Thoughts on the FOMC's Monetary Policy Strategy Review -- Part 3 of 4: Policy Tools
To meet their objectives of low, stable inflation combined with high, stable employment, and a resilient financial system, monetary policymakers use various tools to influence financial conditions. In the first two parts of this series on the FOMC’s monetary policy strategy, we discussed the asymmetries of the current framework (here), and the inflation target itself (here). Now we turn to policy implementation, arguing that the FOMC should consider modifying its current toolkit as part of a new strategy.
Specifically, we ask whether the federal funds rate remains the appropriate target interest rate and how it is that policymakers should view asset purchases. We also address possible restrictions on balance sheet policy that the FOMC could consider. We will discuss forward guidance – the remaining policy tool – in the fourth part of this series that focuses on communication.
Before we start, it is worth recalling that monetary policymakers aim to secure their stability goals by stimulating or restraining aggregate demand. They affect aggregate demand by influencing financial conditions. A more accommodative policy increases asset prices. This raises wealth, makes borrowers more creditworthy, lowers the cost of external funds (both debt and equity), and increases lenders’ capacity to lend. Overall, these easier financial conditions increase economic activity, raising inflation.
Importantly, it is not the overnight interbank lending rate – the FOMC’s traditional target – that matters for any of these things. For example, in the housing market it is mortgage rates that drive individuals to purchase homes. For large firms, it is corporate bond rates and equity prices that influence decisions to invest. What policymakers are really doing is using their control over short-term rates to influence the prices of long-term assets, as it is the latter that matters for growth, employment and inflation.
In addition to interest rates, central banks use their balance sheets as a policy tool. The origins of this are straightforward. If the interest rate target hits its effective lower bound (which may be slightly below zero) and policymakers wish to provide further accommodation, then a natural thing to do is to purchase securities. Since the central bank generally purchases long-term securities, this can have a direct impact on long-term yields, equity prices, and credit supply, among other things (see, for example, Drechsler, Savov, Schnabl, and Supera for the impact of Federal Reserve mortgage purchases and sales on housing credit).
We recommend that in their strategy review, the Committee ask both whether using the federal funds rate as its interest rate instrument remains fit for purpose, and how to best use asset purchases. On the former, our suspicion is that the answer is no, and that the Committee should seriously consider shifting to the U.S. Treasury repo rate. On the latter, our reading of the evidence suggests that large-scale asset purchases should be confined to times when policy rates are at their effective lower bound, and that banking system demand should determine the size of the central bank’s balance sheet.
Turning to specifics, prior to the 2007-09 financial crisis, the federal funds market had an average daily volume in the range of $200 billion. Since this is an unsecured market, it could provide early warnings of problems in the banking system. Lenders might even refuse to lend if they became suspicious of specific banks before supervisors did.
But the financial crisis made everyone wary of unsecured lending, so the federal funds market collapsed. Daily volumes of $200 billion in 2007 fell to nearly $50 billion by 2012 (see here). Even today, the volume is only about one-half of the pre-crisis norm. Moreover, the current market has some very peculiar properties. Rather than banks, the primary lenders today are government-sponsored entities (GSEs) – the Federal Home Loan Banks, Fannie Mae and Freddie Mac – and the primary borrowers are foreign banks operating in the United States. Since 2016, GSEs account for an average of 96 percent of the lending while foreign banks receive 84 percent of the loans (see following chart). In short, this market now provides little information, if any, about the health of domestic banks.
Federal funds market: Lenders, borrowers, and volume (quarterly), 2016-2024
Source: Anderson and Na, Figure 3; and FRED.
As Keating and Macchiavelli describe in detail, the current patterns of lending and borrowing in the federal funds market arise from an arbitrage. While the GSEs have accounts at the Fed, they are not interest bearing. So, when a GSE, such as a Federal Home Loan Bank, has cash, they lend it to a bank that can hold it in an interest-bearing reserve account at the Fed. But domestic banks face a deposit insurance premium on all their liabilities, including the short-term borrowing that they simply deposit at the Fed. In contrast, foreign banks are exempt from these FDIC charges, so they face virtually no cost when borrowing from a GSE and depositing the proceeds in a reserve account. In practice, the lenders and the borrowers split the deposit insurance premium when they engage in this transaction, keeping the difference between the effective federal funds rate and the interest on reserve balances (IORB) virtually constant at 7 basis points (see here).
