Author: Brian Poi, Nomis Data Scientist
With the recent re-introduction of quantitative easing by the Federal Reserve, now is a good time to review some questions that arose in the aftermath of the Great Recession. When the Fed first embarked on quantitative easing, I was asked how such actions would affect bank deposits. Some bankers surmised that quantitative easing might increase aggregate deposit balances, and other bankers suggested that as the Fed unwound or sold off its bond portfolio deposits would shrink. In short, I do not believe quantitative easing has much impact on deposit balances, either as quantitative easing is implemented or as it is unwound. Before I explain why that is the case, it is helpful to have some background on quantitative easing and how it affects the economy. I then present my case.
In normal times the Federal Reserve adjusts the target level for the federal funds rate, the interest rate at which banks lend excess reserves to one another on an overnight basis. This short-term interest rate affects longer-term interest rates, so changes in the federal funds target rate can help speed up or cool the economy. Once the federal funds rate has been reduced to zero, the Fed must become more creative. While some central banks in Europe have experimented with negative interest rates, the Fed has indicated that it does not wish to do so, and I consider the possibility remote.
In the aftermath of the Great Recession the Fed began directly purchases government and other bonds. In March 2009 the Fed began what has come to be known as QE1, purchasing $600 billion in mortgage backed securities. In November 2010 QE2 began, with the Fed purchasing Treasury securities; and in September 2013 QE3 kicked off with the Fed began purchasing agency MBS.
In just the past month the Fed has announced additional actions that could easily amount to $5 trillion in an effort to buffer the economic shock caused by the COVID-19 virus. Various facilities will purchase Treasury securities and agency MBS. Expanding the traditional scope of the Fed, it will also purchase commercial paper, even to companies that have recently lost their investment-grade ratings. Another program will purchase short-term securities issued by some municipalities. Other actions are being performed in conjunction with the Treasury Department to ensure small businesses have access to credit, even beyond the provisions of the CARES Act.
Quantitative easing has had three main purposes. One is to ensure the economy has sufficient liquidity to function. That is why the Fed has decided to purchase commercial paper and municipal securities and to provide (indirectly) loans to small businesses. Without access to funding, some businesses and municipalities would be unable to retain workers on payroll, further exacerbating what is already a dire unemployment situation. The second purpose is to remove toxic assets from bank balance sheets, as was the case with QE1; with banks on sounder ground now, this purpose plays only a minor role.
The final purpose of quantitative easing is to stimulate the economy. The Fed’s purchase of bonds will drive prices up and yields down, in turn leading to lower interest rates and increased borrowing. QE3 is a prime example; the purchase of agency MBS was designed to lower mortgage rates and stimulate the housing market which at the time was still lackluster. Lower interest rates also increase the attractiveness of stocks, since lower interest rates make future cash flows more valuable today. Higher stock prices increase consumption through the so-called wealth effect.
Even if the Fed only purchases Treasury securities, it will still have an impact on other interest rates and, potentially, on economic activity. Mortgage securities, corporate bonds, and other debt instruments trade at a risk premium to Treasury yields; if the risk premium holds constant, a decline in Treasury yields will cause yields on those obligations to fall a similar amount. In fact, risk premia may actually fall as investors “reach for yield.” In all cases, the path to economic stimulation is the same. Lower interest rates make investment more attractive, whether that investment is residential housing, new plants and equipment, or research and development. Of course, the key question is by how much interest rates decline in response to quantitative easing. Estimates vary widely, but the literature suggests that QE1, QE2, and QE3 combined lowered Treasury yields by somewhere between 25 basis points on the low end to more than 100 basis points on the high end.
Strictly speaking, the Fed only trades bonds with a network of 24 primary dealer banks who trade with other banks and bondholders, but for our purposes we can ignore this level of intermediation. When the Fed purchases bonds, we can think of two different classes of sellers: banks and non-banks.
First suppose a bank sells a bond to the Fed. In exchange the bank receives cash that it could use to make new loans, or it could simply hold that extra cash at the Fed as additional reserves. The only way this type of transaction could stimulate the economy directly is if the bank makes a new loan or if it lends its excess reserves to another bank that makes a new loan. Without a bank making a new loan, aggregate deposits would remain unchanged.
Now suppose a non-bank entity such as a pension fund sells a bond. The pension fund would receive cash that in theory it could deposit at a bank and earn interest. Much more likely, however, would be for the pension fund to purchase another bond or invest in another asset class such as equities or real estate. Again, the proceeds from the Fed’s purchase of the bond is unlikely to flow toward bank deposits.
Some salient charts
Figure 1 shows total assets of the Federal Reserve System along with aggregate excess reserves held by banks on deposit with the Fed, and Figure 2 shows total assets along with aggregate loans and leases outstanding.1 As we explained above, if a bank sells a bond to the Fed, one alternative for the bank is to hold the proceeds as reserves. The vertical lines in the figures represent the starting dates of QE1, QE2, and QE3. These figures strongly support the proposition that banks did not use their proceeds from bond sales to increase lending until at least 2014, and it took five years for excess reserves to return to pre-QE3 levels. The path from quantitative easing to lending is slow at best.
Figure 3 shows total assets along with the M2 monetary aggregate, which is predominantly composed of bank deposits. Considering the long-term trend in M2, Figure 3 suggests that bank deposits are not affected by quantitative easing. Of course, simple graphical analyses do not control for intervening effects and cannot prove that one series causes, or does not cause, another series. Nevertheless, the graphs shown here strongly suggest that quantitative easing has very modest effects on lending and no discernable effect on deposits.
Several years ago my colleagues Mark Zandi, Samuel Malone, Tony Hughes, and I wrote a white paper (downloadable here) examining in detail the effects of quantitative easing on bank deposits. We considered simulations from a widely used large-scale Keynesian macroeconometric model and various vector autoregression models. Japan began using quantitative easing in 2001, and so we exhaustively examined the behavior of bank deposits in that country as well. We wrote that paper in response to a question posed by a large money center bank worried that the unwinding of quantitative easing would lead to an outflow of deposits as investors purchased securities sold by the Fed. Our analysis there confirms what I have argued above. Namely, quantitative easing does not have any appreciable impact on bank deposits.
When the policy rate is at the zero lower bound, the Federal Reserve must use alternative measures to stimulate the economy, and those decisions must be made in real time. As a result those measures could have undesirable side effects that only become known much later. Some industry analysts have expressed concerns that quantitative easing could influence bank deposits. As the Fed purchases bonds, investors’ proceeds would then flow into the banking system. Similarly, as the Fed unwinds quantitative easing by selling bonds, investors could shift funds out of bank deposits. A closer examination of the mechanics of quantitative easing, along with empirical research, shows that those concerns are misguided.
1All data were obtained from the Federal Reserve Bank of St. Louis FRED economic database. The data originate from various statistical releases of the Federal Reserve System Board of Governors. The small jumps seen in the loans outstanding series are an artifact of the graph; loans outstanding are reported on a quarterly basis, while we used a monthly series for the balance sheet variable.