US Bank Reserves, Excessive or Exhausted?

If you have been anywhere near financial markets during the past year, you have not been able to avoid the discussions on the (non-)workings of US money markets. We covered this issue rudimentarily a year ago in “A Fed Watch-Out”. Since then, I/we have hopefully learnt and understood more, and as the ruptures have now also caused the Federal Reserve (FED) to start growing their balance sheet again and buying treasuries directly, it might be a good idea to take another look at the situation.

At the heart of the discussion is the simple fact that the Interest on Excess Reserves (IOER) has proven, with increasing frequency, unable to anchor the effective federal funds rate (EFFR) within the target interval[1]. Put another way: Banks are becoming increasingly reluctant to lend (cash/reserves[2]) to each other at the rates that FED has set. Spikes in demand for reserves have thus become increasingly accompanied by stark rises in banks’ funding costs (in the chart below, I have used the treasury weighted average rate, but the Secured Overnight (O/N) Financing Rate (SOFR) as well as EFFR and other policy-related rates show similar developments):

[1] Given ample liquidity, this floor rate – the IOER – is the FED’s main policy tool, even though the FED’s Fund Rate (FFR) is still the official policy rate.

[2] Reserves limit how banks “create money” (c.f., endogenous money)


To many, including yours truly, these developments are surprising, especially given the FED’s still bloated balance sheet (and historically still very high reserves). However, it does not per definition imply that there needs to be something wrong in the “plumbings of the financial system” as many pundits are suggesting.

A number of plausible explanations have been brought forward, e.g.; large corporate tax payments and high treasury issuances are weighing on banks reserves, which are both perfectly valid explanations. About a month ago, however, the actual funding costs for banks (this is why I use the Treasury Weighted Average Rate in the chart above) started to continuously, and dramatically, exceed the FED’s target interval. Consequently, the Federal Reserve New York branch (with some difficulties, I might add) restarted its Temporary Open Market Operations (TOMOs) for the first time since before the financial crisis.

*Note: In the graph above, I have simply assumed that the volume of TOMO’s should be stable as share of nominal GDP.

Before continuing, I think this is a good place to underline that TOMO’s were a quite normal occurrence before the financial crisis of 2007/08. In the post-crisis years, we have become so accustomed to excess(-ive) liquidity that a whole generation of market participants know no other state of affairs. The only caveat here is, perhaps, that the resumed TOMO-volumes are a bit high, at least if we think they should develop in line with nominal GDP.


In conjunction with reserves still being on what can only, at least from a historical perspective, be described as lofty levels, the above developments have given rise to a number of competing narratives; from the more sanguine: ‘Reserves are starting to become scarce again, what about it?’, all the way over to: ‘A few banks are now being singled out as not trustworthy and have therefore been forced to pay a hefty premium for borrowing. A systemic crisis is to follow.’

While all explanations are possible, only one is plausible

Before we start delving into the official, and more sanguine views, I think it is important that we all agree on the current excess reserve measure (see the  graph above) is quite obsolete. In essence, this measure only takes into explicit consideration the Federal Reserve’s reserve requirements for banks’ deposit liabilities but *not* any other of the new rules and regulations being imposed on the banking sector in the years after the crisis.

One way to highlight this is to look at the FED’s Survey of Market Participants where a question on the expected size of excess reserves have been posited since late last year.


As can be seen the level of maintained balances with FED (black line) have converged rapidly with market participants’ views on what size of reserve they think the bank needs. Add to that a need for a healthy buffer between the actual reserves deposited with FED and those implied from a host of new rules and requirements, and you probably end up with an estimate somewhere close to current levels of excess reserves.

At the beginning of the year, researchers at the FED also made some rough estimates of what reserve balances would be needed to satisfy the liquidity coverage ratio during times of heightened systemic stress at the top 8 US banks. They came up with a number just shy of 1 trillion USD, a number which can of course be assumed to be considerably higher for the banking system as a whole.

Put another way, the above discussion points to a distinct possibility that we have moved from a situation of ample to scarce liquidity over the past year. But are there other, perhaps better, indications of this being a broad, system-wide, lack of reserves?

