Shrinking the Balance Sheet: Where Fed Officials Stand

Balance Sheet Normalization, FOMC, Interest Rates, Janet Yellen, IOER
Author's own creation.

In March, the Federal Open Market Committee (FOMC) signaled it could begin shrinking the Fed’s balance sheet sometime later this year. However, with limited official details about what that means and none forthcoming from last week’s FOMC press release, many questions remain:

  • How will the Fed decide exactly when to begin shrinking its balance sheet, and will the move be data or date dependent?
  • Once the wind-down begins, how rapidly will the balance sheet shrink and to what new normal level?
  • How will the Fed dispose of its assets: by simply refraining from reinvesting the proceeds from maturing securities, passively shrinkage of the balance sheet, or by actively disposing of some assets to ensure a smoother path for balance sheet reduction?
  • And would asset sales, should they occur, include both mortgage-backed securities (MBS) and Treasuries or would the Fed initially focus on a single asset class?

Back in September 2014, the FOMC released its Policy Normalization Principles and Plans (henceforth “the Framework”), its official statement outlining a three-step normalization strategy, including balance sheet reduction. First, the Fed would raise policy rates[1] to “normal levels.” Second, the Fed would begin to shrink the balance sheet in a “gradual and predictable manner” by ending the reinvestment policy. And third, the wind down would continue until the Fed holds only enough securities to conduct monetary policy “efficiently and effectively” with a portfolio consisting primarily of Treasuries. There is, of course, a caveat that the Fed can deviate from the Framework as economic conditions change. Since December 2015 the Fed has raised policy rates three times, but it has yet to update the Framework to provide further details on the next steps for balance sheet normalization.

With only the broad principles in the Framework as yet available, more detailed information must be gleaned from elsewhere. Fortunately, nearly every Federal Reserve official has discussed the balance sheet to some extent recently; but while their attention may be uniform, the policy discussion is not. Some officials have said nothing beyond the Framework, while others, particularly those regional bank presidents that do not vote on this year’s FOMC, have offered additional comments about the timing, speed, and ultimate target size associated with reducing the balance sheet. This essay examines the views of FOMC permanent voters first, then regional Fed presidents voting in 2017, followed by non-voting regional presidents.

Balance Sheet Normalization, FOMC, Interest Rates, Janet Yellen, IOER


Federal Reserve Chair Janet Yellen

Chair Yellen has said very little beyond the Framework, and, as the leader of the Fed, keeping to the official talking points is no surprise. In a March speech, Yellen reiterated that the balance sheet would remain elevated until “sometime after” rates rise, though she declined to add specific benchmarks. When asked for additional clarity during the March FOMC press conference she said only that shrinking the balance sheet is not predicated on a pre-specified level for the federal funds rate and that overall monetary policy normalization would be “well under way” before shrinking the balance sheet commenced. George Selgin did not think much of her remarks.

New York Federal Reserve Bank President & FOMC Vice Chairman William Dudley

Dudley, a dove who is a close ally of Chair Yellen, gave a slight preview of the Framework in a May 2014 speech indicating that he wanted to see rates quite a bit higher before the cessation of reinvestments. This was a break from the 2011 Exit Strategy Principles that had called for ending reinvestments first and raising rates secondarily. In that talk, Dudley downplayed the potential adverse consequences of the larger for longer balance sheet approach, believing it prudent to tolerate those risks as the Fed moved off the zero lower bound. Dudley’s preferred order, to raise rates before touching the balance sheet, is, of course, the order now in the Framework.

More recently Dudley discussed balance sheet actions beyond the Framework. In  March he mentioned that shrinking the balance sheet and raising interest rates are, “…two different, yet related, ways of removing monetary policy accommodation.” Because ending reinvestments could act similarly to a rate hike, Dudley cautioned, “…when we begin to end reinvestment, we will have to consider the implications for the appropriate short-term interest rate trajectory.” He has also commented on the mechanics of how to shrink the balance sheet, saying he does not see “a strong need to differentiate between mortgages and Treasuries” as the reinvestment policy ends, which he believes might end this year or in early 2018. Nonetheless, the New York Fed’s trading desk has conducted very small MBS sales to test the operational readiness of such transactions.

