“The advantages of competition and deregulation must be sought in ways that do not impair the payments system or subject it to interruptions and breakdowns.” James Tobin, “Financial Innovation and Deregulation in Perspective,” BOJ Monetary and Economic Studies, September 1985.
How will financial innovation alter the role of central banks? As the structure of banking and finance changes, what will happen to the mechanisms and frameworks for setting monetary and financial policy? Over the past several decades, with the development of inflation targeting, central banks have delivered price stability. And, improved prudential policies are making the financial system more resilient. Will fintech—ranging from the use of electronic platforms to algorithm-driven transactions that supplant the traditional provision and implementation of financial services—change any of this?
This is a very broad topic, some of which we have written about in previous posts. For example, we are skeptical that peer-to-peer lending will replace significant portions of the credit system. And, while we have expressed a hopeful view of the blockchain technology—a reasonably efficient way to track ownership—it is hard to see private digital currencies like Bitcoin making significant inroads (aside from as a vehicle for the evasion of capital controls, money laundering rules, and the like). We have also expressed our view that paper currency is a key aspect how we organize our society, so we should think long and hard before we get rid of it. But, in the end, there is still much that we do not know and still hope to understand.
With that in mind, this post considers an innovation suggested by Barrdear and Kumhof at the Bank of England: that central banks should offer universal, unlimited access to deposit accounts. What would this “central bank digital currency” mean for the financial system? Does it make sense for central banks to compete with commercial banks in providing deposit accounts? We doubt it. It is not an accident that—at virtually every central bank—only commercial banks today have interest-bearing deposits. And there are reasons that central banks that once offered private accounts no longer do. Changing this would pose a risk of destabilizing the financial system.
To understand our assessment, we start by looking at financial intermediaries in general. Money and banking textbooks (like ours) list the roles played by banks: they pool savings to supply credit; provide safekeeping and accounting services (giving their customers access to the payments system and allowing them to track their income and expenditures); supply liquidity (so that depositors can transform their financial assets into money quickly, easily, and at low cost); diversify risk; and collect, process and use information to ensure that savings are transformed efficiently into investments.
A central bank that provides deposits will not supply this full set of services, but it will have to offer accounting, payments system access, liquidity, and the tracking of information. As a consequence, the central bank will need to have compliance and risk management functions—including systems that prevent money laundering, tax evasion and other illegal activities potentially aided by finance. Like banks, they must know their customer.
This brings us to the first conclusion: the cost of providing these services is significant, averaging between 2 and 3 percent of assets for U.S. banks (see Table 3 here). Even in the presence of scale economies, it is hard to see the Federal Reserve gaining much of an advantage over the Bank of America, which has deposit liabilities of $1.2 trillion (roughly 10 percent of non-currency M2). In other words, the central bank is unlikely to provide the services associated with deposit accounts more cheaply than the private sector.
That said, if the central bank displaces private deposits, the cost of private banking is likely to rise. Today, commercial banks are the only entities allowed to issue checking accounts (technically known as demand deposits). The profit they obtain from doing so is what economists call “rent” and business people refer to as “franchise value.” The provision of bank-like services by nonbanks (like money market mutual funds) already has eroded this rent. Central bank entry into the business could eliminate it.
The Barrdear-Kumhof proposal bears similarity to at least five others: Tobin’s deposit currency accounts; Garratt, Martin, McAndrews and Nosal’s segregated balance accounts; Greenwood, Hanson and Stein’s suggestion that the Fed offer large quantities of reverse repurchase agreements (RRPs); Cochrane’s advocacy of narrow banking; and his proposal that the U.S. Treasury issue variable rate perpetuities in single dollar amounts. A common thread in most of these is their aim to promote financial stability and reduce the risk of runs or limit their economic impact. (Tobin, for example, observed the entry of less-regulated nonbanks in the provision of nonbank services and wished to secure the means of payment.)
