February 09, 2024 | 12:43
QT, Quo Vadis?
QT, Quo Vadis?
In the press conference after the recent FOMC confab, Chair Powell indicated quantitative tightening (QT) will be discussed at the next meeting. “That’s what we’ll be talking about at the March meeting. A whole range of issues will be briefed up, and the Committee will get into all of the issues that will be arising over the course of the next, let’s say, year or so.” The biggest issue is when to end QT and then when to start tapering its pace.
The current QT campaign started on June 1, 2022, to reinforce the rate hikes that started almost three months earlier. Monthly caps were established, above which maturing Treasury securities and MBS principal payments would still be reinvested, but below which these amounts would roll off the Fed’s balance sheet. The caps were initially set at $30 billion for Treasury coupons (notes, bonds, and TIPS) and $17.5 billon for MBS. They were doubled to $60 billion and $35 billion, respectively, as of September 1, and have stayed there since. Note that for months in which maturing coupons fall below the cap, Treasury bills are not being rolled over to make up most of the shortfall.
The Fed’s balance sheet peaked during April 2022, with total assets at $8.97 trillion and securities holdings amounting to $8.50 trillion (Chart 1). Recall the quantitative easing (QE) program that started at the pandemic’s onset was still being wound down in early March 2022, just days before rate hikes began. The balance sheet has since shrunk to $7.63 trillion (down $1.33 trillion) with securities holdings at $7.11 trillion (down $1.39 trillion) .
A look back at the previous QT episode (2017-2019) provides some perspective on the current case. Back then, the caps were initially set at $6 billion for Treasuries and $4 billion for MBS and they were raised by these amounts each quarter until reaching $30 billion and $20 billion, respectively, after a year. This time around it took just three months and one iteration to get to the much-larger caps, indicating the Fed has grown more comfortable making abrupt balance sheet maneuvers (as the preceding QE program also showed). Current QT also started less than three months after QE ended, vs. a three-year interval in the first episode.
The previous QT program started in October 2017, with the full caps in place as of October 2018. In May 2019, the cap for Treasuries was halved to $15 billion with QT scheduled to end by October 2019. The MBS cap remained, although effective October, the up-to-$20 billion monthly amount would be reinvested in Treasuries. This reflects the Fed’s goal to hold only Treasury securities over time. In July 2019, it was announced that QT would end by August, two months early. The reason for the premature end, along with how the contours of QT1 were crafted in the first place, mirrored what was happening on the other side of the Fed’s balance sheet. This has been and will continue to be the case for QT2.
Balance Sheet Management
The Fed can control the size of its balance sheet, both total assets and total liabilities (including capital), but it can’t control the mix of its liabilities (Table 1). As stated in May 2022’s QT gameplan: “Over time, the Committee intends to maintain securities holdings in amounts needed to implement monetary policy efficiently and effectively in its ample reserves regime. To ensure a smooth transition, the Committee intends to slow and then stop the decline in the size of the balance sheet when reserve balances are somewhat above the level it judges to be consistent with ample reserves.”
The “somewhat above” criterion reflects the uncertainty surrounding banks’ equilibrium demand for reserves in an ample reserves regime. Each bank is different, reflecting factors such as liquidity and risk management policies along with arbitrage considerations (borrowing more cheaply than the interest rate paid on reserve balances and then parking the funds at the Fed). The criterion also reflects the volatility in items such as reverse repos and the Treasury General Account (more on these later).
Interestingly, this criterion was not part of June 2017’s QT gameplan. As QT1 unfolded, reserves continued to drift lower (Chart 2). By March 2019, the Fed was getting the sense where ‘equilibrium’ reserves could be and started tapering QT in May with an eye to ending it in October.
However, lack-of-liquidity pressure owing to a potential ‘scarcity’ of reserves was already starting to mount. A key barometer such as the secured overnight funding rate (SOFR) was averaging more regularly above the midpoint of the fed funds target range (Chart 3). Recall, fed funds involves unsecured lending, so it should be more expensive than secured lending. The early ending of QT turned out to be too late. During September 2019, SOFR spiked as high as 5.25% compelling the Fed to augment market liquidity with repos and T-bill purchases. The “somewhat above” criterion is designed to prevent this from reoccurring.
The Repo Factor
Since 2021, another factor has entered the Fed’s QT calculus, activity in the overnight reverse repo (ON RRP) facility. This is mostly money market funds (MMFs) investing in overnight repo (which are reverse repos from the Fed’s perspective). These institutional investors are ‘buying’ Treasury securities from the Fed and then ‘selling’ them back the next day to invest their cash on an overnight basis. Typically, these aggregate transactions were small as recently as early 2021, but they surged to average above $1 trillion by August and $2 trillion by June 2022 (Chart 2 again). They peaked at $2.4 trillion to start 2023 but have since plummeted to average under $530 billion.
