Will The Fed Start Reducing Its Balance Sheet? Here Is Goldman's Answer

With the fate of Fed's balance sheet suddenly under Wall Street's spotlight, following last week's hints by several Fed presidents that a runoff in the balance sheet may be on the horizon and prompting various sellside analysts to share their thoughts. Overnight, Goldman too decided to opine on the rising debate of what happens next to the Fed's $4.2 trillion balance sheet, and cutting to the case, says that it continues to expect full reinvestments to end in the middle of 2018 (i.e., no runoff for at least 18 months), but adds that while "we would be very surprised to see a discussion of asset sales under Chair Yellen’s leadership" a shift to "more active management of the maturity of new Treasury purchases could be an option; shortening the duration of new purchases would quicken portfolio runoff once it begins."

Goldman also confirms what other analysts have said previously, namely that "ending reinvestments would result in an increase in MBS issuance to private investors. For Treasuries, the impact on duration supply will depend on how the incoming administration chooses to adjust its sources of financing."

However, should inflation indeed spike up and surprise to the upside as Jeff Gundlach recently hinted, the Fed may have no choice but to engage in just this kind of balance sheet deleveraging, which many have said should have taken place prior to the Fed's launch of rite hikes in December of 2015.

For more details on Goldman's opinion, read the full Goldman Q&A on the Fed’s Balance Sheet

  • Recent public comments from Fed officials have renewed interest in the outlook for the central bank’s balance sheet. Here we tackle the most common questions from investors.
  • At the moment, the Fed fully reinvests principal payments into Treasuries and agency MBS, and we still expect this to continue until the middle of 2018. The committee could decide to end full reinvestment sooner due to (1) unexpectedly strong growth, (2) concern about excessive dollar appreciation, and/or (3) a desire to ensure a smooth transition to the next Fed Chair. However, we would expect the FOMC to proceed cautiously after 2013’s “taper tantrum”, and it may need to consider how its actions intersect with debt management and regulatory goals.
  • We would be very surprised to see a discussion of asset sales under Chair Yellen’s leadership, but a shift to more active management of the maturity of new Treasury purchases could be an option; shortening the duration of new purchases would quicken portfolio runoff once it begins.
  • Ending reinvestments would result in an increase in MBS issuance to private investors. For Treasuries, the impact on duration supply will depend on how the incoming administration chooses to adjust its sources of financing.
  • For broader financial conditions, the impact will likely depend on how the committee communicates the end to reinvestments. One of the important lessons from the Fed’s experience with QE has been that signaling channels appear most important. The same probably goes for winding down the balance sheet: the market implications will likely depend on what these steps tell us about policymakers’ broader intentions.

Q: What is the current status of the balance sheet, and what has the FOMC said about its outlook?

A: The Federal Reserve currently reinvests all principal payments from its Treasury, agency debt, and agency MBS portfolios, thereby holding the nominal size of its securities portfolio unchanged. The Federal Reserve Bank of New York conducts agency MBS purchases in the open market on behalf of the FOMC. For maturing Treasury securities, the Fed rolls over all proceeds into newly-issued Treasuries at auction (Exhibit 1). These purchases do not compete with other investors: the Federal Reserve submits noncompetitive bids, and the Treasury increases the size of its auctions to match the amount of the Fed’s request. At the moment, the Fed does not actively manage the duration of its purchases: it simply allocates its bids proportionally to Treasury’s public offering amounts.

Exhibit 1: Fed Replacing Maturing Treasuries at Auction

Source: Treasury, Goldman Sachs Global Investment Research

In its December 2015 statement, the FOMC indicated that full reinvestment of the securities portfolio would continue until “normalization of the level of the federal funds rate is well under way”. In comments shortly after that meeting, New York Fed President Dudley indicated that the reinvestment decision would hinge mostly on the backdrop for growth: “Now the words ‘well underway’ in the FOMC statement are vague … If the economy were growing very quickly and the risks of an early return to the zero lower bound for the federal funds rate were deemed to be low, then I could see ending reinvestment at a relatively low federal funds rate. In contrast, if the economy lacked forward momentum and the risks of a return to the zero lower bound were judged to be considerably higher, I would want to continue reinvestment until the federal funds rate was higher.”

Market participants have interpreted “well under way” to mean 3-4 additional funds rate increases from current levels. In its regular surveys the New York Fed asks for the expected level of the funds rate “when the Committee first changes its reinvestment policy”. The median response from primary dealer economists in the December survey was 1.38%, and the median response from the separate investor survey was 1.56%. In both cases respondents thought reinvestment policy would change about 18 months from December, or around the middle of 2018.

