Dear Ms. Misback:
I appreciate the opportunity to comment on the Federal Reserve Board"s Request for Comment on Potential Federal Reserve Actions to Support Interbank Settlement of Faster Payments. I am the director of the Cato Institute"s Center for Monetary and Financial Alternatives. The Cato Institute is a public policy, non-profit, non-partisan research organization dedicated to the principles of individual liberty, limited government, free markets, and peace. I write to address the specific question, "Should the Reserve Banks develop a 24x7x365 RTGS [real-time gross settlement] settlement service?," concerning which the Board seeks outside input.1
As the Board"s request for public input itself makes clear, the Federal Reserve can support the development of faster retail payments arrangements in numerous ways. The Reserve Banks" direct provision of faster payments services to banks, through their development and implementation of new settlement facilities designed especially for that purpose, is one such alternative. I strongly believe, however, that such an approach, which would place the Federal Reserve in direct competition with both existing and potential future private-sector providers of faster payments services, is more likely to hamper than to expedite the development, rapid uptake, and continued improvement of an efficient faster payments system in this country.
1. Established Rules for Limiting Direct Fed Provision of Payments Services are Well-Founded
The Board of Governors has long been committed to the principle that the Reserve Banks should refrain from directly providing payments services, other than those essential to fulfilling its mandate, unless in doing so they satisfy three strict requirements. Those requirements are: (1) that, in accordance with the 1980 Depository Institutions Deregulation and Monetary Reform Act (DIDMCA), the fees it charges for such services are sufficient to recover their full (including imputed) cost; (2) that the service in question is one from which the public will clearly benefit; and (3) that the service be "one that other providers alone cannot be expected to provide with reasonable effectiveness, scope, and equity." A 2017 U.S. Treasury White Paper also endorses these criteria.2
The established requirements for the Fed"s direct provision of new payment services are well-founded. They are designed to prevent the Fed from employing resources inefficiently, particularly be preventing it from competing unfairly with private-market payments service providers. They also serve, as Federal Reserve advisors Edward Green and Richard Todd observed some years ago, to maintain "an overall relationship of mutual deference between the Fed and the private sector" that contributes to "the Fed"s reputation as a trustworthy and neutral organization focused on broad public objectives."3 That relationship, Green and Todd go on to observe, is threatened "if banks, other commercial firms or the general public perceive the Reserve Banks to be encroaching on activities that the private sector can perform efficiently and equitably."
Conditions may indeed exist in which unregulated market providers, left entirely to their own devices, are unable to supply particular payments services efficiently — that is, in which some sort of market failure warrants some form of government involvement. But even when some of these conditions exist, the feared inefficiencies may be best addressed through interventions other than direct provision of new payments services by the Federal Reserve. With regard to faster retail payments, convincing grounds for assuming that the private market has failed are lacking; and to the extent that market failure remains a potential problem, direct Fed competition, far from being an efficient solution to the problem, is likely to prove counterproductive. I believe that the Fed can more effectively assist in addressing potential market failures by other means.
2. No Prima-Facie Case Exists for Assuming Market Failure in Fast Payments
Although there are a number of reasons for supposing that private market failures pose a serious barrier to the development of an efficient U.S. fast retail payments system, not all of these reasons are compelling. For example, the fact that fast and "ubiquitous" (that is, universally accessible) faster payments systems already exist in several other countries is sometimes taken to mean that U.S. providers have been slow on their feet. But the fact that the U.S. has for some time had an especially well-established and widely-used bank- and credit-card based retail payments system has also made the demand for alternative payments arrangements less urgent here than elsewhere.
The fact that faster payments services exhibit "network effects," so that their value to individual users rises as more users take part, also does not mean that such services cannot be provided, and provided efficiently, by the private sector. As an article in the FRB Richmond Economic Quarterly noted two decades ago, so long as a private payments network gains more by adding a new participant than it costs to add that participant, that network "will have an incentive to compensate the added participant…even in the absence of government intervention."4
It is also possible that the private sector cannot provide efficient fast payments services because declining average costs make the service a natural monopoly. But even where a particular fast payments service has the properties of a natural monopoly, with substantial fixed investment costs, a private firm"s provision of that service can still be efficient provided that the firm faces either actual or potential competition from private-market suppliers of alternative services. A private fast payment system provider in the U.S. today faces competition both from suppliers of "legacy" retail payments services, including well-established credit and debit card networks; from closed payments networks like PayPal; from non-bank mobile payments services such as Google Pay, Apple Pay, and Venmo; and from blockchain-based payments services. As Green and Todd observe, "even when only a single firm is actively providing a service," the presence of such alternatives can suffice to prevent that firm "from setting prices significantly above competitive norms," from "skimp[ing] on the quality and reliability of their services," or from "discriminat[ing] among customers to a greater extent than is required for economic efficiency."5
3. An Already-Established Private Arrangement Appears Capable of Meeting the Faster Payments Task Forces" Goals
But by far the most compelling evidence against the presumption of market-failure consists of the fact that the private sector has actually established a fast retail payments system capable of meeting the same ends the Fed itself might hope to meet by setting up its own RTGS system. That system is the RTP ("Real-Time Payments") network, established by TCH (The Clearing House Payments Company) in 2014, the year before the Fed established its Faster Payments Task Force, in which TCH also took part.
