“[Quantitative easing] tends to be deployed in response to bad news, but isn’t reversed when the bad news ends. As a result, the stock of bonds held by central banks ratchets up, expanding their balance sheets into the longer term.”
Former Bank of England Governor Mervyn King, Baron of Lothbury, July 20, 2021.
When it comes to quantitative easing (QE), where you stand definitely depends on where you sit. That is among the conclusions of the important new report of the Economic Affairs Committee of the UK House of Lords.
The report provides an excellent survey of how it is that central banks now use their balance sheets. Its key conclusions are the following. First, central bankers should clearly communicate the rationale for their balance sheet actions, stating what they are doing and why. Second, policymakers should provide more detail on their estimates (and uncertainties) of the effectiveness of their various actions, especially QE. Third, they should be aware that the relationship between central bank balance sheet policy and government debt management policy poses a risk to independence. Finally, and most importantly, central bankers need an exit plan for how they will return to a long-run sustainable level for their balance sheet.
We discussed several of these points in prior posts. On communication, we argued that central bankers should be clear about their reaction function for both interest rate and balance sheet policies (see here). On the justification for policymakers’ actions, we emphasized the need for clear, simple explanations tied to policymakers’ objectives, distinguishing carefully between the intended purposes (such as monetary policy, lender/market maker of last resort, or emergency government finance; see here). And, on the relationship between QE and fiscal finance, we noted how the ballooning of the U.S. Treasury’s balance at the Fed in the early stages of the pandemic looked like monetary finance, putting independence at risk (see here).
In this post, we turn to the challenge that Lord King highlights in the opening quote: the need to ensure that central banks do not see bond purchases as a cure-all for every ill that befalls the economy and the financial system, causing their balance sheets repeatedly to ratchet upward.
Central banks originally were created to finance the government. Later, they became lenders of last resort to stabilize the banking system. What we know as conventional monetary policy—the use of interest rates to achieve price stability with full use of the economy’s productive resources—became the most visible activity of central banks only after World War II. About 30 years ago, a broad consensus emerged that central banks should be independent of fiscal authorities, stop direct government finance, and have a clear inflation objective.
The financial crisis of 2007-09 brought with it another set of changes. To stabilize financial markets, central banks increased both the scale and the scope of their activities. The list of actions is long. To pick just a few, the Fed agreed to provide dollars to a number of other central banks, lend to support dealers in their bond market making activities, and create a set of special purpose vehicles to purchase assets (like commercial paper) directly.
In all these financial stabilization efforts, which involved a combination of lending and market making, balance sheet quantities rose temporarily before returning to zero (or nearly zero). The following chart of the Federal Reserve dollar liquidity swaps is representative. The level went sharply up and then quickly down on three separate occasions. The Lehman bankruptcy in late 2008 triggered the first, and largest, episode. The second peak comes during the euro crisis in mid-2012. And the third is in April 2020, during the financial stress that accompanied the early stages of the pandemic. In each case, there is a spike—the amount lent jumps and then plummets. One can also see this spike pattern--where Fed holdings or loans surge and then plunge—in lending to primary dealers, liquidity provision to support money market mutual funds, commercial paper holdings, and a host of others.
Central bank dollar liquidity swaps (weekly, billions of dollars), 2007-2021
Source: FRED.
However, for monetary policy actions aimed at easing financial conditions to stimulate aggregate demand, the balance sheet path is fundamentally different. The following chart shows Federal Reserve holdings of securities minus currency outstanding. (Growth in U.S. dollar currency in circulation requires the Fed to increase its assets even if all other central bank liabilities are held constant, so we remove that.) Note first that the sustained use of balance sheet tools is closely linked with the effective lower bound on nominal interest rates: when central banks ran out of conventional interest rate space, they turned elsewhere.
Federal Reserve securities held outright minus currency in circulation (month-end, trillions of dollars), 2007-July 2021
Note: Through April 2009, we have truncated the series at zero. Source: FRED.