– This has actually been addressed on a number of occasions by, e.g., (the almost mythical) Simon Potter, former head of Market Operations at the New York FED (and who was ousted from the FED just a few months ago). Among other things, he has suggested that an increased scarcity (of liquidity) should be reflected in overnight (unsecured) funding rates rising above, and staying above, the IOER, as more and more banks would be forced to borrow reserves from other banks:

Indeed, some time during early 2019 most O/N rates turned from comfortably below the IOER to now being clearly above the IOER. Even when looking at the lower percentiles (in the O/N interest rate distribution, also available in the Macrobond application), we can see similar developments. Another signal, Potter states, would be an increase in the daylight overdrafts, implying that more banks want to avoid using O/N liquidity (which would drain reserves):


Indeed, we see some increased management of intraday liquidity, even though it is admittedly far from being eye-catching (it seems as if tri-party repos are becoming more common as a tool for managing “time-critical” liquidity).

The most persuasive argument for the current volatility being caused by a shortage of liquidity is, in my opinion, that we are again seeing a relationship between the changes in reserves and interest rates – this has not been the case for several years (as liquidity has been abundant):

*Note: With some more work you could perform the regressions on weekly data, which would have improved the test statistics, but to do that you first have to adjust for the month-end effects in EFFR (or OBFR).


To me, the above analysis quite convincingly shows that recent volatility on funding markets are simply an effect of reserves becoming more scarce. And as the interest rate floor-system that the FED operates necessitates ample reserves, we now experience spikes in rates and rate volatility (I remain clueless as to why the FED has not opted for a corridor-system in the likes of the ECB or the Riksbank, instead of the recently decided floor system).

Financial crisis 2.0?

That said, there are also some issues that hint at more severe problems in the funding market and which warrant a more sinister interpretation. The first question mark pertains to the simple fact that reserves are still very high from a historical perspective. Relatedly, when the FED asked banks ‘what they believe is the lowest comfortable level of reserve balances’ in the recently released Senior Financial Officer Survey (the SFOS cover ca. 80 banks), their actual reserves are close to double (!) that size, which suggests there should still be ample reserves in the system and that the recent woes might be attributable to something else, e.g., problems within a single (or a few) institutions.

Source: Federal Reserve
*Note: The results of the SFOS rhyme very poorly with the results in the SMP (cf. above).


Another question mark emanates from the skewed distribution of interest rates in recent O/N transactions – it is apparently a very small number of institutions that have been forced to pay prohibitively high premia to borrow from their peers. Why is that? – It could quite simply be that they are the ones who had bigger funding needs and that that comes at a cost in a scarce liquidity environment. But it could, of course, also be a harbinger of dire news from a bank or even a group of banks.


As can be seen, it is obvious that high rates were concentrated to a few market participants. The SOFR depicted also includes tri-party repo transactions, and rates for all but the lowest percentiles of such transaction were noticeably high during September.

The risk of any of the major bank corporations being in funding or any other sort of trouble had most analysts holding their breath this week, as banks’ Q3 reports have come pouring in, and in particular the only tri-party dealer left, BNY Mellon. Hence, it has been alleviating to learn that no bank has lifted up anything out of the ordinary and, from what I have heard, banks’ results are generally quite upbeat – also for BNY Mellon.


The FED’s balance sheet will, and should, remain bloated

From my perspective, and after looking through the data and analysis above, the conclusions are quite straight-forward albeit, perhaps, less exciting:

  • Most major banks – including BNY Mellon – are reporting strong results, and nothing that has come out so far would suggest funding issues for any single market participant (a la Lehman);
  • While it is true that overall bank reserves are high from a historical perspective, quite a few indicators show that recent stress in funding markets is rather due to excess reserves being depleted[1];
  • Among other things, we can see that changes in reserves are again having an impact on interest rates, which would only happen if reserves are becoming more scarce across the board;
  • Thus, the most pressing response to the current volatility on funding markets is not trying to find out which particular bank might be in trouble, but simply to find some sustainable solution to the demand for liquidity in the floor system the FED has chosen;

On that last note, the recently announced TOMOs and “reserve management purchases of Treasury bills” will unfortunately only (and very reactively) provide temporary relief. From my (and many other analysts’) perspective, a standing repo facility or, preferably, a corridor system like those the ECB and Bank of Canada employ would be a much better way to mitigate inherent risks in the current monetary operational framework. I would not be surprised to see that discussion surface again soon, also within the FED.

[1] Yet another way of looking at the issue of ample vs scarce liquidity, and one that might actually be able to reconcile both views, is if some institutions have become “reserve-flooded” while others are simultaneously being “reserve-strapped” as the flow of reserves and collateral have come to a halt.

Scroll to top
Thank you! Your subscription has been confirmed. You'll hear from us soon.
Subscribe to our latest posts