The RTP system relies on a Fed Master Account that is jointly-owned by the participating banks, of which TCH is the sole custodian. The participants must prefund their portions of the pooled account, in amounts established by TCH rules. Payments made on the RTP network are cleared and settled almost instantaneously on the RTP account ledger, with accounts replenished (or excess credits transferred to individual bank accounts) through Fedwire. The system is free of any settlement risk, operates 24x7, and is open to all banks, with a fixed cost of 4.5 cents to payment senders — that is lower than the 5.2 cent cost of same-day (but far from near-instantaneous) ACH payments.6
At present the RTP network consists of 11 banks, accounting for nearly 50 percent of total U.S. bank deposits; and although many smaller depository institutions have yet to join, its flat-fee arrangement is specifically designed to attract such banks by allowing them to effect payments on the same terms as their larger rivals. The fee arrangement also appears to meet the Task Forces" equity objective. Although the RTP network is as yet far from "ubiquitous," it is just as capable as a Fed-provided RTGS system would be of becoming so, and perhaps more capable, other things equal, owing to its first-mover advantage.
4. There is Risk of Unfair Fed Competition
Although it is generally desirable that other firms should be free to offer payments services that are close substitutes for, and compete directly with, RTP"s services, it does not follow that allowing the Federal Reserve itself to do so is in the public interest. The reason for this is that, while the Fed might potentially be an efficient provider of such services, there is also some risk that its unique privileges will allow it to successfully compete with RTP and other private-market fast payments service suppliers despite being less efficient than they are.
The history of the Fed"s involvement in the clearing of paper checks illustrates this danger. Prior to the passage of the DIDMCA, the Reserve Banks made use of cross-subsidies — together with their unique ability to compel member banks to settle checks it mailed to them immediately and at par — to compete successfully with efficient private-sector clearing systems. The Fed thus became the dominant provider of a payments service that had once been provided entirely by private-sector providers, and that is still privately provided in many parts of the world.7 The Fed"s status as this country"s only provider of nationwide check-clearing services in turn allowed it to compete successfully with private providers of other payments service markets, including ACH payments, that the private sector alone might otherwise have served efficiently.8
While the passage of DIDMCA limited the Fed"s ability to cross-subsidize its check-clearing services, and a 1994 reform allowed private check presenters to share in the privilege of same-day par settlement provided their checks were presented by 8 AM, the Fed still enjoys the advantage of being able to demand same day payment for any checks it present before 2 PM. It is mainly owing to this "six-hour monopoly" that the Fed was able to continue to dominate the check-clearing industry long after the DIDMCA took effect.9
The history of the Fed"s involvement in check clearing should serve as a cautionary tale to those who suppose that the DIDMCA"s provisions can be relied upon to prevent the Fed from competing successfully against more efficient private fast payments providers. There are, furthermore, reasons for doubting that the Fed always complies adequately with the DIDMCA"s cost-recovery provisions. The difficulty of achieving full compliance is especially acute with regard to "imputed" costs the Fed is supposed to take into account, including taxes that it would have to pay were it also a profit-making private-sector provider. Among other problems, the Fed is not required to take account of many of the regulatory compliance costs that private-sector payments service providers incur. That the Fed"s internal cost-accounting system has not been reviewed by an external auditor in several decades supplies that much more reason for fearing that it might offer fast payments services for less than their true cost.10
The possibility that the Fed might compete unfairly with private sector service providers is bound to discourage private investment in fast payment service infrastructure. Yet, as Jeffrey Lacker and John Weinberg have observed, the same privileges that could allow the Reserve Banks to engage in such unfair competition may also "inhibit their [own] pursuit of improvements in products and processes." 11 Pointing to the example of the U.S. Postal Service, Lacker and Weinberg note furthermore that it can be especially hard "for large organizations, particularly public institutions, to respond nimbly to new technological opportunities." 12
5.Direct Fed Provision of Faster Payments Services Could Well Prove Counterproductive
Owing to the presence of network externalities, the efficiency of any payments network necessarily depends on its success in attracting members. Furthermore, the Faster Payments Task Force"s "ubiquity" goal treats nearly universal network participation as a fast payments network desideratum. Where rival networks exist and cannot easily be made fully inter-operational, anything less than full participation in at least one network will mean fewer benefits than could be realized by a single universal network, even setting aside the wasteful replication of sunk costs that multiple networks entail. It is, nevertheless, essential that payments service markets be contestable, so that established networks are not capable of prevailing despite ceasing to operate efficiently. Sound public policy must therefore seek to ensure the contestability of markets for fast payments services, even when doing so may invite occasional, disruptive innovations.