Once balance sheet policies become the norm, we see the ratchet effect. Outstanding amounts rise markedly, and then fall only modestly before rising again. For the Fed, the first step-up occurred in 2009, boosting non-currency related holdings above $1 trillion. Step two, from October 2010 to July 2011, drove the level to $1.6 trillion. Step three, from December 2012 to September 2014, boosted securities minus currency to $2.9 trillion. Finally, step four—easily the largest—reflects the Fed’s response to the pandemic and its aftermath. At this writing, the Fed holds $7.6 trillion in securities and $2.2 trillion in cash is outstanding, so the total in the chart is $5.4 trillion. But, since Fed purchases are continuing at a rate that far exceeds the growth in currency, we don’t know when this will stop.
Why is there a ratchet? Where does it come from? One possible explanation is rooted in the observation that the impact of asset purchases on financial conditions and aggregate demand is limited. In fact, research produced outside of central banks suggests that there may be no impact at all (see Fabo et. al.). To be sure, large-scale asset purchases likely have a significant impact when there are dislocations in markets that prevent arbitrage. But when markets are working well—which is most of the time—purchases may operate only by adding to the credibility of central bank forward guidance (see here). From the perspective of aggregate demand management, this “signaling only” impact means that there is little short-run risk of doing too much and a strong incentive to err on the side of doing more. A second cause of the ratchet is the one Lord King highlights in the opening quote: the belief that central banks can counter the impact of bad news on the financial system and the real economy by buying more bonds.
Not only are central bank balance sheets now routinely ratcheting higher, but the range of their actions also is considerably broader. Again, focusing on the Fed, the pandemic triggered interventions in corporate bond markets, municipal bond markets, and bank business lending (see our summary here). In most cases, the totals remain quite small—just over $30 billion for all credit facilities combined—and the facilities are now winding down. However, the same cannot be said of mortgage-backed securities, where Fed holdings increased by $1 trillion over the past 18 months and continue to rise.
We see numerous risks in the expansion of the scope and scale of central bank balance sheet actions. First, as we wrote in an earlier post, to ensure that central banks retain the independence to do well what they are designed to do, we need to impose clear boundaries on the scope of what they are authorized to do, limiting both what they can buy outright and to whom they can lend. As it stands, many people believe that, regardless of conditions, central banks can stabilize any market, rescue any borrower, and neutralize any bad economic or financial shock. This simply is not true.
Second, to ensure the sustainability of their balance sheets, central bankers need to counter the QE ratchet. This means providing clear guidance on their view of what is a normal balance sheet (in size and pace of expansion) and how policymakers plan to get there.
For example, suppose that the Fed wishes to supply reserves comfortably (but not excessively) above the level needed to keep overnight interest rates near their target without frequent open-market operations. The events of September 2019 give us a hint about what scale might fit this “ample reserves” regime. As we wrote at the time, from January 2018 to September 2019, the Fed contracted its securities holdings by $650 billion, allowing commercial bank reserves to fall by nearly $800 billion. In retrospect, they went too far, as the overnight Treasury repo rate spiked to 6% on September 17. From this experience, in the absence of structural or regulatory changes to the financial system, one could conclude that the level of securities minus currency in summer 2019—something in the range of $2 to $2½ trillion—is close to the sustainable level. (For reserves, the level is probably in the range of $1½ to 1¾ trillion, less than one-half the current $4 trillion).
Returning to the report from the UK House of Lords, we agree with the fundamental message that QE has gone awry. Over time, central banks will need to pull back. More immediately, they should provide much greater clarity on how they link their balance sheet actions quantitatively to financial conditions and aggregate demand. Most importantly, policymakers need to tell how they will exit when they near their inflation and employment goals.
They should answer the following questions: What will lead you to stop purchasing securities? What are the conditions that will lead you to sell securities? What is the long-run normal and sustainable size of your balance sheet? In normal conditions, how rapidly will your balance sheet grow? When economic stability is restored, how quickly will you shrink it back to the normal level?
Obviously, any answers to these questions will be conditional, so they can change as the economic and financial environment evolves. But policy planning is always conditional, so it is natural that the explanations will be, too.