QE Or Not?

For the first time since 2014, the Federal Reserve is growing its balance sheet. Some market observers are calling the Fed's resumption of balance sheet expansion quantitative easing 4 (QE4). We think this is partly misleading. The mechanics of the operations may look similar at first instance, but the intention and operational nuances are different from the balance sheet expansion of 2008-2014.

Under quantitative easing, or "QE," the central bank deliberately expands the size of its balance sheet by acquiring assets (usually government debt securities, but in principle any asset) paid for by creating reserves. (1) The QE episodes in the U.S. since the global financial crisis were designed to ease monetary conditions beyond what the conventional interest rate allowed for (constrained by zero lower bound). In contrast, the current balance sheet expansion is supposed to prevent unwanted tightening of monetary conditions rather than ease monetary conditions--a fine distinction, but an important one.

The announced $60 billion of monthly purchases at least through spring 2020 are almost as large as the $85 billion a month of assets the Fed bought during QE3. But unlike QE3, which increased the supply of bank reserves beyond the minimum needed to meet the interest rate target, the current expansion is merely to stabilize the supply of reserves to restore a healthy buffer needed to control the policy rate--consistent with the Fed's post-crisis operational regime of an "ample" level of reserves. (2) Together with a commitment to carry out overnight and term repo operations at least through January 2020, it is designed to prevent a repeat of the mid-September squeeze in the overnight repo market. (3)

What's more, the purchases will put the New York Fed's SOMA (System Open Market Account) asset holdings above $4.0 trillion by next summer, reversing half of the reduction of the past two years. But, purchases this time around will be of Treasury bills (versus long-dated Treasuries and mortgage-backed securities in the case of QEs over 2008-2014).

The Fed's asset purchase program in 2008-2014 was aimed at extracting duration from the market to push down longer-term yields when the fed funds rate was zero. In the process, the Fed increased the average maturity of its Treasury holdings. In contrast, the current purchases of bills will reduce the average maturity of the Fed's Treasury holdings, and in isolation, this action should also not affect term premium at the longer end of the yield curve--both of which are opposite of previous QEs' outcomes.

Our calculation suggests that the Fed is likely on the trajectory to add roughly $530 billion to its total asset holdings over the next 12 months, front-loaded in the next six months. This would account for:

• The roughly $80 billion organic growth of the balance sheet--as currency liabilities rise, and
• The assumption that $450 billion in additional bank reserves are needed to get back to "ample" levels.

The additional reserves we have penciled in are on the high side because we think the Fed would rather err on the side of too much liquidity rather than too little.

The added reserves would cover the early September levels of $1.5 trillion that the Fed now views as the minimum plus a comfortable buffer of $350 billion, which we presume will allow for distributional issues and one-off fluctuations, such as in the Treasury account. (Also, recall that at year-end 2018, repo rates spiked with $1.7 trillion in reserves.) (4) The addition would bring back reserves from their nadir near $1.4 trillion in mid-September to $1.85 trillion by September 2020.

In our estimation, the size of the Fed's securities held outright (in SOMA holdings) is likely to stabilize near 18% of GDP, about 1.5 percentage points above third-quarter 2019. In comparison, the Fed's securities were 5%-6% of GDP over 2002-2007 and 24% at their peak in 2014 (see chart 1).

Although the Fed will be increasing the size of its balance sheet, in comparison with other major central banks, it will still be rather moderate. In Japan and Switzerland, the expansion has been far bigger, at 80%-100% of GDP, which suggests the Fed has more room to expand its balance sheet (see chart 2).

Chart 1

Chart 2

The Fed will also continue to roll over all its maturing Treasuries and conduct its Treasury purchases related to reinvestments of prepayments of its mortgage-backed securities (MBS) portfolio at a maximum rate of $20 billion per month. The distribution of purchases will match the maturity distribution of Treasuries outstanding (i.e., 15% Treasury bills and 85% notes and bonds).

These two "rolling over" and "swap" operations do not lead to expansion of reserves or the balance sheet overall. Still, at the margin, swapping MBS for Treasuries should lead MBS spreads to drift upward--again, opposite of QE. (The maturing amount above $20 billion will be reinvested in MBS.)

All said, in terms of the mix of assets it holds as well as their duration, the Fed is still normalizing its balance sheet. The accelerated purchases of Treasury bills are unlikely to lead to a material (downward) revision to our forecasts for longer bond yields. They will, however, gradually shorten the weighted-average maturity of the Fed's Treasury holdings to more closely reflect the overall Treasury market (a market-neutral strategy), which would likely have steepening pressure on the yield curve, all else equal.