As I"ve noted, the Fed differs from private payments service providers in being capable of successfully entering a market for payments services, not by offering a superior product, but by taking advantage of its unique powers to underprice its services. No less importantly, unlike private sector providers the Fed does not have to convince shareholders that it can recover the costs involved in any new payments venture it undertakes. Instead, it only faces an ex-post obligation to do so, which it may or may not meet.
Thanks to its unique powers, the possibility that the Fed might compete head on with the private RTP network is likely to have an exceptionally chilling effect on RTP"s ability to attract new members, and to do so even despite RTPs first-mover advantage, and also despite any real efficiency advantages it might enjoy. Banks faced with substantial network-interface investment costs will hesitate to join RTP until they are certain of the Fed"s intention. Even if the Fed does not ultimately enter the market, this hesitation will itself be costly, because it will delay the achievement of a ubiquitous system. The likelihood that the Fed will take several years to establish its own RTGS system will compound this delay. The case of rival private-sector entrants differs, both because such entrants are only likely to contemplate entering the market if their stakeholders believe them to be capable of operating more efficiently than established rivals, or of offering services that are clearly superior to theirs, and because they must in fact offer superior or less expensive services to gain market share.
Should the Fed actually choose to compete with RTP (and, therefore, with other actual and potential private suppliers of faster retail payments services), the potential welfare costs may be far greater. Those added costs will include the costs of building the Fed"s alternative fast payments infrastructure — which will presumably detract from Fed remittances to the Treasury — as well as extra network connection costs borne by banks that choose either to join both networks or to switch from the RTP network to the Fed network. Should all present RTP member banks elect either to join both networks or to abandon RTP, these extra costs will be considerable. Yet alternative possibilities would be still less desirable, for in those cases neither network will be truly ubiquitous. Consequently, despite any expenses incurred, the result will fail to achieve one of the Fast Payments Task Force"s principal objectives.
Equitable payments fees are, of course, another crucial Fast Payments Task Force objective. But it is difficult to imagine a scheme more equitable than the RTPs fixed sender fee, which favors smaller banks. That this fee is less than the present fee for slower ACH payments also makes it highly unlikely that the Fed could offer its rival service for less while still abiding by the DIDMCA cost-recovery requirements. The result — assuming that the DIDMCA rules are strictly adhered to — could be capital expenditure on the Fed"s part that it is not able to recoup, owing to insufficient network subscription, so that it must abandon its retail payments effort, while forcing the Treasury to bear its sunk investment costs.13
6. The Fed Needn"t Compete with Private-Market Suppliers to Guard Against Market Failure
While there are serious drawbacks to having the Fed compete directly with private-sector providers of fast payments services, it is also possible that, in the absence of direct Fed competition, RTP or some other dominant private-sector provider would take advantage of its position to engage in inefficient or inequitable provision of its services — or to otherwise fail to achieve the goals set-forth by the Fast Payments Task Force. A crucial question, therefore, is whether the Fed can effectively guard against such abuse without establishing its own fast payments network.
I believe that it can. As I noted previously, avoiding an established provider"s abuse of market dominance requires, not that there be multiple firms actively providing identical or very similar payments services, but that the fast payments service market be contestable, where the relevant contests consist not only of that between or among potential suppliers of very similar services but also that among alternative suppliers of services that, while not perfect substitutes, are each nevertheless potential substitutes to some considerable degree. Provided there are enough such potential substitutes, market-discipline can still be relied upon to check any one private-sector provider"s abuse of its dominant position.
