The Fed's Reverse Repo Facility & Instability Risks
The Fed's Reverse Repo Facility & Instability Risks
Yes, it creates safe assets, yet also...
The Fed rightly doesn’t think about QE from a pure “quantity” of reserves perspective, yet the stacks of reserves are the byproduct of its large-scale asset purchases. Because not all reserves are eligible for the interest rate paid on bank reserves (a rate which sits toward the top end of the fed funds target range) and because the Fed cares about the broad suite of money market rates, the Fed needs to reinforce the floor on its target rate. Enter: its reverse repo facility—available to primary dealers, government-sponsored enterprises, many money market funds, and banks.
Post-Pandemic Evolution
The Fed continually lifted the per-counterparty ceiling on the RRP as reserves—and government MMFs—grew following the pandemic. Because reserves can only live within the banking system, they can weigh on banks’ non-risk-weighted leverage ratios—particularly, banks’ supplementary leverage ratios (SLR). Early in the pandemic, the Fed exempted reserves (and Treasuries) from the SLR, but it let that exemption expire after March 31, 2021. Yet, at its FOMC meeting in mid-March, the Fed began raising the RRP per-counterparty limit—from $30 billion to $80 billion, effective March 18, 2021.
Then, in its March 19, 2021 press release announcing that the SLR exemption would expire as scheduled at the end of the month, the Fed noted:
To ensure that the SLR—which was established in 2014 as an additional capital requirement—remains effective in an environment of higher reserves, the Board will soon be inviting public comment on several potential SLR modifications.
But it’s been crickets since then. Now that Michael Barr is installed as the vice chair for supervision, however, we may hear more about this soon.
Banks defended against the possible capital penalties of carrying reserves by discouraging deposits and encouraging clients to “deposit” with the banks’ money market funds instead. This means that reserves that would otherwise not pose a problem for monetary policy rate implementation were being pushed out into the MMF complex because of QE’s demands on bank capital. Many MMFs have access to the RRP, which could soak up reserves and relieve the banking system of its reserves quantity problem. (In April 2021, the Fed expanded counterparty eligibility as well.)
This was effectively sterilization via another Operation Twist: the Fed was buying longer-term assets while providing the market with a greater supply of short-term assets to soak up the reserves (though, without actually meaningfully releasing collateral to the market).
As RRP usage has increased, the Fed has said the reverse repo is working as intended, helping to enforce the floor on short-term interest rates. (Keeping otherwise immobile bank reserves in banks’ deposit accounts at the Fed would also help pull short-term rates up.)
Right now, with the bottom of the target range for the fed funds at 2.25%, the Fed’s overnight RRP is offering 2.3%, up to $160 billion per counterparty (not per sponsor/parent)—and the limit “can be temporarily increased at the discretion of the Chair.” This has likely accommodated the run-up of MMF participation at the RRP:
And here’s the June 30, 2022 RRP operation results:
Just as the Fed’s QE affects the banking system immediately, so has its quantitative tightening. MMFs’ allocation to RRP has continued on an upward trajectory; meanwhile, as Reuters recently reported:
For now, bank reserves are still considered abundant at $3.3 trillion, but the decline has been rapid, some market participants said. From a peak of nearly $4.3 trillion in December last year, bank reserves have declined about 23%. In the Fed's previous quantitative tightening (QT), $1.3 trillion in liquidity was withdrawn in five years.
Quantity vs. Quality of Reserves
As long as capital charges remain on reserve balances, banks may not aim to undo their redirection of reserves to MMFs. And, regardless, it might be an awkward conversation not worth bugging a client over again: “So, I know a year ago we told you your money’s no good here, but…”
This makes sense for an individual bank: the cost to carry the reserve balances on an ongoing basis might be too high relative to the option value of being able to put the reserves to work if there’s a market liquidity shortage. But, for the banking system, it’s less optimal.
For one, it may limit the Fed’s ability to be as hawkish as it wishes. The effective strike price of the so-called “Fed put” has fallen from the equity market down to simply financial stability. And there’s a “Fed call” on interest rates: as long as inflation persists, the Fed will likely take as much tighter policy as it can until financial instability forces it into reverse. (Though, even in such a situation, the Fed may have the tools to enable continued tightening.)
That instability episode may come sooner to the extent reserves are roped off in the money market fund complex.
For one, the entire real economy is made up of clients of the banking system. Thus, banks are in a position to quickly both see and respond to unanticipated liquidity shortages. They even pre-position for such contingencies in some cases, with committed lines of credit and the like. Pre-positioned or not, the point remains that a bank can respond in real-time to a client’s unforeseen liquidity squeeze: a loan, a collateral upgrade, a letter of credit, a guarantee of a trading position, etc. To the extent such client assistance has to be funded, however—say, a bank meeting a client’s margin call on an exchange without pressing the client itself to come up with the liquidity—it is important for reserves to be accessible to that bank at that moment’s notice.
Money market funds have a much smaller field of vision. Relative to banks, there’s likely much less overlap between their “depositors” and their borrowers. They also see only (parts of) the money markets, and have limited visibility elsewhere. Moreover, they can use their balance sheet for far fewer things than a bank—namely, only fully funded, short-term, safe loans. Even if the matching between the borrower facing a liquidity gap and the lender (the MMF willing to direct reserves away from the Fed’s RRP) could occur, a new repo (if the borrower has desirable collateral…) or commercial paper deal would take precious time to ink and operationalize.
Thus, the less reserves present in the banking system, the sooner the Fed tightening runs up against its financial stability put.
Going Forward
It’s possible the now-Barr-equipped Fed could quickly exempt reserves from the SLR (and do the pursuant recalibration). But even if that got done tomorrow, the banks—as alluded to above—may not rush to bother their clients to bring their deposits back. And even if bank deposit rates become market-competitive with their MMFs (since there’s no longer a need to bake in a capital charge), they’re unlikely to offer such a sweet deal that deposits flood back overnight.
As such, this may be the world we’re stuck with for this cycle. And this implies a larger terminal size for the Fed balance sheet. Or, at least “larger for longer.”
That said, with SLR relief in place, the Fed could start gradually reducing its ceiling on per-counterparty RRP exposure from the current $160 billion. But the only recent update we have is from the June FOMC meeting minutes, wherein the Fed portfolio manager “noted that, if ON RRP usage continued to rise, it may be appropriate at some point to consider further lifting the per-counterparty limit.”
Comments are turned off here but more than welcome via email (steven.kelly@yale.edu) and Twitter (@StevenKelly49).