Federal Reserve Vice Chairman Stanley Fischer

In a February 2016 speech, Fischer said that because the federal funds rate is now adjusted using two new tools, interest on excess reserves (IOER) and overnight reverse repurchases (ON RRP), the Fed can change the size of the balance sheet independently from interest rate policy. At that time, Fischer saw benefits to maintaining a larger balance sheet, remarking that when to “…begin phasing out reinvestment will depend on how economic and financial conditions and the economic outlook evolve.”

In November, Fischer reiterated the Framework position, saying that shrinking the balance sheet would commence when “…the short-term interest rate approaches more normal levels.” However, he also offered a position different from Dudley’s, explicitly stating that the Fed would begin by ending reinvestments on mortgage-backed securities while continuing to roll over Treasuries. Just last month, Fischer said he does not expect significant market disturbances, such as another taper tantrum when reinvestments end, given the muted market responses to Fed officials’ discussions about shrinking the balance sheet, thus far.

Federal Reserve Governor Jerome Powell

Powell said in a recent interview that he wants the Fed “well into the normalization process” before the balance sheet begins to shrink. With rates far from zero, “removing accommodation” by ending reinvestments would then proceed in “a very predictable almost automatic way.”

Federal Reserve Governor Lael Brainard

Though Brainard is widely considered to be the most dovish Federal Reserve official, she voted with the rest of her colleagues to raise interest rates at the March FOMC meeting. She has also signaled a willingness to increase the speed of rate increases provided the new administration makes good on its campaign pledges of expansionary fiscal policy.

Brainard offered more details about the normalization strategy than her colleagues on the Board when she identified two available strategies in a recent speech. The first is the complementarity strategy, in which balance sheet adjustments would be viewed as an independent and thus second tool for conducting monetary policy. As Brainard says, “Under this strategy, both tools would be actively used to help achieve the Committee's goals…to take advantage of the ways in which the balance sheet might affect certain aspects of the economy or financial markets differently than the short-term rate.” The Fed might deploy the balance sheet to affect term premiums on longer-term securities and use the policy rates to affect money markets. The second option is the subordination strategy, in which the policy rates would remain the primary tool for the Fed’s conduct of monetary policy. Once normalization of short term rates was “well under way” the balance sheet could begin to shrink in a “gradual [and] predictable way.” When reinvestments end, the balance sheet would then shrink on “autopilot.”

Brainard is an advocate of the subordination strategy and supports the automatic process that Powell discussed, though she does maintain that were the economy to be hit with a large adverse shock restarting reinvestments could be prudent in order to preserve traditional policy space in the federal funds rate.


Minneapolis Federal Reserve Bank President Neel Kashkari

Kashkari made national headlines when he posted an essay explaining his dissent at the March FOMC meeting, where all his colleagues voted for a rate increase. In dissenting, he noted that a 2% inflation target was no reason to raise rates as though 2% was a ceiling . His preferred strategy was for the Fed to publish a detailed plan for shrinking its balance sheet, allow some time to gauge the market reaction, and then continue to use short term rates as the primary policy lever. Kashkari supports Brainard’s subordination strategy when he says, “…we can return to using the federal funds rate as our primary policy tool, with the balance sheet normalization under way in the background.”

Philadelphia Federal Reserve Bank President Patrick Harker

Harker, an engineer by training, has been more precise than his colleagues. In January he said, “When we are at or above 100 basis points — and we are moving toward that — I think it is time to start serious consideration of first stopping reinvestment and then over a period of time unwinding the balance sheet.” In March, Harker said that the right number for interest rates could be 1.5%, but that balance sheet reduction is not going to be dependent on a trigger or a target and that it will also depend on the “momentum” of the economy — a position similar to Chair Yellen’s at the March FOMC press conference. Harker does prefer the “Treasury-heavy” portfolio called for in the Framework, though he is not sure that the Fed should completely get out of the MBS market.