It's About Operational Control

To keep the interest rate where it wants, the Fed's post-2008 operational regime requires maintenance of "ample reserves," compared with "tight reserves" prior to 2008. (5) Under the current system, the Federal Open Market Committee aims to control its policy rate--the federal funds rate--by keeping an abundant level of reserves that does not require daily active management. In contrast, the open-market operation framework that was in place prior to the financial crisis was predicated on a scarcity of reserves.

Prior to 2008, reserves were "tight" in the sense that they were small relative to GDP, and there was a small amount in excess of banks' required reserves. The demand for reserves moved daily with changing demands for liquidity and required reserves. To ensure that the federal funds rate traded close to the target rate, the Open Market Desk at the New York Fed made small, frequent changes to the supply of reserves within a managed corridor via Open Market Operations (OMO)--corresponding to horizontal shifts in the supply of reserves (in chart 3).

Since the Great Recession, as the Fed's balance sheet grew substantially, bank reserves also exploded in volume. Because the abundance of liquidity rendered the pre-2008 corridor system ineffective, the Fed has used interest on excess reserves (IOER) and an overnight reverse repurchase program (ON RRP) to put a two-tiered floor under the short-term interest rates.

Both IOER (upper bound) and ON RRP (lower bound) are rates the Fed pays holders of reserves--hence the fed funds rate "target range." These holders (potential lenders) have no incentive to part with their cash for less than they get from the Fed, so long as they have access to the programs. The floor system ensures short-term interest rates rise even when reserves are ample.

Control of the federal funds rate is straightforward: The Fed announces the target for the federal funds rate, which shifts the demand curve up or down through arbitrage by banks that earn the interest rate on excess reserves. When reserves are above the ample level (right of the "kink" on the flat portion of demand curve in chart 4), changes in the supply of reserves have little or no impact on the federal funds rate. However, if reserves were to fall below the satiation level, the market price--the effective federal funds rate (EFFR)--would respond as indicated by the negatively sloped portion of the demand curve to the left.

Chart 3

Chart 4

The Federal Reserve has been generally successful in using its postcrisis tools to pull up short-term interest rates across a variety of financial instruments. But that's not to say there weren't any hiccups, such as when surging Treasury security issuance, from a significantly increased budget deficit, led to temporary spikes in Treasury repo rates throughout 2018 and 2019, which, in turn, caused the EFFR to drift toward the upper bound of the target range (instead of the midpoint). (6) (See "Appendix: Strains In The Repo Market Offered The Fed An Incentive To Act Quickly.")

But the mid-September funding stress in the repo market that triggered a sizable increase in the interest rate toppled the proverbial apple cart that had already shown signs of modest funding pressures throughout 2018 and 2019. It suggested that the level of reserves in the system, which had fallen to $1.38 trillion, was no longer "ample." (7) In economic parlance, the reduced aggregate supply of bank reserves hit a point on the demand curve for reserves that was steeply upward sloping--or highly price inelastic.

The Fed was at risk of losing its operational control over the fed funds rate, its key monetary policy implementation tool.

Management Of Bank Reserves

A key issue for the Fed is the appropriate scale of bank reserves, since this affects efficient control of the federal funds rate, liquidity in the banking sector, the transmission of interest rate policy decisions to broad financial conditions, and behavior in money markets both in normal times and during stress episodes.

The mid-September squeeze in the repo market--at the margin--was indicative of a larger underlying trend that has been leading to changes in bank reserves.

As soon as the Fed's aggregate liabilities stopped rising in 2014, reserve supply peaked (at $2.8 trillion) and began to decline (see chart 5). The growing U.S. budget deficits, organic growth of currency in circulation, and the Fed's decision to begin shrinking the balance sheet in October 2017 gradually drained reserves from the system. Consequently, large temporary fluctuations in the supply of reserves, which would have had virtually no impact even a few months earlier, triggered sizable upward interest rate fluctuations.

Three things, in large part, influence the overall level of reserves in the banking system: reserves go down when the central bank reduces its assets, when banknotes (or currency) go up, and when the government deposits go up (see "What Determines The Level Of Bank Reserves?"). Other smaller but common forms of central bank liability, including reverse repurchase agreements, term deposits, and central bank bills, when created, all involve a corresponding reduction in reserves--meaning they involve changing the form of central bank liabilities, but not the amount.