In sum, the federal funds market is now a small, highly idiosyncratic market that has little connection with the financial system or the real economy.
So, a key question for FOMC strategy is which short-term market provides the best alternative to the federal funds market for setting the interest rate target.
In our view, the most obvious target today is the U.S. Treasury repo rate. The repo market – the secured market for short-term funding – is a large, dense market that is critical for the operation of the financial system. The market is divided into two segments – one portion is centrally cleared through the Fixed Income Clearing Corporation (FICC); the other goes through a third party (called tri-party repo). Importantly, daily repo volume is currently around $4 trillion (excluding the Federal Reserve’s facilities), about 40 times larger than the federal funds market (see the following chart). Moreover, the repo rate directly affects the financing costs of both banks and nonbanks, which have become large providers of credit. Given its importance for the funding of many short-term market participants, the repo rate is far more closely tied to financial conditions than the federal funds rate.
Repo transactions volume (Trillions of dollars, daily) 7 May 2018-2 August 2024
Note: Centrally cleared data combines delivery versus payment (DVP) for specific securities with general collateral repo. Tri-party repo, which is not centrally cleared, excludes the Federal Reserve’s Overnight Reverse Repo Operations. Source: Office of Financial Research and FRED.
Now, the Federal Reserve currently operates both a standing repo facility (SRF) and an overnight reverse repo facility (ONRRP) – the former takes in securities and provides cash, while the latter does the opposite. Currently, the ONRRP is merely a device to buttress the impact of the IORB for the world of nonbanks. However, it would be only a small step for the FOMC to announce that it will operate a corridor system for the repo rate by setting a target repo rate that is between the two SRF and ONRRP rates. If they do so, the Committee may wish to consider widening the gap from the current 20 basis points to something like 50 or even 75 basis points.
Such a corridor approach also could aid short-term financial stability. To see what we mean, consider the case of September 17, 2019, when the upper limit of the federal funds target range was 2¼ percent. During the day, a shortage of cash drove general collateral repo rates to 5¼ percent, and some rates went even higher. That is, actual rates in this key short-term market were 300 basis points above the top of the federal funds target range. Clearly this liquidity problem created concern among policymakers, who acted quickly to stabilize the repo market using ad hoc means. Going forward, a policy framework consciously based on a corridor system for the repo rate would reduce, if not eliminate, such temporary spikes, improving the stability of short-term funding markets. (The Fed also may wish to consider various repo market reforms, such as a shift from next-day to same-day settlement that can reduce ordinary liquidity needs, further limiting the disruption from a temporary liquidity shock.)
To continue, what is the role of the Fed’s balance sheet as a tool for monetary policy? We divide this question into three parts: asset purchases aimed at stabilizing the financial system, balance sheet policy at or near the effective lower bound (ELB) for the policy rate, and policy in normal times.
On the first of these, central banks should and do use their capacity to purchase securities as a financial stability tool. Recent examples are easy to find: the Eurosystem’s purchase of sovereign securities starting in mid-2014, the Federal Reserve’s purchase of U.S. Treasury securities at the onset of the pandemic in March 2020, and the Bank of England’s purchase of U.K. gilts in September 2022, to name just a few. In addition, in recent years, central banks often purchased corporate bonds. As these “market-maker of last resort” (MMLR) operations become more common, there is need to develop a framework for when and how they should proceed. Together with Willem Buiter, Kathryn Dominguez and Antonio Sanchez Serrano, one of us (Steve) worked to develop a set of principles for such interventions (see here). We hope that clarifying the framework for MMLR operations is on the agenda for the FOMC’s upcoming strategy review.