Chicago Federal Reserve Bank President Charles Evans

Evans, who originally gained national prominence when the Fed began to employ Forward Guidance, is one of the more dovish members, believing that only two hikes in 2017 are possible, while, by contrast, Eric Rosengren of Boston is predicting four. When it comes to shrinking the balance sheet, though, Evans is one of the few to comment, not on the timing, but on a new target size. Recently, he estimated a target size for the balance sheet of $1-1.5 trillion, requiring as much as $3 trillion of securities to roll off. That is drastically different from former Federal Reserve Chairman Ben Bernanke’s estimate for a new normal balance sheet of $2.5-$4 trillion. Despite the potential reduction, Evans has yet to say when reinvestments might actually end.

Dallas Federal Reserve Bank President Robert Steven Kaplan

Kaplan has become more vocal on balance sheet action throughout this year. In January, he said that 2017 would be a good year to discuss a “plan of action” to “slim” the balance sheet, but that nothing should actually be done until rates hikes were “further along.” Kaplan echoed those sentiments in February: “…as we make further progress in removing accommodation, I believe we should be turning our attention to a discussion of how we might begin the process of reducing the size of the Federal Reserve balance sheet.”

After the March rates hike, Kaplan went even further, saying that as rates rise the Fed should publish a plan to shrink the balance sheet. He added that he does not want balance sheet normalization to “unduly affect” financial market conditions, suggesting that securities rolling off ought to be kept to a percentage of daily trading volumes in MBS and Treasuries. Such a strategy would require more active management of the balance sheet than the autopilot strategy proposed by Brainard and Powell. For Kaplan, one of the most important considerations as the balance sheet begins to shrink is to “minimize disruption” to markets.


As mentioned, the most varied opinions about the next move for the balance sheet come from the regional bank presidents who do not have a vote on the FOMC in 2017.

St. Louis Federal Reserve Bank President James Bullard

Bullard is known to be the low dot on the “dot plot,” as he believes the economy is stuck in a low rate regime likely to persist for years. He differs from many of his colleagues in other important ways. For example, Bullard believes that the policy rates are currently at the appropriate levels and that the Fed has, “…delayed a little bit too long in reducing the size of the balance sheet.” While he doesn’t necessarily oppose another hike this year, Bullard thinks the FOMC’s priority should be reducing the balance sheet in an effort to increase the Fed’s ability to react to the next downturn.

San Francisco Federal Reserve Bank President John Williams

Recently, Williams offered perhaps the most comprehensive assessment of the future of the Fed’s balance sheet, with a call for the reinvestment policy to end this year. Like Evans, Williams offered a target, saying that a balance sheet around $2 trillion is likely appropriate, though added that no decision had been made. But, unlike Evans, Williams also offered a timeframe, remarking that getting to a balance sheet that size would likely take 5 years. Williams also believes that with the policy rate and the balance sheet moving contemporaneously, the path of each one will be slower than if they were operating alone, similar to Dudley. He thinks the Fed will raise rates twice more this year, though leaves open the possibility for a third hike if the data support it — a position held by his colleague in Boston.

Boston Federal Reserve Bank President Eric Rosengren

Rosengren is now one of the leading hawks, having announced in a recent speech that he anticipates three more rate hikes this year, likely at every other FOMC meeting. While Dudley and Williams believe shrinking the balance sheet might slow rates hikes, all else equal, and Bullard thinks balance sheet reduction can replace a rates increase, Rosengren believes the path for rate increases is not affected much by gradually shrinking the balance sheet and that the process can begin soon. As identified by Ben Bernanke, Rosengren also differs from his colleagues by being the very rare Fed official to discuss asset sales — though he stopped short of actually advocating them in the speech. However, Rosengren also thinks it is likely that the Fed would resume asset purchases during future recessions, “…unless they are very, very mild.”

Kansas City Federal Reserve Bank President Esther George

George is the most hawkish member on the FOMC, having said that the Fed was behind the curve in raising rates in December 2015, repeatedly voting to trim assest purchases during QE3, and having far and away the most dissenting FOMC votes — now that Jeffrey Lacker has stepped down. And yet, at a recent event George indicated that she did not think that any decision regarding the balance sheet would be made soon. She wants the Fed to spend more time analyzing its path toward normalization, stating that in the meantime the size of the balance sheet is not likely to change. This is a change from her position back in 2014, when she thought it was appropriate to begin shrinking the balance sheet via “passive runoff” before the first rate hike, following the policy articulated in the original 2011 Exit Strategy Principles.

