller banks seem all the more obvious. Consequently, the Fed and others argue, having FedNow directly compete with RTP is the surest way to keep RTP from reneging on its flat-fee commitment.
But closer consideration of TCH's general pricing practices, along with some history, suggest that the Fed's entry is more likely to have just the opposite consequence. Regarding TCH's practices, in seeking a statement from the Justice Department's Antitrust Division "of its present intention not to seek any enforcement action against" the RTP system it was then developing, TCH explained that it
operates on a "utility" model, charging fees only to cover the costs incurred in operating its CHIPS, EPN, and check imaging systems and to support future innovation, and does not pay dividends to its owner banks.13 Accordingly ... TCH owner banks … will benefit by participating in the RTP system and enhancing their abilities to compete more effectively among themselves and with non-TCH owner banks and nonbank payment service providers.14
The veracity of TCH's claims is attested to both by the known pricing practices of its established payment systems and by the Justice Department's conclusion that RTP did not in fact pose "significant anticompetitive threats."15
FedNow, in contrast, does pose such a threat, as is clear from what happened in the case of ACH payments. The Fed competes with TCH, and in the past competed with other private-sector providers, in providing ACH payment services. TCH initially charged flat ACH fees. But during the 1990s, the Fed, in an effort to compete more aggressively in an increasingly national ACH market, resorted to volume-based ACH fees.16 The Fed's move compelled TCH to follow suit to avoid losing the business of its larger ACH customers. Yet TCH's ACH prices are still more favorable to small banks than those charged by the Fed, which charges many smaller banks five times the per-transaction fee it charges its largest customers.17
It was to protect itself from such potential Fed competition, and not (as Fed officials have suggested) to be able to ultimately resort to discriminatory pricing, that TCH made its flat-rate commitment contingent on the Fed's not entering into competition with it. Were TCH not to do this, it would risk having FedNow bid away its large participants.
To avoid having volume-based pricing undermine the goal of equitable real-time payments, Congress must do more than merely trust the Fed not to engage in such pricing. At very least, it should insist that the Federal Reserve Board follow TCH's example by making a public commitment to refrain from offering volume-based discounts on FedNow or, at very least, by publicizing a specific, anticipated FedNow pricing policy, such as it presumably employed in assessing the new service's feasibility and desirability. As then Richmond Fed economist John Weinberg observed some years ago, "When the Fed is one of several competitors, it can contribute to the efficiency of the market by adopting a clear pricing policy to which other sellers can react."18
Prejudicial Treatment of Balances in Jointly-Held Fed Accounts
The third danger stems from the Fed's ability to refuse to classify bank balances held in jointly-owned Fed accounts as reserves, and to do so even when the accounts in question are "intended to facilitate settlement between and among depository institutions participating in private-sector payment systems."19
Fed balances classified as "reserves" earn interest, while those not so classified do not. Consequently, by refusing to classify the jointly held Fed balances held by banks participating in a private payments network as reserves, the Fed adds to the cost of participating in that network, and hence to the relative attractiveness of other networks, including those it itself operates, that aren't subject to the same "reserve tax." The Fed's status as bank regulator can thus allow it to compete unfairly by "raising [its] rivals' costs."20
Although the Fed allows "only an institution eligible to have a Federal Reserve account under the applicable federal statute and Federal Reserve rules, policies, and procedures" to be a joint account holder, it reserves the right to determine whether balances in joint accounts count as reserves on a balance-by-balance basis.21 Today, the Fed administers three joint accounts serving to facilitate settlements among participants in TCH's CHIPs, RTP, and EPN networks.22 So far as I'm aware, it has not yet chosen to treat balances in any of these accounts as reserves. Consequently those balances neither bear interest nor qualify as "High Quality Liquid Assets" that can satisfy Basel's LCR ("Liquidity Coverage Ratio") requirements.