Q: What has changed more recently?

A: Fed officials have begun discussing the balance sheet more often in their public remarks. Exhibit 2 summarizes their comments so far, including the brief mention of the balance sheet in the December FOMC meeting minutes. None of these comments on its own would be particularly noteworthy. For example, Boston President Rosengren has been discussing options for the balance sheet for some time, and the remarks in the minutes and from Governor Brainard simply restate the committee’s existing guidance—that the time for ending full reinvestments could change as the outlook evolves. However, taken together, recent remarks suggest that officials may be starting to fine-tune their views now that the committee has gotten a couple rate hikes under its belt. They may also be getting ahead of potential criticism from the incoming administration, as some of the economic advisers to President-elect Trump appear to favor a smaller Fed balance sheet with a shorter duration.

Exhibit 2: Officials Addressing Balance Sheet in Public Comments

Source: Federal Reserve, CNBC, Reuters, Bloomberg, Wall Street Journal

Q: Have you revised your forecasts for the Fed’s balance sheet?

A: No, we continue to expect full reinvestments to end in the middle of 2018. First, we are still relatively far from levels of the funds rate that most forecasters see as consistent with the committee’s “well under way” guidance. While policymakers could always signal that reinvestments will end at lower levels for the funds rate, they may view such a change as risky after 2013’s “taper tantrum”. Moreover, for the most part recent public comments do not indicate a shift in views along these lines—Presidents Rosengren and Bullard have proposed genuine alternatives, but other speakers have only highlighted that the balance sheet outlook will depend on the economy, and may need to be discussed this year.

Second, the committee may need to consider how its plans for the balance sheet intersect with the administration’s debt management and regulatory goals (see here for more background). For example, excess reserves created by the expansion of the Fed’s balance sheet serve as a bill-like instrument in the financial system, providing an asset that helps depository institutions meet their regulatory requirements. Shrinking the balance sheet would reduce the stock of excess reserves and therefore the supply of risk-free assets available to investors. The Treasury could theoretically offset this decline by increasing net issuance of bills, but it may be constrained from doing so sustainably because of the debt ceiling, which could lead to sharp cutbacks in bill supply twice this year (see here for details).

Our forecast that reinvestments will continue into 2018 reflects a view that the committee will remain focused on executing the early stages of funds rate normalization this year, and that reinvestments will end a few months after the new Fed Chair takes over. Reinvestment could end sooner if we were to see rapid growth and there appeared to be little risk of returning to the zero lower bound—the conditions President Dudley laid out early last year.

Q: Besides a booming economy, why might full reinvestment end sooner?

A: We can see two main arguments. First, substituting funds rate increases for balance sheet runoff could have smaller effects on the exchange rate (a point noted by some Fed officials and consistent with our research). If policymakers thought that the tightening in financial conditions was becoming imbalanced—with too much dollar strength but still low long-term rates—then they could consider changing the mix of monetary tightening. Second, addressing the balance sheet could ease the transition process to the next Fed Chair. The FOMC may change significantly over the coming 18 months due to open Board positions and expiring terms. As a result, Chair Yellen may want to set in place a framework for shrinking the balance sheet before she steps down.

Q: Could concern about net interest and remittances be driving the recent debate?

A: We highly doubt it. In Q3 2016 (the latest quarter for which data are available), the Fed earned an annualized $109bn in interest on about $4.2 trillion in security holdings, which equates to an effective yield of 2.6%. In order for interest expense to reach the same level, the yield on the Fed’s interest-bearing liabilities would need to rise to around roughly 4.4%. Because about $1.5 trillion of the Fed’s liabilities consist of non-interest-bearing Federal Reserve Notes (i.e. cash currency), increases in the interest on excess reserves (IOER) rate are very unlikely to bring remittances to zero. A recent comment by Federal Reserve Board economists makes this point using simulations of the Fed’s FRB/US model.

Q: Is the FOMC likely to sell assets?

A: Not during Chair Yellen’s term. Several years ago the FOMC indicated that it would begin selling agency MBS “sometime after the first increase in the target for the federal funds rate”, and that sales would aim to eliminate the Fed’s holdings “over a period of three to five years, thereby minimizing the extent to which the SOMA portfolio might affect the allocation of credit across sectors of the economy”. The committee later revised this guidance, saying in June 2013: “a strong majority now expects that the Committee will not sell agency mortgage-backed securities during the process of normalizing monetary policy, although in the longer run, limited sales could be used to reduce or eliminate residual MBS holdings”. These remain the guiding principles around asset sales.