Turning to the use of the balance sheet for the purposes of monetary policy, central banks typically aim at influencing aggregate demand (see here). In the aftermath of the pandemic, people asked whether asset purchases are a tool that should only be deployed when policy rates hit the effective lower bound, or whether they have a role in normal times?
To answer this question, one must first evaluate the aggregate demand impact of large asset purchase programs (quantitative easing or QE). Do these work, and if so, when and how? Nearly a quarter of a century after the Bank of Japan first initiated a policy of balance sheet expansion, and over a decade after both the Federal Reserve and the Eurosystem began asset purchase programs with rates at or near zero, we still lack convincing answers to these questions. In fact, in a post in mid-2021, we wrote: “When it comes to quantitative easing (QE), where you stand definitely depends on where you sit.” As Fabo, Jančoková, Kempf and Pástor show, when the analysis of QE (or QT) is done by someone in a central bank, it is much more likely to conclude that balance sheet policy is effective. That said, a very recent academic study (see Selgrad) concludes that QE can have a material impact on corporate bond financing costs.
Importantly, most of this research regards what happens at the ELB. What about the steady-state size of the Fed’s balance sheet?
The following chart shows the evolution of the size of the Fed’s liabilities since 2007. The peak in mid-2022 is nearly $9 trillion – roughly 10 times its pre-crisis level. Since then, it has declined by nearly $2 trillion. Importantly, nearly all the shrinkage is in overnight reverse repo (ONRRP, in orange). Since the Fed began actively shrinking its balance sheet in mid-2022, bank reserves (in blue) averaged $3.27 trillion, remaining within a narrow range of only $3.0 to $3.5 trillion. Extrapolating this pattern, as Fed ONRRP disappears, the system will approach a point where the size of the balance sheet size is consistent with meeting the banking system’s demand for reserves at the going market rate. (That is, reserve supply will be at the point where reserve demand just starts to have negative, rather than zero, slope.)
Consolidated balance sheet of the Federal Reserve System: Liabilities (trillions of dollars, weekly), 2007-2024
Source: Board of Governors of the Federal Reserve, H.4.1 release and Federal Reserve Bank of New York.
Allowing for fluctuations in the demand for currency, the size of the U.S. Treasury’s account, and the accounts of foreign central banks, the current balance sheet is more than sufficient to meet banks’ demand for reserves. Consequently, we see no reason for the FOMC to target an even larger balance sheet.
What about using the balance sheet as a tool for influencing financial conditions (and aggregate demand) in normal times? Here there are two issues. First, when the Fed purchases debt securities, it can only influence interest rates to the extent that investors’ view Treasury (and fully federally insured agency) bonds of different maturity as imperfect substitutes (see the extensive discussion here). While there are term premia that reflect differences in the riskiness of long-term vs short-term bonds, the Fed’s ability to compress those term premia, thereby flattening the yield curve, can be unpredictable. Indeed, there are many other factors—including the creditworthiness and convenience yield of Federal debt—that matter for the term premia (see Gómez-Kram, Kung, and Lustig).
Second, a large Fed balance sheet creates political risks. The primary of these is the risk associated with losses that arise when interest paid on reserves exceeds income from the Fed’s bond portfolio (see here). While we support the use of asset purchase programs to ensure financial stability and provide further policy accommodation at the ELB, our view is that the balance sheet size (and composition) should be largely determined by what various liability holders demand during normal times. (Note: we do not discuss negative interest rates, because we are doubtful about the utility of this tool.)
To conclude, we recommend that the FOMC’s strategy examine two aspects of its toolkit. First, should the Committee shift their interest-rate policy target away from the federal funds rate that has served them so well for the past 40 years? Second, how should the Committee view the size of its balance sheet? Our answer to the first is that the U.S. Treasury repo rate appears to be a better target. Second, we suspect that the best choice is to let the scale of the balance sheet be demand determined and to restrict large-scale purchases to times when policy rate is stuck at the ELB and other tools – such as fiscal policy – are not available to stimulate aggregate demand.