The Big Idea
Looking for QT to taper in July
Stephen Stanley | April 5, 2024
This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors. This material does not constitute research.
Fed officials in recent months have publicly commented on the best tactics for normalizing the balance sheet, and the March FOMC took up the topic in detail for the first time. Chair Powell reported that the consensus favors soon slowing the pace of Treasury redemptions to better manage the transition to the ultimate proper level of the Fed’s balance sheet. However, Powell quickly added that such a decision should not be taken as a sign that the balance sheet will “ultimately shrink by less than it would otherwise,” suggesting that the Fed envisions a longer taper period than financial market participants generally expect. Still, taper looks likely to start in July and run through 2025.
Learning lessons from 2019
The Fed has only used balance sheet expansion as a means to ease policy twice in modern history, in multiple rounds of Quantitative Easing (QE) after the Global Financial Crisis and then again during and after the pandemic. That gives us limited history for gauging the unwind, but the 2017 to 2019 balance sheet reduction offers lessons that loom large in the minds of a number of policymakers.
The original balance sheet expansion ultimately extended from 2009 to 2014. The Fed had not, in modern times, used the balance sheet as an independent policy lever prior to 2008 for two reasons. First, the zero bound on policy rates never became a binding constraint. In addition, until Congress gave the Fed the authority to pay interest on reserves in the fall of 2008, QE was never a viable option because the Fed had to carefully calibrate the level of bank reserves to hit its fed funds rate target. The ability to pay interest on reserves severed that tie and gave the FOMC the option of buying securities in order to add liquidity to the banking system as well as to push asset prices higher.
From 2009 through 2014, the Fed ultimately bought about $3.7 trillion of securities. When the FOMC finally decided that it could safely begin to reduce the balance sheet in 2017, the tricky issue was that there was no easy way to gauge how many reserves the banking system would demand in equilibrium. That magic number was clearly far higher than before 2008, as a new regulatory regime imposed after the financial crisis meant that banks needed (and wanted) far more reserves than in the pre-crisis days (as it turns out, the Fed only reversed about $700 billion of the $3.7 trillion accumulation of securities).
As a result, the Fed took an extremely cautious approach to runoff. At first, the Fed permitted only $6 billion a month of Treasury runoff and $4 billion a month of agency MBS runoff. The pace of redemptions gradually ramped up over time and eventually by late 2018 reached monthly caps of $30 billion for Treasuries and $20 billion for agency MBS.
The Fed had already begun to slow the pace of runoff by the spring of 2019 and at the July 2019 FOMC meeting decided to end balance sheet reduction on August 1. Going forward, all Treasury maturities would be rolled over, and agency MBS redemptions would be rolled over into Treasuries.
Even so, a unique set of circumstances combined to create a problem that the Fed failed to properly manage. Congress and the Trump Administration went through a debt ceiling standoff in the spring and summer of 2019. The debt ceiling suspension expired March 1 and a deal was not struck until late July. During debt ceiling episodes, when Treasury is forced to use extraordinary measures, the cash balance by law has to be drawn down below the levels that Treasury prefers.
This actually makes a critical difference for managing the Fed’s balance sheet because the Treasury cash balance represents a drain on bank reserves. That is, for a consistent level of liquidity provision from the Fed, every dollar increase or decrease in the Treasury’s cash balance subtracts or adds a dollar from the private sector. This could take the form of a tax payment or an increase in Treasury borrowing, either of which must pull resources from someone in the private economy. So, when Treasury ran down its cash balance in the spring and summer by nearly $300 billion, the indirect (and apparently largely ignored) result was an injection of reserves into the financial system of that amount, leaving the illusion of plentiful reserves at a time when the Fed may already have overshot the proper balance sheet reduction.
Then, once the debt ceiling impasse resolved and Treasury normalized its cash balance, liquidity quickly drained from the financial system (Exhibit 1). When the tide receded, it turned out that the financial system did not have as much liquidity as it needed to operate smoothly, and the repo market melted down in September, with the overnight repo rate spiking by as much as 300 bp.
The Fed was forced to open a repo facility that dealers could tap, which they did in late 2019 ultimately to the tune of over $200 billion.
Exhibit 1: Treasury cash balance in 2019
A number of Fed officials took away from that experience that, given the uncertainty regarding the “correct” level of bank reserves and liquidity needed, the Fed should tread carefully in normalizing the balance sheet.
Avoiding past mistakes
The scale of balance sheet expansion during the pandemic dwarfed the 2009-to-2014 period. In just two years, the Fed bought nearly $5 trillion in securities. So, the unwind this time would by necessity also need to be larger, though, again, it is impossible to know with any precision the proper ultimate size of the balance sheet.
There were varying schools of thought regarding how careful the Fed needed to be. Some officials see an excessively large balance sheet as mostly harmless as long as the extra liquidity is being sopped up, for example, by the RRP facility. Others surmise, as I do, that excess liquidity provision by the Fed inevitably permeates the financial system and makes for relatively easy financial conditions. As I argued in a recent piece, there is a strong likelihood that the still-immense amount of excess liquidity sloshing around in the financial system is inflating asset prices and frustrating the Fed’s efforts to moderate the economy through an elevated fed funds rate target range.
The other issue on which opinions range is the downside of an overshoot. Some are laser-focused on avoiding an episode like the events of September 2019, while others believe that the Fed has already taken care of that concern by establishing a standing repo facility. If the Fed’s provision of liquidity becomes too low to meet the demand of the financial system, pressures would naturally send dealers to the standing repo facility and the sort of abrupt meltdown seen in 2019 would presumably be avoided.
Broadly, comments from Fed officials on the balance sheet strategy were rare up until a few months ago, but most policymakers, when they were asked, noted that the reductions had gone smoothly so far and that there was still a long way to go to get the balance sheet to the proper level.