The catalyst for the surge was the reduction in T-bills that had originally financed the massive buildup in Treasury’s cash account at the Fed (the Treasury General Account or TGA). The TGA rocketed from averaging under $400 billion in March 2020 to above $1.8 trillion by July to pay for the federal government’s massive pandemic relief measures. These measures left a legacy of excess savings that found their way into bank deposits and MMF balances, with the latter typically parked in repos (with banks) and T-bills. As the government disbursed its funds and the TGA was drawn down, corresponding T-bills were reduced. There were now less T-bills for MMFs to invest in while banks, already flush with deposits, were less interested in conducting MMF repos. MMFs then turned to the ON RRP facility.
Interestingly, the debt ceiling dramas during the autumn of 2021 and spring of 2023, which caused the TGA to be depleted to its bare minimum and corresponding T-bills to be reduced further, also augmented reverse repo activity. This emphasizes how potential volatility in the TGA will be a consideration for when to end QT. The “somewhat above” criterion is partly designed to absorb this volatility.
With respect to the $1.8 trillion plummet in reverse repos from their peak, as the Fed raised policy rates by 525 bps between March 2022 and July 2023, money market investors have presumably found more lucrative places to park their cash. We reckon this trend will continue with aggregate transactions returning to their earlier lesser amounts. Some concern has been raised that after a near three-year run of elevated ON RRP activity, a return to smaller amounts could undermine banking system/money market liquidity, potentially expediting the timetable for tapering and ending QT. However, with $3.6 trillion of reserves remaining, we judge this shouldn’t be an issue.
Where We Stand
The combination of reserves and reverse repos is a measure of how much liquidity the Fed is maintaining in the banking system. It currently stands at $4.1 trillion, steadily trending down from the recent high of $5.7 trillion hit last March as falling reverse repos have more than offset mildly rising reserves (Chart 2 yet again). So, what is the level consistent with ample reserves, along with a cushion?
Given that ‘ample’ will grow over time, say, alongside broader economic growth, one way to gauge liquidity is relative to nominal GDP (Chart 4). In 2023Q4, reserves and reverse repos averaged 15.1% of GDP. We know that 6.9% in 2019Q3 was way too low. And it’s safe to assume that the first post-QE peak of 16.2% in 2014Q3 was too high (let alone the second post-QE peak of 22.9% in 2021Q4). Interestingly, the median level over the full QE era is 11.9%, which seems like a reasonable first pass at a figure. This results in a $3.3 trillion liquidity target for ending QT as of 2023Q4. Although this figure would grow in line with GDP, we’ll still use it as our working assumption for now given that the Fed has the Standing Repo Facility (SRF) at its disposal. This was introduced in 2021 to address pressures in the repo market of the sort we saw during 2019. So, other things equal, the Fed could end QT with a smaller above-ample cushion than would be the case in the absence of the SRF.
This means there’s about $800 billion in QT still to go. If we assume reverse repos move to their earlier lesser amounts, accounting for more than $500 billion of liquidity drain, this puts the reserves target at about $3.3 trillion. At the current ‘full-cap’ pace, it would take about nine months to achieve this. But allowing for tapering and continued sub-cap MBS payments, QT should continue into 2025. We assume tapering will start when rate cuts begin July, with the Treasury coupon cap cut in half to $30 billion and the MBS cap staying at $35 billion (but the actual roll-off running around $20 billion). QT should end by the spring of 2025, with the MBS cap remaining in place and any amounts below it being reinvested in Treasuries.
Importantly we view this taper timing partly as a token move; the Fed could technically wait until closer to QT’s end before tapering. However, given the current QT pace was advertised as reinforcing rate hikes, a reduction when rate cuts start would help the Fed in portraying continued QT as a balance sheet normalization strategy instead of an effort to potentially temper the impact of rate cuts.
 There are two things to note. First, the reduction in securities holdings has been less than the total amount of ‘cap room’. As of last month, the latter amounted to $1.76 trillion. With T-bill roll-off making up most of the shortfall of Treasury coupon maturities, the culprit is the perennial shortfall of MBS principal payments below their cap. In the wake of surging mortgage rates (from record lows to two-decade highs), and amid still lofty prices, home sales recently hit their lowest volumes in more than 12 years, So, much fewer mortgages are being paid off as folks sell and originated as they buy. Meanwhile, refinancing became very unattractive. By the end of 2022, the volume of mortgage applications for refinancing hit its lowest level since mid-2000, remaining in a relatively low range since.
Second, on the balance sheet and excluding securities holdings, remaining assets have risen over the QT interval. This solely reflects the impact of the Bank Term Funding Program (BTFP), which was introduced amid the regional banking crisis last March. Take-up quickly pushed above the $100 billion mark by early June before stabilizing. In recent months, however, activity picked up again, to around $165 billion, as some banks were taking advantage of the program’s relatively attractive funding cost (which the Fed has since increased). The BTFP will end next month with loans lasting for up to one year. [^]