I can think of no economic reason why the Fed should not classify all Federal Reserve bank balances held in joint accounts used in settling payments as reserves, and to accord such balances the same privileges as other reserve balances. RTP account balances, for example, are no less liquid than banks' regular Fed account balances, and serve the same purpose of supplying their owners with means for settling payments. That banks choose to fund their RTP accounts rather than their individual Fed accounts, so as to allow them to make real-time payments instead of relying on slower ones, should not subject them to any avoidable penalties.
Moreover, by refusing to treat RTP balances as reserves the Fed may complicate its monetary policy operations unnecessarily by creating a new "autonomous" determinant of the total stock of bank reserves. As the Fed itself explains:
if joint account balances are not treated as reserves, they are a factor affecting the supply of reserve balances, meaning, all else equal, movements in joint account balances have similarly sized but opposite effects on the supply of reserve balances, which the Federal Reserve will need to offset to provide the appropriate level of reserves in a scarce reserve regime.23
In short, the Fed's ability to refuse to classify balances held in joint accounts "intended to facilitate settlement" on private payments system with which it competes represents a clear conflict of interests. To resolve this conflict, and thereby assure that the Fed competes fairly with rival payment service providers, Congress should compel the Fed to classify all balances held in joint Federal Reserve bank accounts as reserves, provided only that the accounts in question are designed to facilitate settlements on private payments networks. Congress should also have the Government Accountability Office (GAO) occasionally review the Fed's handling of applications for such joint accounts, to ensure that it continues to abide by its current guidelines for granting them.
Abuse of Monetary Control Act Loopholes
Finally, I wish to point to the risk that the Fed will take advantage of loopholes in the 1980 Monetary Control Act (MCA) to charge prices for its FedNow services that fail to cover their full costs, as that act requires. Thanks to its monopoly of paper currency, the Fed earns substantial "seigniorage" revenue it can use to cross-subsidize its other payment services to the extent that MCA loopholes allow it.
Although the MCA is supposed to rule out such cross-subsidies, there are at least two defects in its provisions that can prevent it from doing so. One concerns the Act's requirement that the Fed's service fees cover its costs "over the long run." Because it fails to define "the long run," the Act as written allows the Fed to interpret the phrase as it pleases. In contrast, private-sector payment service providers must generally be able to recover the cost of new services rapidly enough to achieve a positive present value for those services.
Fed officials claim that they generally endeavor to recover the Fed's expenditures for established services within a ten-year period, but that they expect FedNow's "first instance of long-run cost recovery to occur outside" that 10-year cost recovery period.24 However, they do not say how far outside, and the Fed incurs no penalties for failing to recover its costs within any specific length of time.25 It follows that the Fed's investment in FedNow needn't have a positive present value, so that it can set FedNow fees below what a private-sector provider of an equally costly service could afford.
A second MCA loophole leaves to the Fed itself the choice of an internal cost accounting system by which the Fed allocates its expenditures among its various activities, while failing to provide for periodic and systematic external reviews of that accounting system to assure its adequacy. In consequence the last external review of the Fed's cost accounting system took place in 1984! External assessments of the Fed's success in complying with the MCA's cost recovery provisions, such as that undertaken by the GAO in 2016,26 are therefore only as accurate as the Fed's own internal audits — a highly unsatisfactory circumstance.
By closing these MCA loopholes, Congress can prevent the Fed from underpricing its payments services, including FedNow. To do so, it should insist that the Fed offer compelling proof that it will be able to recover the costs of FedNow rapidly enough to give that project a positive present value using an equitable and competitive fee structure. Congress should also follow the GAO's 2016 recommendation that it provide for periodic independent reviews of the Fed's cost-accounting practices.27 Together these changes should go far in assuring that the Fed competes fairly with private payment service providers.
I conclude my testimony by observing that none of the steps I have recommended to Congress would prevent the Fed from doing all that it can possibly do to facilitate faster payments in the U.S. My recommendations will only serve to make sure that in competing with private-sector payment service providers, the Fed plays by the rules, as it must if it is to contribute to rather than hinder the speeding-up of U.S. payments. A well-intentioned Fed should therefore have no objection to them, while an ill-intentioned one will make them indispensable.