Cleveland Federal Reserve Bank President Loretta Mester  

In three recent speeches Mester has shown an increasing comfort level with shrinking the balance sheet this year. She wants to end reinvestments in 2017 and believes this move is consistent with the Framework, putting the Fed on a path towards a balance sheet consisting primarily of Treasuries. And just yesterday, Mester supported her colleagues’ notion to announce a plan for balance sheet reduction, which will take “several years,” as well as a return to using the federal funds rate as the “main tool” for monetary policy. Mester added that the balance sheet will eventually be “considerably smaller than it is today.”


How Federal Reserve officials view the balance sheet will change as new data come in. There are also potential shifts at the Fed via new personnel. With the retirement of Governor Tarullo in April, President Trump can appoint three new Fed Governors. Additionally, Rafael Bostic will assume leadership of the Atlanta Fed in early June and sit on the FOMC next year while the Richmond Fed is continuing its search for Jeffrey Lacker's successor, who will have a vote in 2018 as well.

Whoever comes to the Fed and however the views of those already there change, the important questions about the balance sheet will remain. These questions can be grouped into four buckets: Timing, Mechanics, Interest Rates and the Endgame.

On timing, the most important question is when the reinvestment policy ends. There is a growing chorus suggesting that 2017 will see the end of the reinvestment policy, as laid out in the Framework. However, many officials condition their balance sheet remarks as data dependent. It is unknown how much the data would need to soften to move a Fed official’s view away from Mester’s position and towards George’s.

The mechanics of the balance sheet wind down are extremely uncertain. Will the Federal Reserve simply allow for passive shrinking when securities mature, or will they actively manage the process and shrink the balance sheet on a smoother path, perhaps limited by trading volume ratios as Kaplan suggested? These questions require clear answers in the kind of public, detailed plan called for by Kashkari and Kaplan. Another mechanical issue to address is distinguishing between Treasuries and other securities. Is that distinction less important, as Dudley has implied, or will the Fed start by paring back its MBS holdings, as Fischer has suggested?

Related to the mechanics is how shrinking the balance sheet will affect the path of interest rates. Will the Fed adopt the subordination strategy advocated by Brainard? Or will the balance sheet runoff tighten financial market conditions such that the paths for rates hikes and shrinking the balance sheet could be slower together, as Dudley and Williams have considered? Or could ending reinvestments be a substitute for a rates hike, as Bullard prefers?

And lastly, what is the Fed’s endgame when it comes to balance sheet normalization; what is the proper size? Many Fed officials have noted an elevated demand for currency, compared to what existed before the crisis, but only a few have offered specifics as to the balance sheet’s final size. Will the balance sheet stay quite large, something Ben Bernanke advocates, or will it pare down to $2 trillion, as Williams suggests, or even beyond that to $1.5 trillion, as Evans estimates?

As most officials concede, the Federal Reserve is about to take actions with which it has virtually no experience. Providing further details on how and when they will normalize the balance sheet would go a long way to reducing uncertainty. But even then, it will remain critical to track where Fed officials stand on this issue and how those views evolve with the data.

[1] The Framework discusses, “…steps to raise the federal funds rate and other short-term interest rates to more normal levels…” That language, however, is ambiguous as the federal funds market has shrunk dramatically in a financial system awash in reserves. Consequently, interest rate policy is now conducted using two new policy rates to create a federal funds rate target “range:” the interest paid on excess reserves (IOER) creates the target ceiling while the overnight reverse repurchase (ON RRP) rate creates the target floor. Both rates are set administratively by the Fed. For further reading on the Fed’s new monetary control mechanism using IOER and ON-RRP for a federal funds rate range, see “A Monetary Policy Primer, Part 9: Monetary Control, Now” by George Selgin.