The first four proposals allow individuals to obtain deposits at the central bank through intermediaries, while the fifth cuts out the Fed as an intermediary between individuals and the U.S. government (the debt issuer). In each of them, a customer would place funds into an account with the understanding that the bank would deposit the full amount at the central bank. For the currency accounts, segregated balance accounts and narrow banks, this would occur directly. In the case of the RRPs, the transaction is slightly more complex, as the bank or mutual fund would promise to engage in an RRP with the Fed for the full value of the deposit. While these are analogous to the central bank offering universal access, there is the important distinction that a private intermediary is providing the typical banking services described earlier.
Because they are inherently safer than those offered by their private competitors, we would expect direct deposit accounts at the central bank will be in great demand. How big might the funds shifts be if the Federal Reserve today offered demandable deposits without limit? We can’t know precisely, but to give some sense, we note several things. First, there are roughly $2 trillion in demandable deposits in the U.S. banking system. The bulk of that could move quickly. Second, of the $12 trillion in non-currency M2, roughly $5 trillion is uninsured, as is another $2 trillion in money-market funds that are outside of M2. And finally, even some of the insured time deposits might move.
This suggests to us that the shifts will be very large at the start of a new regime, even in the absence of financial stress. But that will just be the beginning. Imagine what happens when there is even the slightest disturbance in the financial system. Suppose you are the chief financial officer (CFO) of a firm that still has large uninsured deposits at a bank? This is the cash that you are using to meet your payroll, pay your suppliers, and the like. How would you react if you were told that you could either keep your firm’s operating funds in the bank, with the small possibility that it will fail, or you can move them to the central bank? Even with some interest rate premium, would you really stay put? We wouldn’t! (In today’s financial system, a nervous CFO could switch to short-term Treasury bills, but a checking account at the Fed would offer even greater liquidity.)
This leads us to our main concern about the proposal: this mechanism would promote, rather than diminish, the risk of runs. During times of stress, if the central bank offers unlimited deposits universally, funds that have not already exited the commercial banking system will flee. Under such circumstances, the only thing that would save the financial system would be a broad extension of government guarantees to other private bank liabilities, or a heroic expansion of lending by the central bank. In other words, to prevent a crisis, the central bank would be forced to take over much of the private financial system, replacing commercial bank deposits with its own liabilities.
From this, we come to our last point. Elastically supplied central bank digital currency raises profound political economy issues. By creating accounts for a broad swath of individuals and firms, the central bank becomes a massive state bank. Would such a bank really restrict its relationships with nonbanks to the liabilities side of its balance sheet? Under conditions of sufficient stress, even laws could be changed to allow the state bank to become the economy’s primary supplier of credit, not just liabilities. It is not difficult to imagine political pressure on the central bank to supply credit to a failing General Motors, Volkswagen, Marks and Spencer, or what have you. That means a government bank that not only controls money but also directs credit.
Worse, following a crisis, would the state bank tolerate (let alone encourage) renewed competition from private-sector banks and lending institutions? The U.S. experience in housing finance, which remains dominated by the wholly-owned government-sponsored enterprises long after the crisis of 2007-2009, clearly suggests otherwise. And, as the prime banking supervisor, the central bank could do something about any competitive threat!
We are pleased with the idea that central banks should experiment with digital currency. There is much to learn. And the pace of technical change in the financial sector is clearly accelerating. Central banks will need to adapt to secure their traditional goals of stabilizing prices, activity and the financial system. (See the discussion here and here.) It also is vital that central banks secure the benefits of seignorage for taxpayers, rather than let them be captured by a privileged few (like the founders of Bitcoin).
But we are very cautious about wholesale reforms that would profoundly alter the balance between the private and public provision of financial services. We worry especially about technologies that would add to, rather then diminish, the risk of bank runs. And, about changes that would lead the central bank to become a lender to broad swaths of the economy. Until ways can be found to address these concerns, we will remain skeptical about the elastic, unlimited central bank provision of universal deposit accounts.
Acknowledgments: We thank our friends Charles Goodhart, Jeremy Stein and Paul Tucker for very helpful suggestions.