Banks and Money, Or Watch out What You Wish For

“[S]overeign money is an unnecessary and dangerous experiment, which would inflict great damage on our country." Thomas Jordan, Chairman, Swiss National Bank, Speech at the Swiss Institute of Banking and Finance, 3 May 2018

On 10 June 2008, Switzerland held a vote on the proposal that all commercial bank demand deposits be held at the central bank. It failed by a 76% to 24% vote.

This Vollgeld referendum was another incarnation of the justifiable public revulsion to financial crises and the bailouts that inevitably accompany them. Vollgeld proponents claimed that a system in which the central bank is the sole issuer of “money” will be more stable.

Serious people debated the wisdom of this proposal. One of Switzerland’s premier monetary economists, Philippe Bacchetta, wrote passionately in opposition. Martin Wolf, chief economics commentator at the Financial Times, argued in favor. And Swiss National Bank Chairman Thomas Jordan discussed the many dangers in detail.

It should come as no surprise that, had we had been among the Swiss voters, we would have voted “no.” In our view, the Vollgeld (sovereign money) initiative combined aspects of narrow banking with those of retail central bank digital currency. We see these as misguided, distorting the credit allocation mechanism and more likely to reduce than improve financial stability (see here and here). In the remainder of this post, we explain why.

Before getting to any of the details, we start with the textbook definition of money: an asset that is generally accepted as payment for goods and services or repay­ment of debt. So far so good. But moving from a theoretical definition to a practical measure is not so easy. In doing so, people focus on various commercial bank liabilities in deciding what qualifies as money and what does not. Yet, as James Tobin pointed out more than half a century ago, the dividing line between transactions money and other less liquid assets is inherently arbitrary.

For example, should checkable money market funds (MMMFs) be treated as money? In the United States, these function virtually as bank demand deposits. In practice, the Federal Reserve includes them as part of the broad aggregate M2, but not in the narrower transactions measure, M1. M1 is the sum of currency plus demand deposits. M2 adds time deposits, as well as some other highly liquid instruments (like retail MMMFs, but not commercial MMMFs).

To provide some background to the discussion, we turn to the data. The following chart displays the M1 and M2 money multipliers for Switzerland and the United States. These are the ratios of the monetary aggregates to the size of monetary base (MB). MB (the denominator of the ratio) is equal to currency (again) plus commercial bank balances held at the central bank (reserves). The multipliers tell us the ability and willingness of the banking system to convert reserves, supplied by central bank, into what we commonly think of as money. Because it is the monopolist supplier, the central bank can control MB. In practice, most central banks prefer to set an interest rate.

Switzerland and the United States: M1 and M2 money multipliers, 2005 to February 2018

Sources: Swiss National Bank and Board of Governors of the Federal Reserve.

Sources: Swiss National Bank and Board of Governors of the Federal Reserve.

Compared to the historical average, it is perhaps surprising that in both countries the M1 multiplier is so close to one―slightly above in Switzerland (1.2) and slightly below in the United States (0.9). Since currency accounts for only about 15% of Switzerland's monetary base and 12% of M1, Swiss reserves are currently backing over 80% of customer demand deposits. (In the United States, reserves back 95% of total checkable deposits.) In other words, while a forced partitioning of Swiss commercial bank balance sheets―at least of their domestic currency deposits held by domestic residents―would generate some immediate dislocations, it doesn’t look like a terribly big deal. At least, not for now. (The Vollgeld initiative explicitly exempted time deposits.)

Over time, however, a requirement that only the central bank can create demand deposits (either directly or through a segregated commercial bank account) would have pushed the Swiss financial system in two somewhat divergent directions. One would be a further increase in the size of central bank’s balance sheet, and the other would be to drive intermediation into quasi-bank intermediaries (or shadow banks).

Granted, two provisions in the Vollgeld initiative aimed at reducing the incentive to shift funds toward central bank liabilities. First, unlike most narrow banking proposals, under the Swiss proposal savings and investment accounts maintain deposit insurance of CHF 100,000. And second, the proposal granted the SNB the statutory authority to set a minimum holding period for accounts that are not required to be deposited at the central bank. A short minimum holding period would initially limit the volume of central bank liabilities, but the close substitutability of the short-term instruments would facilitate an enormous shift if commercial banks came under stress. Regardless, the combination of guarantees and exemptions could allow conventional banks to maintain at least a portion of their current liabilities.

