The continued unwinding of large central bank balance sheets in the post-pandemic period and its impact on financial markets and the broader economy have presented several policy and monetary policy implementation questions in the U.S. and abroad. To help address this, the Federal Reserve Bank of New York and Columbia University’s School of International and Public Affairs hosted a workshop on “Unwinding Large Central Bank Balance Sheets” in May. The workshop’s three panels brought together academics, industry and public policy experts, and current and former central bankers. In this article, we highlight the main themes from the day’s discussions.
Quantitative Easing and Quantitative Tightening: Impacts and Differences
With the global financial crisis and COVID-19 pandemic in recent memory, participants assessed the impacts of central bank balance sheet changes on economic and financial market conditions. Panelists discussed how quantitative easing (QE)—or when a central bank increases its balance sheet through asset purchases—is most beneficial when addressing deep downturns. QE can boost growth and inflation when inflation is below target, the policy rate is at the effective lower bound (often near zero), and stimulus is needed. It also potentially improves the long-term fiscal outlook, as increased tax revenues from the economic recovery may exceed future central bank losses on bond holdings. While QE may still be beneficial during shallower downturns, it’s less so given the greater chance for inflation to overshoot and for the central bank to realize relatively large losses if the economic recovery is strong. Central banks should therefore strive to adapt quickly as the outlook changes, participants said.
Panelists also discussed how the impacts of quantitative tightening (QT), or when a central bank reduces its balance sheet, can be symmetric to those of QE, depending on the channels through which they operate. For example, the impacts are largely symmetric within the portfolio rebalancing channel—when central bank bond purchases or sales affect the term premium, or the compensation investors demand for holding a long-term bond, rather than a short-term one. In the U.S., for instance, the effects of surprises to balance sheet size are broadly comparable in QE and QT periods. Although QT announcements have generally been associated with much smaller yield movements than QE announcements, this is largely because they have been well anticipated and priced. In other ways, however, QT is different from QE: QT is usually carried out in calm markets, away from the effective lower bound, with its size reasonably clear from forecasts, and with minimal intended signaling regarding the path of future policy rates, all in contrast to QE. The housing market also introduces a key asymmetry: QE reduces long-term yields, resulting in low long-term mortgage rates that households can lock in. QT cannot unwind this effect.
Determining Demand for Reserves
As reserves—that is, deposits that commercial banks hold at the central bank for activities like managing liquidity, making payments, and meeting requirements set by bank regulations—have fluctuated with recent balance sheet changes, participants discussed methods for evaluating reserve conditions. Some panelists emphasized the importance of the repo market—with its large size, diversity of participants, and rapid reaction to changes in the supply and demand for funding—for assessing financial conditions and whether reserves are at an ample level. In fact, some participants interpreted recent volatility in repo markets as an early indication that the U.S. is approaching an ample level of reserves. While some volatility is acceptable—particularly if it occurs at predictable dates with the causes well understood—very large or persistent increases in volatility are not desirable.
Other panelists noted that, for large banks, the constraint on their activity in money markets is generally not how many reserves they hold, but rather their need to meet prudential liquidity requirements, including the liquidity coverage ratio and internal liquidity stress tests. In general, acquiring more reserves by borrowing overnight in the federal funds or repo market is not helpful in meeting these requirements. Rather they may need to issue term debt or convert non-high-quality liquid assets into high-quality liquid assets—for example via a sale.
Other panelists discussed the reserve-demand curve, which traces the rates at which banks are willing to trade their reserve balances and how that varies with the aggregate supply of reserves provided by the central bank. Some panelists proposed an approach that models the underlying structure of the market for reserves and estimates the model with microstructure data. Panelists also emphasized that, for central banks, a point estimate of reserves demand is not sufficient, given that demand can move in response to shocks. Policymakers must take account of uncertainty and variability in reserve demand and ensure that increases do not lead to a shift from ample to scarce reserves.
Estimating the optimal supply of reserves was also the focus of a keynote address by New York Fed President John Williams. President Williams presented a model in which aggregate demand for reserves is nonlinear and subject to shocks. Shocks that reduce the demand for reserves cause short-term rates to fall relatively modestly, since they shift the short-term rate along the gently sloping segment of the reserve demand curve. By contrast, shocks that increase reserve demand cause short-term rates to increase by more, since they shift the short-term rate along the steeper segment of the reserve demand curve. To avoid the large increases in market rates associated with positive reserve demand shocks, a central bank targeting short-term interest rates will supply more reserves when facing uncertainty. The existence of an effective central bank lending facility reduces the optimal supply of reserves, since positive demand shocks can be accommodated by use of the facility, dampening spikes in money market rates.
Different Frameworks for Rate Control Share Commonalities
In looking at monetary policy implementation frameworks across countries, some common themes emerged. Balance sheets are expected to remain larger than they were before the global financial crisis, driven by greater demand for central bank liabilities. But at the same time, they are expected to shrink meaningfully as a share of GDP compared with the recent peak. Most panelists also emphasized the need for central banks to be responsive in supplying reserves if needed, either through demand-driven facilities or discretionary central bank actions.
Panelists also discussed factors that contribute to differences in a central bank’s implementation framework. These include the structure of financial markets, especially the footprint of banks versus nonbank financial institutions; policymaker views on the cost of a large balance sheet and large-scale participation in repo markets by the central bank; and the degree of stigma in standing lending facilities. Some participants also emphasized the importance of past policy decisions. Central banks that expanded their balance sheets to provide accommodation during past crises have tended to retain ample reserves frameworks, while several central banks that did not engage in large-scale asset purchases have continued to use scarce reserves frameworks.
Focusing on the U.S., System Open Market Account Manager Roberto Perli noted that the Federal Reserve’s framework for implementing monetary policy has delivered strong rate control by appropriately transmitting the stance of monetary policy to other money markets and the broader economy in the midst of changing economic and financial market conditions. The Federal Reserve is currently reducing the size of its balance sheet, but demand for Federal Reserve liabilities, including reserves, has increased substantially over recent years. This means that the balance sheet will need to remain larger in nominal terms than it had been prior to the pandemic. Perli further noted that the standing repo facility (SRF) would grow in importance as reserve levels declined, and that for the SRF to be effective, primary dealers and banks must be willing to use it when it makes economic sense for them to do so. To support this, the Federal Reserve has been working to reduce frictions to SRF use, including introducing morning-settling SRF operations in the near future.
Overall, the frameworks used by advanced economy central banks all achieve strong rate control by supplying an ample quantity of reserves to accommodate increased demand for central bank liabilities.
Marco Cipriani is head of Money and Payments Studies in the Federal Reserve Bank of New York’s Research and Statistics Group.
Richard Finlay is an advisor in the New York Fed’s Markets Group, on secondment from the Reserve Bank of Australia.
The views expressed in this article are those of the contributing authors and do not necessarily reflect the position of the New York Fed or the Federal Reserve System.