With the Fed choosing to not change the overall size of the assets, at $4.5 trillion over 2014 to fourth-quarter 2017, the rise in nonreserve liabilities (largely currency) gradually displaced reserves. From October 2017 until July 2019, when the Fed decided to shrink its overall balance sheet by allowing some of its maturing assets to expire without replacement (see chart 6), that further accelerated the decline in overall reserves.

Chart 5

Chart 6

Against that backdrop, the funding squeeze the week of Sept. 16 had a couple of well-documented reasonable suspects at the margin: a large volume of corporate tax payments and a spike in settlements for Treasuries (from heavy treasury issuance). The confluence of the two created a scarcity of cash relative to supply of Treasury collateral, which led to a jump in Treasury repo rates.

Corporate tax payments drained funds out of money market mutual funds and banks that could have otherwise been used to support the Treasury repo. At the same time, a large quantity of new Treasury securities settled, decreasing supply of cash for repo financing by the institutions that held the securities. Thus, heavy Treasury issuance and quarterly corporate tax payments together increased the Treasury's general account balances at the Fed and simultaneously drained bank reserves. (8)

What Determines The Level Of Bank Reserves?

In simplified and identity form, a central bank's balance sheet can be written as:

Assets (A) = Reserves (R) + Banknotes (BK) + Government Deposits (GD) + Capital (C)

The most common assets a central bank holds are government debt securities and various claims on banks (repurchase agreements, loans, etc.), often using government debt securities as collateral. Reserves, sometimes called "central bank money" (not to be confused with foreign exchange reserves, which sometimes make an appearance on the asset side), are the deposits at the central bank of banks and other financial institutions that maintain an account with it. The interest rate that the central bank sets on reserves anchors and influences the whole of the term structure of interest rates.

Together with banknotes, these two components make up what is commonly known as "base money" or the "monetary base." The central bank is the banker to the government or its "fiscal agent," so the government maintains a deposit account with the central bank, too, meaning that all fiscal transactions of the government with the public involve the central bank's balance sheet (a very important but usually overlooked fact). The capital account of the central bank is not very consequential because the central bank is usually a part of the government (what is known as the "consolidated government"). The capital of the central bank is little more than a residual accounting item, which appears on both sides of the consolidated government's balance sheet--that is, the liability (right-hand) side as the central bank's capital and the asset (left-hand) side as the Treasury's ownership stake--and so, in effect, cancels out.

Rearranging the central bank balance sheet identity, ignoring equity to simplify, and expressing it in changes, not levels, yields:

Three things influence the overall level of reserves in the banking system: reserves go up when the central bank adds to its assets (this is how it pays for them), when banknotes go down, and when the government spends more than it takes in (from the public) (runs a budget deficit). Because at any point in time the latter two are exogenous to, and observable by, the central bank, it can determine the level of reserves simply by adjusting its asset holdings accordingly (e.g., via open-market operations), increasing its assets if it wants to increase reserves, shedding assets if it wants to decrease reserves.

Source: "QExit Q&A: Everything You Ever Wanted To Know About The Exit From Quantitative Easing," published Aug 17, 2017.

Keeping An Eye On The Current Liquidity Regulations…

That said, no one, including the Fed, knew the precise level of reserves that would trigger such interest rate fluctuation. But an episode like in mid-September had to happen to reveal the point of inadequate reserves. After all, sizable excess reserves are a new phenomenon in the U.S., and assessing the line between adequate and inadequate reserves was bound to be tricky, especially given the new liquidity regulations put in place since the global financial crisis.

On the surface, the spike in repo rates should have created an arbitrage opportunity for providers of liquidity, but some large banks that normally provide overnight liquidity hesitated to fully meet market demand. This indicated that perhaps banks had shrunk their own reserves to their minimum desired level, after accounting for liquidity regulations and nonobvious reasons to prefer holding cash relative to Treasury reverse repo.

The liquidity coverage ratio (LCR) and certain liquidity measurements associated with resolution planning have significantly raised the amount of liquidity the large banks under enhanced regulation, in particular, must maintain on a consolidated basis and at their key subsidiaries. We believe banks also now aim to hold larger cash balances on an intraday basis rather than paying to borrow to meet intraday needs--as they may have done more frequently before the financial crisis.

In addition, banks have other regulatory constraints, such as the leverage ratio and the global systemically important bank buffer, which limit their willingness to grow their balance sheets, particularly at or near quarter-end.