That said, we still believe a significant fraction of funds could have migrated under the Vollgeld regime. As we discussed in a recent post on central bank digital currency, uninsured depositors with access likely would shift more and more of their deposits into these ultra-safe central bank deposits. The Vollgeld initiative was clear that these retail accounts at the Swiss National Bank (SNB) would have zero interest, but the SNB’s overnight repo rate (SARON) recently has been −0.75%. If the SNB’s liabilities grew, its domestic assets would have to grow as well. Over time, we have argued that such a balance sheet expansion would drive a central bank to lend directly to the private sector, eventually becoming a state bank that substitutes for the private allocation of credit (see here). The alternative of financing a government that then recycles the funds through its own credit supply or through its state banks would have the same impact.

As an aside, it is worth noting that the zero interest rate requirement written in to the initiative would have clearly constrained Swiss monetary policy. Today, the SNB’s target range for Swiss Franc LIBOR is −1.25% to −0.25%. Forcing the central bank to pay zero on retail customer deposits creates a hard floor. If the Vollgeld initiative had passed, current SNB policy would no longer be possible. That is, the effective lower bound on the safest nominal interest rate in Switzerland would be the zero lower bound. And, as SNB Chairman Jordan points out, the proposal also would dramatically limit the central bank’s capacity for foreign exchange intervention, something that they have done in large volume in recent years (see here).

What about funds that do not go into central bank money? We have experience with the economic and financial impact of regulatory caps on the interest rate that banks can pay on demand deposits. From 1933 to 1986, the Federal Reserve’s Regulation Q placed a ceiling on deposit rates that ranged from 2½% initially to 5¾%. In the early 1970s, as market interest rates began to rise far above this cap―the 3-month Treasury Bill rate peaked at over 17% in late 1980―innovative bankers created MMMFs. These nonbank entities hold short-term commercial paper and offer retail customers liquid (fixed net asset value) liabilities. Unlike bank deposits, MMMFs are not covered by deposit insurance. They also do not have official access to the lender of last resort (that is, excluding their emergency bailouts following Lehman’s 2008 failure for which they paid no ex ante insurance fee). At their peak in early 2008, retail money funds included in M2 gathered over $1 trillion, equivalent to 3.3 times commercial bank demand deposits. 

We are not sufficiently familiar with Swiss securities law to anticipate what precise forms of nonbank competition for liquid deposits would arise under a Vollgeld regime. But we suspect that someone would find ways to create various forms of mutual funds that function very similarly to banks. That is, they would transform relatively safe, liquid, short-term liabilities into riskier, less liquid, longer-term assets.

Importantly, these quasi-banks would neither have access to the LOLR nor qualify for standard government guarantees. Yet, to the extent that they engage in bank-like transformation of credit, liquidity, and maturity, they would be subject to bank-like runs and to failure. Indeed, lacking a government safety net, runs would be more likely and more severe than for banks. To the extent that these quasi-banks became a substantial part of the credit system, governments would almost surely be forced to bail them out should they fail. Put differently, illiquidity plays the same role in a world of mutual funds with many investors as it does in the classic Diamond-Dybvig model of a bank run (see our discussion here).

So, we remain supporters of conventional banks. These institutions serve capitalist economies in ways that are costly to replicate. First, they are experts in providing liquidity both to depositors and to borrowers. For the former, it is deposits and for the latter, lines of credit (such as home equity loans and commercial paper backstop facilities). Second, they have expertise both in screening potential borrowers (mitigating adverse selection) and in monitoring those to whom they make loans (reducing moral hazard). That is, they specialize in curbing the information problems that plague all financial transactions. And they are driven by the profit motive to allocate capital to what they perceive to be the most productive opportunities.

But the fact that we believe banks to be indispensable does not mean we support the current system. As we said at the outset, we see the public’s reaction to banking stress and bailouts as something that authorities can and should address. We continue to find reforms like the one proposed in the 2016 Minneapolis Plan attractive. That proposal includes both a large further increase in capital requirements and a tax on borrowing by shadow banks to contain the resulting incentive for regulatory arbitrage by nonbanks.

That is, to make the financial system safe, reforms should focus on making banking activities safe, regardless of the intermediary that performs them. At the same time, reforms should encourage a reliable supply of credit that is efficiently allocated. Vollgeld wouldn’t qualify on either account.