  1. Really good analysis by Mr. Lacey. My preference would be a balance sheet in the $2 trillion range, other factors being considered. For example, currency outstanding regularly is $1.5 trillion plus now, and the Fed holds "only" about $2.4 trillion of Treasuries with which to cover Treasuries. Minimum operating amount probably about $1.8 trillion in the current environment. I would sell every last mortgage-backed security if I could. The Fed never should have bought them in the first place and had to engage in high-level legal weaselry to find the authority to do so after Fannie and Freddie were placed into conservatorship in the summer of 2008. The half-life of mortgage-backed securities is about seven years, I am told. If the Fed had begun even in 2011, well past the "need" for crisis-level purchases of MBS, to allow its existing portfolio of MBS to roll off as they matured, the $1.7 trillion plus of MBS holdings already would have shrunk by the better part of $1 trillion. Also, the mortgage market itself would be healthier, in the sense that it would not be sustained artificially by the Fed's continued purchases. Right now is already rather late for starting the non-rollover process of MBS. I would hold onto the Treasuries until the MBS are all gone, but I would allow any Treasury maturing in more than five years to roll off also, which would make everyone's life a little easier (but might cost the Treasury a little more in interest payments). Remember in any case that, on a balance sheet-to-GDP ratio, the Fed since mid-2008 created enough new reserves to support an economy four times as large. We are only about 50 percent larger than we were then; the rest of that expansion capacity is tied up in various financial assets, not the least of which are excess reserves. I think I'm right that the "normal" range of Fed balance sheet vs. GDP ratio is between 6 and 8 percent (about 6.8 percent in mid-2008). Currently we run about 23 percent, and we have been as high as 25 percent. If the Fed tried to shrink even to 10 percent of GDP, the balance sheet would come in at around $1.9 trillion. Just so everyone would understand how truly bloated and bubbled we really are as the Fed allegedly tries to start the shrinkage process. There are plenty of other artificially induced financial market distortions to worry about, but the root of nearly all of them would be addressed if the Fed finally, really, truly, and deeply, began to shrink its balance sheet, MBS first. — Walker F. Todd, Middle Tennessee State University, American Institute for Economic Research (Trustee), and resident of Chagrin Falls, Ohio May 9, 2017

    1. I agree the analysis is good, and your comment is very interesting as well. Unfortunately, evidence largely shows that absent hyperinflation, money is neutral, short-term and long-term, so the ratio of Fed balance sheet to GDP is irrelevant. It is as if a large gold or silver hoard was found (think: Spanish discovery of the New World, South African find, CA 49'ers, or Klondike): aside from the seigneurage that the precious metals issuing entity enjoys (and leaving aside the long-term deleterious effects like Dutch disease), prices adjust to the 'new money' and there's really no change in real output in an economy.

      1. "Unfortunately, evidence largely shows that absent hyperinflation, money is neutral, short-term and long-term." You keep saying this, but the vast majority of monetary economists don't believe it, and the mass of evidence doesn't support it. The high correlation of nominal and real GDP over shorter time horizons is itself difficult to square with it, as is the short-run inverse correlation between nominal interest rates and (especially unanticipated) central bank asset-purchases. These are but tiny tips of a very large empirical iceberg.

        1. Do I dare engage in a battle of wits with the long-armed George Selgin? Yes I do.

          Evidence of short term money neutrality:

          [1] Blanchard's econ textbook cites: Lawrence Chambers, Martin Eichenbaum and Charles Evans, “The Effects of Monetary Policy Shocks: Evidence from the Flow of Funds”, Review of Economics and Statistics, Feb. 1996, 78-1. – a mere 60% confidence interval (i.e., very wide band) from what the Fed does and what the economy does, showing the Fed is largely impotent

          [2] "Measuring the Effects of Monetary Policy: A Factor-Augmented Vector Autoregressive (FAVAR) Approach * Ben S. Bernanke et al (2003) “Apart from the interest rates and the exchange rate, the contribution of the [Fed] policy shock is between 3.2% and 13.2%. This suggests a relatively small but still non-trivial effect of the monetary policy shock. In particular, the policy shock explains 13.2%, 12.9% and 12.6% of capacity utilization, new orders and unemployment respectively, and 7.6% of industrial production” (page 24) Bernanke et al paper: “the data span the period from January 1959 through August 2001.”

