Kenneth D. Garbade
The Federal Reserve announced on November 3, 2010, that in the interest of stimulating economic recovery, it would purchase $600 billion of longer-term Treasury securities. The announcement led some commentators to conjecture that the Fed’s large-scale asset purchase (LSAP) program—popularly known as “quantitative easing”—is more likely to trigger inflation than stimulate recovery. This post discusses why those concerns may be misplaced, and also why they are not without some basis. A recent Liberty Street Economics post by James J. McAndrews—“Will the Federal Reserve’s Asset Purchases Lead to Higher Inflation?” addressed the same issue from a broader perspective and came to a substantially similar conclusion.
A Simple Money Multiplier Model
Let’s begin by recalling the textbook account of what happens when the Fed purchases Treasury securities and pays for the purchases by adding to bank reserves.
If the banking system initially holds just enough reserves to satisfy its reserve requirements, the addition of new reserves will leave it with an excess of reserves. Prior to the fall of 2008, reserves did not earn interest, and as a consequence, banks holding excess reserves had an incentive to convert them to interest-earning loans. The loans would be extended in the form of credits to borrowers’ bank accounts and would, therefore, increase aggregate reserve requirements. This process of expanding loans and expanding reserve requirements would continue—possibly over a prolonged interval of time—until all of the reserves added to the banking system by the Fed’s original purchases had been absorbed.
To see this point through a numerical example, suppose that banks are required to keep 5 percent of their deposit liabilities in the form of reserves, and that the Fed bought $1 billion of Treasury securities. The additional reserves would support $20 billion of additional deposits, because 5 percent of $20 billion equals $1 billion.
The consequences of an expansion in bank loans and deposits depend on the level of economic activity during and following the expansion. If activity is slack, the additional loans can be expected to stimulate consumption (if the loans are to consumers) and investment (if the loans are to businesses), and hence stimulate aggregate production. However, if there is little slack in the economy, the additional demand from consumers and business will translate largely into inflation. Because the expansion of loans and deposits may take place over a prolonged interval, monetary stimulus (to either production or inflation) is likely to have long and variable lags.
Some commentators are concerned that the Fed’s large-scale asset purchases will trigger an expansion of bank lending and stimulate consumption and/or investment exactly when the economy is recovering for other reasons, and hence will trigger inflation. However, this concern assumes that the extra bank reserves will actually result in increased bank lending, which is far from obvious. To see why, we need to consider several alternative ways for the government to buy Treasury bonds from the public.
Some Alternative Bond Purchase Mechanisms
Suppose first that the Treasury, instead of the Fed, buys Treasury bonds from the public (in buyback operations like those undertaken in 2000 and 2001) and that it pays for the bonds by issuing thirteen-week and twenty-six-week Treasury bills in conventional bill auctions. In this scenario, the government’s debt management operation would have no effect on bank reserves, so there is little reason to think it would stimulate economic activity through a money multiplier effect. (However, the reduction in the supply of longer-term Treasury debt, together with the increase in the supply of shorter-term Treasury debt, can be expected to flatten the yield curve from bills to bonds, reducing bond yields relative to bill rates.)
In the second scenario, the Fed, rather than the Treasury, buys Treasury bonds and finances the purchases by issuing short-term interest-bearing debt in the form of “Fed” bills. Since the public is left in exactly the same position as in the first scenario (it holds more bills—now Fed bills instead of Treasury bills—and fewer bonds), the same economic consequences can be expected.
The Money Multiplier Model When the Fed Pays Interest on Bank Reserves
Now consider a third scenario, in which reserve accounts at Federal Reserve Banks earn interest—as they have since the fall of 2008—and the Fed finances its purchases of Treasury bonds by crediting those reserve accounts instead of by issuing Fed bills. As long as the interest paid on reserve accounts is high enough to deter banks from converting their new reserves to interest-earning loans to individuals and corporations, this third scenario is functionally equivalent to the two preceding scenarios.
Since this third scenario essentially describes the Fed’s LSAP program, there is no substantial basis for believing that the LSAP program will necessarily trigger inflation. More particularly—and this is the key point of the present post—paying interest on reserves can immobilize reserves in a way not contemplated by the traditional pre-2008 money multiplier model.
There is, however, a caveat to this conclusion: the Fed has to ensure that it keeps the interest rate on reserves high enough to deter banks from lending out their excess reserves. (This is functionally equivalent to saying, in the second scenario, that the Fed has to be willing to pay enough interest on Fed bills to keep the public willing to hold those bills.) One can reasonably ask how the Fed can tell when the interest rate on reserve balances is too low (relative to market rates of interest) and banks are actively seeking to lend out reserves in excess of required levels. This question is important because, as noted earlier, stimulative monetary policies can have long and variable lags, so any additional delay in identifying a change in contemporaneous market conditions can exacerbate fluctuations in production and inflation.
The views expressed in this blog are those of the author(s) and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System. Any errors or omissions are the responsibility of the author(s).
Thank you for your comment. Perhaps I could clarify a couple of points. First, paying interest on reserves is a new tool that allows the Federal Reserve to better control the fed funds rate. I would not think of the tool itself as being either contractionary or expansionary. Rather, this tool allows the FOMC to change the stance of monetary policy, making it more or less accommodative as the Committee deems appropriate, in a way that is largely independent of the quantity of excess reserves. Second, paying interest on reserves does not necessarily reduce the supply of loanable funds or absorb saving. Perhaps what you have in mind here is that banks may choose to respond to a quantitative easing program by making fewer loans rather than expanding their balance sheets. While certainly a possibility, the evidence seems to suggest that the Fed’s asset purchases are indeed having an expansionary effect on the economy. See, for example, the earlier post on the Liberty Street Economics blog by Curdia and Ferrero: http://libertystreeteconomics.newyorkfed.org/2011/05/how-much-will-the-second-round-of-large-scale-asset-purchases-affect-inflation-and-unemployment.html
IOeRs are not only contractionary, they exert a depressing effect upon the economy. IOeRs induce dis-intermediation (an outflow of funds) from the non-banks (e.g., MMMFs). IOeRs reduce the supply of loan-funds, increase the cost of loan-funds, and absorb existing savings (preventing savings from being matched with real investment).