AS EXPECTED, the Federal Reserve announced on September 20th that it will soon begin reversing the asset purchases it made during and after the financial crisis. From October, America’s central bank will stop reinvesting all of the money it receives when its assets start to mature. As a result, its $4.5trn balance-sheet will gradually shrink. However, the Fed did not give any clues as to what the endpoint for the balance-sheet should be. This is an important question. There are strong arguments for keeping the balance-sheet large. In fact, it might be better were the Fed not shedding any assets at all. 

Most commentators view a large balance-sheet, which is the result of quantitative easing (QE), as an extraordinary economic stimulus. Janet Yellen, the Fed’s chair, seems to agree: at a press conference after the Fed announcement, she said the balance-sheet should shrink because the stimulus it provides to the economy is no longer needed. But the claim that the balance-sheet is stimulating the economy is far from an established fact. The theoretical case for it is weak (Ben Bernanke, Mrs Yellen’s predecessor, famously quipped that QE “works in practice, but it doesn’t work in theory”). While most studies have shown that QE brought down long-term interest rates, it may have worked by signalling that the central bank was serious about keeping rates low for a long time, or by lubricating financial markets during the credit crunch. If so, the Fed’s balance-sheet is probably not providing much stimulus to the economy today. Markets have not seemed to react much the Fed’s signals that today’s announcement was coming (see article). 

How, then, should we think about the right size for the balance-sheet? An alternative way of viewing QE is as an operation that changes the maturity profile of government debt. The Fed hoovers up Treasury bonds, and replaces them with newly-created bank reserves. Those reserves are just a different type of government liability. Just as the Treasury pays interest on bonds it has issued, the Fed pays interest on bank reserves. The Fed and the Treasury both, in practice, represent the same borrower: the federal government. The main difference between the two types of debt is that Fed reserves are perpetual and instantaneously redeemable—in other words, they are money. Treasuries, by contrast, mature at a known date (though their liquidity and safety makes them resemble money). 

The best size for the Fed’s balance-sheet therefore depends on the best maturity profile for government debt, once the liabilities of the Fed and the Treasury have been combined. If money is more useful than Treasury bills, then the Fed performs a useful service by swapping one for the other. And money is useful. In a paper from 2016, Robin Greenwood, Samuel Hanson and Jeremy Stein argue that abundant bank reserves increase financial stability. They note rampant demand in financial markets for money-like instruments: from 1983 to 2009 one-week Treasury bills yielded, on average, fully 72 basis points less than six-month bills. (For comparison, the difference in yield between a five-year Treasury and a ten-year one is below 50 basis points).

The authors argue that when the government fails to satisfy this demand, the private sector steps in, by issuing very short-term debt like asset-backed commercial paper. This threatens financial stability, because such debt tends to be risky. A run on money-market funds, which had gorged on short-term private debt, was central to the meltdown in financial markets in late 2008.

So why not create plentiful reserves to fulfil the demand for money? One problem is interest-rate risk.  The shorter the maturity profile of government debt, the more painful an unexpected rise in interest rates would be for taxpayers. In practice, this works as follows. To raise interest rates, the Fed must immediately pay more interest on the reserves it has created. Yet it would not receive higher yields on the Treasury securities it owns. So an unexpected rise in interest rates could ultimately threaten the solvency of a central bank with a large balance-sheet. The taxpayer would have to plug the gap.

One solution to this problem, discussed by the authors, is a so-called “barbell” structure for government debt. This means a combination of very short-term liabilities, with the associated benefits for financial markets, and very long-term debt, which guards against interest-rate risk. The Fed could implement a barbell structure by buying up medium-dated Treasury bonds, thereby taking them out of the hands of the public. (This is probably better than the Treasury implementing a barbell structure on its own, because short-term Treasuries, unlike reserves, must be reissued frequently at auction.)

Unfortunately, the Fed is unlikely to give such considerations much weight. Its large balance-sheet is controversial. Many observers fail to see that its liabilities are another form of government debt, claiming instead that they “distort” the market for Treasuries. Similarly, The Fed’s critics claim the interest it pays on reserves constitutes a subsidy to banks. They fail to note that if banks held Treasuries instead, they would earn interest on those, too.

The political power of these worries will help determine the end point for the Fed’s balance-sheet. The Fed will need more assets than it did before the financial crisis, because of increased demand for currency, another central bank liability. There is currently $1.5trn of currency in circulation, up from about $800bn before the financial crisis. Keeping the current system for setting interest rates, which works by saturating banks with reserves and then paying interest on them, would probably require another $1trn of assets. However, the Fed might revert to the system it used before QE, when it controlled rates by keeping reserves scarce and tweaking their supply.

Which system is used, and hence the ultimate size of the balance-sheet, is a decision that will probably be taken after President Donald Trump decides whether or not to reappoint Mrs Yellen as chair. It is an important one. Economists since Milton Friedman have noted the benefits of interest-bearing money like bank reserves. The Fed should make sure there is lots of it about.