The hoarding of central bank reserves for regulatory reasons becomes a matter of attention for policymakers in that it could adversely affect the transmission of monetary policy. Whenever these intermediaries are reluctant, for regulatory or other reasons, to react to a shortage of liquidity elsewhere in the system--even in response to short-term shocks to the supply of and demand for liquidity--upward pressure on short-term rates becomes unavoidable. That said, Chairman Jerome Powell indicated that the Fed would rather raise the level of reserves in the system than weaken liquidity requirements for banks.

…To Pin Down Bank Reserves' Satiation Point

All said, the mid-September disruptions forced the New York Fed (the arm of the Fed that implements monetary policy) to immediately orchestrate large emergency injections of bank reserves into the banking system via overnight and term repo operations. Subsequently, on Oct. 11, the Fed unveiled new measures that would have its balance sheet start expanding again. With the announcement of new measures, the Fed aims to push reserve balances back up to levels above $1.5 trillion that it now views as the floor for "ample" reserve balances.

Still, determining equilibrium demand at any given time can be a difficult exercise. That means a buffer is prudent to have. And to get to the reasonable buffer of reserves above that level (large enough buffer restoring reserves to stable equilibrium), the Fed will be purchasing assets materially faster than growth of currency in circulation in the coming year. The New York Fed had estimated not long ago that a buffer of at least $350 billion above the structural demand for reserves is needed to account for variation in reserve balances (9), which would put the total bank reserves at $1.85 trillion.

The Fed is trying to get to a satiation point where reserves are not only "ample" but also "efficient"--efficient in the sense it addresses the distributional aspect of reserves while also minimizing increased cost in interest paid to banks for overall excess reserves. According to data from the U.S. Federal Deposit Insurance Corp. (FDIC), 86% of excess reserves are held by just 1% of U.S. banks. Four banks alone account for 40% of aggregate excess reserve holdings in the U.S. When a few big banks hoard reserves, the Fed's operating framework becomes inefficient--as a result, pinning down this optimal satiation point is tricky.

For that reason, we think, as an additional step to ensure sufficient liquidity, the Fed is likely to consider also adopting a standing repo facility (perhaps sometime in first-quarter 2020)--a more permanent facility to provide temporary liquidity whenever needed.

This, together with the Fed's recent action, should help put the issue to rest. (10)

Appendix: Strains In The Repo Market Offered The Fed An Incentive To Act Quickly

In mid-September, the secured overnight financing rate (SOFR)--the heir apparent to LIBOR--which is based off of Treasury repo transactions, spiked to 5.25% (volume-weighted average) from just 2.2% a week earlier. The overnight repurchase, or "repo" market, is where banks and nonbanks get short-term funding by swapping more than $1 trillion of Treasuries for cash each day (11), and at a fundamental level, the surge in Treasury repo rates signaled a scarcity of cash relative to Treasury collateral.

A combination of pressure on bank liquidity, a reduction in money fund cash to be invested in reverse repo, and an increase in Treasury securities needing to be financed in the repo market pushed some primary dealers to pay as much as upwards of 9% (some trades reportedly occurred at 10%) for repo agreements (12) (see chart 7).

Chart 7

The rates since then have come back down to normal levels (13) following the Fed's intervention, but the temporary dislocation was concerning. After all, a smooth functioning of the repo market for Treasuries, arguably the most liquid and safe financial asset, is critical to the plumbing of the broader financial system--it underpins much of the short-term funding that financial institutions (banks and nonbanks) partake to manage their daily cash flow needs. Left unaddressed, short-term rates would have been higher than intended, and the lack of liquidity could have spilled over into the real economy via reduced financial intermediation.

As a sign of how the Treasury repo market is interconnected (14) with other money market rates, even the federal funds rate--or the rate at which financial institutions lend bank reserves overnight to one another on an unsecured basis--rose a bit above its then-prevailing target range of 2.00%-2.25% (see chart 8). Oddly enough, the fed funds rate (the Fed's main policy rate) rose even as the Fed had already begun lowering its policy rate in the current cycle, and market participants were expecting further easing ahead.