A move back toward asset sales under Chair Yellen would be very surprising. First, ending reinvestments is a lower-risk strategy that the committee would undoubtedly start with initially. Second, history suggests these actions are rare: according to historian Niall Ferguson and coauthors, most central banks never shrink their balance sheets after large expansions but instead hold nominal values constant and allow the balance sheet to shrink as a share of GDP. Professor Ferguson recorded no examples of active sales of long-term debt in his 100+ year sample (although there are a handful of examples of passive balance sheet shrinking, such as Japan in 2006). Third, selling assets could result in realized losses, which would affect remittances (for details see Carpenter, Ihrig, Klee, and Quinn, 2013).

Q: Are there alternative policies that the FOMC might consider?

A: A more active approach to managing the duration of Treasury purchases could be an option. As noted above, the Fed currently purchases Treasuries proportionally, based on public auction amounts for a given day, and without regard to securities’ maturity. As a result, the weighted-average maturity of its purchases can be quite long: in November, for example, the weighted average maturity of the Fed’s auctions purchases was 9.2 years. As an alternative, the committee could decide to actively manage the duration of its assets—one of the ideas offered by President Rosengren. For example, if the Treasury were offering maturities of 3-years, 10-years, and 30-years, the Fed could purchase only the 3-year note, thereby shortening the duration of its assets and quickening the pace of portfolio runoff once it begins. The downside of this approach would be additional debt management challenges for the Treasury: the Fed’s portfolio would mature faster, increasing Treasury’s financing needs over the near term.

Q: How will an end to reinvestments affect the supply of duration to public markets?

A: For mortgages the answer is straightforward: the portion of gross issuance previously absorbed by the Fed will need to be purchased by private sector investors. This will amount to an increase in the supply of mortgage duration held in investor portfolios, which could affect yields and/or spreads.

For Treasuries the outcome will depend on how issuance changes in response to the Fed’s actions. Fed purchases can be thought of a source of financing for the Treasury: every dollar used to purchase securities at auction by the Fed is a dollar that does not need to be borrowed from other investors. When the Fed ends reinvestment, Treasury will need to decide how to raise those additional dollars. If it chooses to increase bill issuance and/or drawdown its cash balance, then there would be little impact on the net supply of Treasury duration. If instead Treasury increases coupon auction sizes, an end to Fed reinvestments would result in an increase in the net supply of Treasury duration.

Under its current auction schedule for coupon-bearing notes and bonds (including TIPS), the Treasury issues $343bn 10-year equivalents per quarter (i.e. an amount of duration equal to $343bn of the on-the-run 10-year note). Over the next 12 months, we expect the Federal Reserve to roll over an average of $58bn in maturing Treasury securities at auction each quarter. If instead the Fed did not roll over its holdings, public coupon auction sizes would have to increase (without an increase in bill supply and/or drawdown in cash balances), but it would be up to the Treasury to decide by exactly how much for each issue. For example, if Treasury officials increased new issuance proportionally (based on current nominal auction sizes), then monthly gross issuance would increase to $386bn 10-year equivalents per quarter. If instead Treasury increased issuance equally across all maturities, then gross issuance would rise to $406bn 10-year equivalents per quarter (Exhibit 3).

Exhibit 3: End of Reinvestments Likely Implies More Duration Supply

Source: Goldman Sachs Global Investment Research

Q: What does all this mean for financial conditions and the stance of monetary policy?

A: It depends on how the committee communicates the end to reinvestments—with the critical distinction being whether portfolio runoff is intended to complement or substitute for funds rate increases. If the committee intends to deliver more monetary restraint—because of strong growth and/or above-target inflation—then an end to reinvestments should be considered a complement to funds rate increases, resulting in tighter financial conditions. However, if the committee is aiming at shifting the mix of policy tightening—because of concerns about dollar strength and/or a change in view about the appropriateness of maintaining a large balance sheet—then the impact on financial conditions could be more benign. One of the important lessons from the Fed’s experience with quantitative easing (QE) has been that signaling channels appear most important. The same probably goes for winding down the balance sheet: the market implications will likely depend primarily on what these steps tell us about policymakers’ broader intentions.