There are several ways in which the Fed might enhance the contestability of the market for fast payments services short of entering that market itself. It can do so by enhancing its current ("legacy") settlement services to make them more suitable to the needs of private-market fast payments service providers, particularly by reducing the interbank credit risk to which such providers expose themselves when they offer faster payments services, and especially round-the-clock services, while relying on less flexible interbank settlement arrangements. The Fed might do this either by further extending, ideally to 24x7x365, the operating hours of its Fedwire or National Settlement Services, or by otherwise offering private-sector RTGS providers a full-time "liquidity management" service that will allow them to move funds outside ordinary business hours. The Fed can and should also continue to cooperate with NACHA (the National Automated Clearing House Association), as it has been doing, in further enhancing the timing and frequency of its ACH settlements, perhaps by adding still more payment submission "windows."
The Fed can also enhance the contestability of the market for faster payments services by easing entry into legacy payments systems. For example, it might allow certain non-bank providers of "front end" retail payments services to open master accounts with it. Alternatively, it could make it easier for chartered banks to serve as correspondents for, or otherwise form relationships with, non-bank payments service providers, by reducing regulatory burdens stemming from such relationships, and by making it easier than it is at present for would-be bankers affiliated with non-bank service providers to acquire either state of national charters.14
I thank the Board of Governors for the opportunity to share my opinions concerning the means by which the Federal Reserve System should seek to encourage the development of an equitable and ubiquitous faster payments system in the United States.
1 Federal Reserve System, Request for Comment, "Potential Federal Reserve Actions to Support Interbank Settlement of Faster Payments," Federal Register 83, no. 221 (November 15, 2018): 57351, available at https://www.govinfo.gov/content/pkg/FR-2018-11-15/pdf/2018-24667.pdf, p. 57354.
2 Steven T. Mnuchin and Craig S. Phillips, "A Financial System That Creates Economic Opportunities: Nonbank Financials, Fintech, and Innovation," U.S. Department of the Treasury, October 2017, p. 156.
3 Edward J. Green and Richard M. Todd, "Thoughts on the Fed"s Role in the Payments System," Federal Reserve Bank of Minneapolis Annual Reports, May 1, 2001.
4 Jeffrey M. Lacker and John A. Weinberg, "Can the Fed be a Payment System Innovator?," Federal Reserve Bank of Richmond Economic Quarterly, Spring 1998, p. 8.
5 Op. cit., p. 12.
6 For further details see "Real-Time Payments Operating Rules," The Clearing House, October 30, 2017, available at https://www.theclearinghouse.org/payment-systems/-/media/6de51d50713841539e7b38b91fe262d1.ashx.
7 See Jeffrey M. Lacker, Jeffrey D. Walker, John A. Weinberg, "The Fed"s Entry into Check Clearing Reconsidered," Federal Reserve Bank of Richmond Economic Quarterly, Spring 1999, pp. 1-31.
8 Richard Oliver and Stuart E. Weiner, "The Role of Central Banks in Retail Payments: The Central Bank as Operator." Proceedings of the Federal Reserve Bank of Kansas City Research Conference, The changing retail payments landscape: what role for central banks? November 9-10, 2009, p. 208.
9 Lacker and Weinberg, pp. 17-18.
10 For further details see GAO Report, "Payment Services: Federal Reserve"s Competition with Other Providers Benefits Customers, but Additional Reviews Could Increase Assurance of Cost Accuracy," August 30, 2016.
11 Op. cit., p. 22.
13 The argument here resembles that for not having the Fed enter the credit card business. As Kansas City Fed economists Richard Oliver and Stuart E. Weiner point out, although some have argued that by competing with established, private card networks the Fed might help foster "reasonably equal access, efficiency, and integrity" in the credit-card payments market, the fact "that current providers have invested millions in the existing infrastructure and relationships" constitutes a powerful counterargument, because the Fed "would have to raise and invest capital sufficient to provide promising scope and scale economies to be successful over the long run" — which it would be hard-pressed to do whilst still abiding by the DIDMCA requirements. "Without substantial government subsidies," Oliver and Weiner conclude, "it appears unlikely that the Fed could easily or efficiently enter the card market at a scale that would invite long-term success." See Oliver and Weiner, op. cit.
14 See Barak J. Sanford and Daniel Bufithis-Hurie, "Should Non-bank Payment Firms Be Eligible to Open Federal Reserve Bank Accounts," Banking Perspectives, November 25, 2018.