Tactics versus end game
Then, around the turn of the year, Dallas Fed President Lorie Logan asserted that it might make sense to slow the pace of redemptions soon. Financial market participants had long taken the view that the Fed should leave a heightened amount of liquidity in the financial system and took Logan’s remarks, which were also reflected in the December 2023 FOMC minutes, as a sign that the FOMC would be tapering redemptions imminently and ending balance sheet reduction altogether within a few months.
As of late January, when Treasury surveyed primary dealers in advance of the February Refunding announcement, the median expectation for Fed SOMA redemptions of Treasuries was $596 for fiscal year 2024 (October 2023 through September 2024) and $75 billion for fiscal year 2025. Since the Fed was already set to have redeemed well over $200 billion in FY2024 through the end of January, the median estimate implied a total of about $440 billion of further runoff of Treasuries from February 1, with the program ending altogether by the end of this year. Note that with the current pace of runoff of $60 billion a month, by the end of April, the portfolio will already have shrunk by another $180 billion.
To put that projection in perspective, as of late January, the Fed’s securities portfolio was still well above $7 trillion and the liquidity in the financial system, as measured by the sum of the take-up for the RRP facility and bank reserves, was well over $4 trillion, more than $1 trillion higher than plausible estimates of where it should be, as I laid out in a piece earlier this year.
It seemed clear to me at the time that President Logan’s remarks were being misinterpreted. She fleshed out her views in March 1 comments on a paper reviewing the global experience with QT. She noted that it was impossible to identify the proper level of reserves in advance. As a result, she wanted to slow the pace of runoff once the RRP balances “approach a low level.” She noted that significant RRP balances are clear evidence of excessive liquidity provision, but uncertainty begins to ratchet up once that margin dissipates. She consequently advocated slowing the pace of runoff to “reduce the risk of an accident.”
In addition to the quantity of reserves in the banking system, Logan also appealed to the distribution of reserves. In her view, there are frictions in the system that limit the ability of reserves to flow freely from those banks who hold more than they need to those who need more. Slowing the pace of runoff allows the banking system more time to work through those frictions in an orderly manner.
Her conclusion, however, must have been a shock to those in the financial markets who were thinking of a quick end to redemptions: “I want to emphasize that slowing, to me, doesn’t mean stopping. In fact, I believe that proceeding more gradually may allow the Fed to eventually get to a smaller balance sheet by providing banks with more time to adjust.”
In 2019, President Logan was the manager of the SOMA account for the New York Fed. Thus, she has a unique level of expertise on these matters, and it is not surprising that her views appeared to carry the day when the FOMC discussed balance sheet tactics at its latest meeting. The section of Chairman Powell’s press conference statement related to the balance sheet sounds like it could have been taken directly from Logan’s March 1 speech: “the general sense of the Committee is that it will be appropriate to slow the pace of runoff fairly soon, consistent with the plans we previously issued. The decision to slow the pace of runoff does not mean that our balance sheet will ultimately shrink by less than it would otherwise but rather allows us to approach that ultimate level more gradually. In particular, slowing the pace of runoff will help ensure a smooth transition, reducing the possibility that money markets experience stress and thereby facilitating the ongoing decline in our securities holdings consistent with reaching the appropriate level of ample reserves.”
Recalibrating my estimates
The FOMC is clearly pursuing a somewhat different set of tactics than I had assumed back in January when I last visited this topic. At that time, I presumed that, with the ultimate target still clearly very far away, the Fed would continue balance sheet reduction for all of 2024 at the current pace and then gradually taper through 2025.
My assessment of where the Fed balance sheet needs to eventually settle has not changed substantially, but I now have a better sense of the preferred path for getting there, as laid out by President Logan. As a result, I have altered my projected path for redemptions going forward in light of the result of the March FOMC meeting discussion.
Previously, I had the Fed holding the monthly pace of Treasury redemptions at $60 billion through the end of this year, then tapering gradually over the course of 2025, with runoff ceasing at the end of 2025.
Now, I am assuming that the FOMC decides at the June FOMC meeting to begin tapering as of July 1. I have penciled in monthly caps of $45 billion for the third quarter, $35 billion for the fourth quarter, and a continued slow taper through 2025, as before, ending redemptions at the end of 2025.
I expect the pace of MBS runoff to be unaffected for the foreseeable future. The current monthly cap is $35 billion, but actual runoff is running around $15 billion per month, and I look for that pace to continue through the balance of 2024 and much of 2025.
On a technical note, one other balance sheet development that has occurred since January is the expiration of the Bank Term Funding Program in March. At the end of January, the take-up for this program was $165 billion, and it has already declined to about $130 billion. By next March, it will move to zero, subtracting that amount from the asset side of the Fed’s balance sheet, a somewhat quicker pace of reduction than I had implicitly assumed back in January.
In any case, my new path for the runoff of Treasuries implies about $270 billion less in redemptions from July 1 through the end of 2025 than before, partially offset by the quicker rollup of the BTFP. In all, I would expect redemptions of Treasuries from July 1 through the end of 2025 to total about $450 billion, while MBS runoff over that same period may be about $250 billion. Coupled with the end of the BTFP, that would imply that the Fed’s balance sheet by the end of 2025 would be smaller by about $850 billion. If that reduction translates one-to-one to the sum of the RRP facility and bank reserves, then that would imply a combined level still somewhat in excess of $3 trillion, which would be close to the high end of the plausible range that I framed up in the January piece. As President Logan noted, no one knows with any precision where the balance sheet is supposed to settle out, but, especially with the “longer runway” approach that the FOMC has chosen, balance sheet reduction may plausibly extend through all of 2025 and I would not be shocked if it even continues into early 2026.