Chart 8

Related Research

• The Spike In U.S. Repo Rates Reflects Technical Factors, A Smaller Fed Balance Sheet, And Tighter Bank Regulation, Oct. 2, 2019
• QExit Q&A: Everything You Ever Wanted To Know About The Exit From Quantitative Easing, Aug. 17, 2017
• Implementing Monetary Policy: Perspective from the Open Market Trading Desk, May 18, 2017, New York Federal Reserve (https://www.newyorkfed.org/newsevents/speeches/2017/log170518)
• Credit and Liquidity Programs and the Balance Sheet, the Board of Governors of the Federal Reserve System website (https://www.federalreserve.gov/monetarypolicy/bst_fedsbalancesheet.htm)
• Policy Normalization, the Board of Governors of the Federal Reserve System website (https://www.federalreserve.gov/monetarypolicy/policy-normalization.htm)

Endnotes

(1) Reserves, sometimes called "central bank money" (not to be confused with foreign exchange reserves, which sometimes make an appearance on the asset side), are the deposits at the central bank of banks and other financial institutions that maintain an account with it. The interest rate that the central bank sets on reserves anchors and influences the whole of the term structure of interest rates.

(2) If the central bank supplies more reserves than the banks require (QE), that does not mean the banks will lend more as a result. Whether banks lend or not depends on three things: most importantly whether there are willing borrowers; secondly, on whether the banks have enough equity; and, least importantly, (because central banks always supply the necessary reserves) on whether banks have reserves.

(3) These open-market operations also have the secondary benefit of providing insight on real-world demand for reserves and on how the Fed can play a role in the repo market going forward.

(4) The New York Fed had estimated not long ago that a buffer of at least $350 billion above the structural demand for reserves is needed to account for variation in reserve balances.

(5) Back in January, the Federal Open Market Committee (FOMC) communicated its intention to continue to implement policy according to this regime that requires a large Federal Reserve balance sheet and abundant reserves--the same framework that it has used since the global financial crisis. The principal alternative that the FOMC rejected was conducting policy with a small balance sheet and scarce reserves, as it did before the crisis.

(6) Prior to mid-September, there were already signs of modest funding pressures building in the money market throughout 2018. The effective fed funds rate moved noticeably above the midpoint of the fed funds target range, indicating that the demand for reserves was moderately outstripping supply. In response, the Fed lowered the interest it pays on excess reserves relative to the midpoint of the fed funds target rate three times since mid-2018. Following those adjustments, the Fed was comforted that the effective fed funds rate promptly fell in line with the target. Since the Sept. 16-17 turmoil, the Fed lowered interest on excess reserves two more times, the most recent following the Oct. 29-30 FOMC meeting.

(7) In 2017, the New York Fed had its assumption about the level of reserve balances needed in equilibrium at $500 billion. The New York Fed raised its estimate to $600 billion in March 2018 and then to $1.3 trillion in September 2019. https://www.newyorkfed.org/medialibrary/media/markets/omo

(8) Since Treasury cash balances and bank reserves both reside on the liability side of the Fed's balance sheet, and the Fed was holding the overall size of the balance sheet constant (since July 2018), a rise in Treasury balances typically causes a decline in bank reserves. The week of Sept. 16, the Treasury's cash balance rose by $65 billion while bank reserves fell $62.7 billion on a weekly average basis to $1.39 trillion, their lowest since March 2011.

(9) "Implementing Monetary Policy: Perspective from the Open Market Trading Desk," Lorie Logan, May 18, 2017. https://www.newyorkfed.org/newsevents/speeches/2017/log170518

(10) We do not view the latest episode similar to the repo crisis in 2008. What's different this time is that there isn't a widespread asset collapse, like with house prices in 2008. That made the collateral and debt environment completely different and very unstable. There is no such trigger apparent right now.

(11) Every day more than $1 trillion is borrowed and lent by financial firms through repos, which involves posting Treasury securities as collateral.

(12) A repurchase agreement, or "repo," is a sale of a security with a simultaneous agreement to repurchase it at a later date, usually the next day. Consequently, it is effectively a collateralized loan. A reverse repo is just the same transaction viewed from the other side and is an investment.

(13) The New York Fed orchestrated large emergency injections of bank reserves into the banking system via overnight repo operations Sept. 17-20, and it subsequently announced ongoing overnight and term repo operations, at least through January 2020.

(14) It is not difficult to see how disruptions in the repo market would affect the parallel fed funds market. For example, when the primary dealers--mostly large banks--end up with more Treasury bills than they want or need (say, during times when the U.S. Treasury issues a large amount of new securities, such as Treasury bills), they lend these securities through overnight repurchase agreements in the repo market. In such situations when more Treasury bills are chasing smaller supply of cash (bank reserves), the interest rate needed to clear the repo market rises. As it happens, it also starts attracting cash from parallel fed funds market (lenders see higher repo rates as an attractive alternative investment). As capital flows toward the repo market, the reduced supply of lenders in the fed funds market then puts upward pressure on the fed funds rate, all else equal.