US reverse repo volumes and bank reserves to fall in 2023
The way to think about the Fed’s balance sheet in round numbers is to start with its current size at almost $9trn. Of that, there is $2.25trn showing up at the reverse repo facility, $3trn in bank reserves, and most of the rest is cash in circulation (apart from other bits and pieces). So what comes out of reserves has been going into the reverse repo facility. And as the Fed’s balance sheet falls in size through bonds rolling off the front end (soft quantitative tightening), there must be a corresponding fall in bank reserves and/or in usage of the reverse repo facility.
How much balance sheet roll-off is required for better balance? We think US$2trn
The question then is how much balance sheet roll-off is required in order to bring about a sense of equilibrium between collateral and liquidity. A measure of this need is the $2trn of liquidity that routinely gets shovelled back to the Fed in its overnight reverse repo facility. The large use of this facility is reflective of an ongoing liquidity overflow that manifests in market repo rates struggling to match the rate being offered by the Fed (at 5bp above the fed funds floor). In fact, at times, the SOFR rate (effectively the general collateral rate) has been trading below the funds rate floor, which is not a great look. To help rectify the situation, more available collateral will help, and the counterpart to this is a better balance versus liquidity.
The rise in usage of the reverse repo window has coincided with a fall in bank reserves, which are now running at $3trn. These peaked at $4.25trn in the fourth quarter of 2021. The previous low for bank reserves was $1.4trn in 2019, having come from a prior peak of $2.75trn in 2014. Back then, the Fed’s financial crisis-inspired bond-buying programme came to a conclusion (2014), and a bond roll-off then ran through 2018/19. Part of the fall in reserves reflected an uptick in economic activity and an increase in currency in circulation, and a requirement to build a buffer of high-quality liquid assets, the other part was a reduction in the Fed’s balance sheet as bonds rolled off. Fast forward to today and bank reserves are down from the highs, still at a relatively elevated $3trn, but primed to ease lower through 2023.
Reverse repo balances should be the Fed liabilities that shrink the fastest in 2023
Repo could see a material move higher as reverse repo volumes drop
For 2023, we can see the bond roll-off continuing throughout the year. If things get really tough macro-wise, there may be an argument for the roll-off to be put on pause. But barring the unexpected, there is ample room for the Fed to maintain the roll-off. That could have the direct effect of reducing the use of the Fed’s reverse repo facility. It does not have to, but this facility can wind all the way down to zero, which would be a desirable outcome as the market should not require recourse to repo away from the market. Should things get tight liquidity-wise, the Fed now has a permanent repo facility, where liquidity can be supplied back to the market (with bonds posted to the Fed). Ideally, the Fed should not be required to do large volumes through either of these windows. But they are there as a buffer - a buffer in both directions.
2023 should see an accelerated reduction in usage of the Fed’s reverse repo facility
Overall, 2023 should see an accelerated reduction in usage of the Fed’s reverse repo facility. This should coincide with a rise in general collateral rates to above the reverse repo rate, ideally towards the effective fed funds rate. This is typically 8bp above the fed funds floor, compared with 5bp above for the reverse repo rate. Something like 8-10bp over the fed funds floor would be a good area for SOFR to settle at, correlating with a drying up of the usage of the Fed’s reverse repo facility.
2023 should also see bank reserves falling to US$2.5trn, or lower
Beyond that, there could also be pressure for bank reserves to ease lower, but these should ease lower by far less than the contraction in the reverse repo volumes. We think reserves could slip down to the $2.5trn area, and if they go lower, we’d be surprised to see them dip below $2trn. This leaves them likely some $1trn above the lows seen before the pandemic but in any case at least $0.5trn above those lows.
Lower eurozone liquidity will make Euribor fixings more sensitive to credit risk
The end of abundant liquidity
Most of the decisions pertaining to the withdrawal of central bank liquidity were taken in 2022, but the effects should only become evident in 2023. Even with central banks in various stages of the QT process, it is clear that their preference would be for a faster withdrawal of liquidity than that produced by a simple reduction of their bond portfolios.
In cases where some of that liquidity stems from other policies than QE, for instance, in the case of the ECB’s Targeted Longer-Term Refinancing Operation loans to banks, faster liquidity withdrawal is simply a matter of creating incentives for early repayments. The ECB has taken steps to that end at its October meeting and we’re expecting around half of the €2.1tn TLTRO balances to be repaid by March 2023.
You’d be hard-pressed to show the effect of shrinking liquidity in money markets in 2022. This will change
The Bank of England has an arduous task at hand. The basic principle is to introduce new facilities to absorb liquidity from banks. This, in effect, is what the Fed’s reverse repo facility is doing in exchange for collateral. The BoE has taken no such steps yet but the repo rates and short gilt yields' reluctance to fully reflect rate hikes might trigger calls for faster liquidity absorption.
Truth be told, you’d be hard-pressed to show the effect of shrinking liquidity in money markets in 2022. This will change in 2023. Regardless of the currency zone, the liquidity situation can still be described as plentiful. This, in turn, has resulted in suppressed money market rates. In the case of government bonds and repo, these have diverged further, to the downside, from policy rates. In the case of supposedly credit-sensitive money market rates, they have failed to reflect growing systemic risk and the looming recessions.
UK and German bond scarcity is stretching valuations against swaps
Collateral shortage becoming a monetary policy issue
The other side of the abundant liquidity problem is the shortage of high-quality collateral evident across developed markets, but most prominently making the headlines in the eurozone and UK due to ever-widening swap spreads. On one level, collateral shortage and abundant liquidity are two sides of the same coin: too much money chasing too few assets. On another, regulations and falling unsecured interbank volumes mean the availability of collateral is becoming a problem of money market functioning, which is likely to persist even after liquidity is withdrawn.
Both excess liquidity and collateral shortages can be solved with the same tools
The good news is that both excess liquidity and collateral shortages can be solved with the same tools, as the Fed's experience has shown. The BoE and ECB both have securities lending facilities, but their use is more anecdotal, and insufficient to keep repo rates close to the policy rates. There have been calls for more ambitious facilities to be put in place. The BoE can point to the existing standing and special repo facilities although the lending rate would have to be raised and gilts would have to be borrowed from the Asset Purchase Facility (QE) portfolio.
As is the case in the UK, the ECB can also point to efforts by some institutions, more notably the German Treasury, to release more bonds on repo. More is likely to come, including to finance energy-related spending. Combined with QT, and TLTRO repayments, they will chip away at the collateral scarcity in the eurozone, but we expect the effect to be backloaded to the second half of 2023.