          So based on refs [1] and [2], it can be said that though Fed monetarism is 'statistically significant' short term, and does have an effect, the effect is small (a mere 60% confidence interval) and results in a mere 3.2% to 13.2% effect out of 100% on a range of variables. Perhaps for this reason people really don't care that much about nominal price stability and are not clamoring for a return to a gold (or hard money) standard?

          As to what mainstream economists believe, that can change; by analogy I remind you George that just yesterday a meta-study found that contrary to popular erroneous belief, eating cheese, even full-fat cheese, is not per se bad for your heart. Blessed be the cheesemakers. Unfortunately, in economics the thinking of 'don't fight the Fed' is so common I doubt anybody can do an objective meta-study, as there aren't enough papers taking the position that the Fed doesn't matter. Even Bernanke's paper barely got press and I couldn't even find a final draft (the above is from the first draft). Traditions die hard; even Copernicus after showing the earth moved around the sun believed in 'heavenly spheres' that moved around the sun, in perfect orbits, not unlike Ptolemy.


          1. From emphatic declaration that money "is" neutral in the short-run to a couple papers, one unpublished, showing that the short-run effect estimates are subject to wide errors, modest, but "non-trivial." Hardly an impressive counterargument!

            Of course traditions die hard even when the evidence isn't consistent with them; but so long as the evidence continues to support them, they do not die at all–nor should they!

          2. What can I say George, but "you win again"!? BTW, the confidence interval paper is a bit weak to prove anything, since the graph jumped around a bit (in and of itself a confidence interval says nothing, and is not to be confused with a confidence level) so I guess my primary evidence is an unpublished paper by Bernanke and my observation that even diehard new or otherwise monetarists, like Scott Sumner, agree that 'long term' money is neutral. But they are vague in what "long term" means: months, weeks, days, hours? Finally, if monetarism is so important, why don't we have a neo-William Jennings Bryant on the political scene? Is it really lack of education by the voting public, or simply is hard money an idea whose time has come and gone?

  2. I had to look up money neutrality:
    The neutrality of money, also called neutral money, says changes in the money supply only affect nominal variables and not real variables. In other words, an increase or decrease in the money supply can change the price level but not the output or structure of the economy.

    In my model, the currency bankers notices that savings to loans are out of variance and executes an immediate asynchronous, adjusted interest swap. (no time bets) . Winners are those who are close to the new match between savings and loans.

    Does that change output? Well, it was a savings to loan imbalance that only a few bankers noticed, then suddenly the interest swaps, and everyone knows. New monetized information is a change in output. Who changes the output? The original bankers with the inside information. I thought currency baking was all about finding the output structure. Dunno who gets the credit for changing output, but it happens.

    1. I should explain

      If a member bank takes out a three month loan, what does the central banker know? In all likelihood, the member bank is lying, it plans to finish up in two months, and return the money. early and the central bank got snookered. Time cheating makes the mathematics impossible.

    2. For every winner, there is a loser, so output (net) does not change…money is neutral. Even die-hard monetarists believe that, except they will say that money is only neutral long term. In the short term, claim these guys, on scant evidence I might add, there is "money illusion" (information is imperfect, people don't understand inflation) and "sticky prices" (especially wages, though it seems in the real world, absent labor unions, sticky wages are rather hard to find). As Bernanke says in his 2002 FAVAR paper (see my other comment), money non-neutrality is only 3.2% to 13.2% out of 100%, no big deal, albeit 'statistically significant'.

      1. "For every winner, there is a loser, so output (net) does not change…money is neutral." Can't you say something that's true now and then?

        1. It's very truthy George. Keep in mind exchanging something for value is not net zero since both the buyer and seller have different marginal utilities where a sale leaves both of them better off. The same is not true with moving money around and relying on 'money illusion' and other forms of sleight of hand to increase the economic pie. No free lunch…

          1. I like "truthy." Still, even if one accepts the claim that monetary expansion never adds value (I don't), the very fact that there are winners and losers implies…non-neutrality! For if money were strictly neutral its creation would have neither distribution nor output effects.

          2. Coase Theorem, George? Sigh, one of these days I'll win a debate vs Dr. Selgin.

Comments are closed.