The Bank Policy Institute hosted a live webinar on Oct. 5, 2023, discussing the outlook for the Fed’s balance sheet and money market conditions, and for banking conditions and bank regulations. The discussion was organized around two panels joined by 25 additional participants featuring banking professionals, academics and researchers from the public sector. The agenda and participant list are copied below.
BPI began these roughly annual discussions, mostly on the intersection between Fed balance sheet policy and bank regulations, in February 2019, when money market conditions were beginning to show signs of reserve scarcity much earlier than anticipated, a result attributed in part to bank liquidity requirements.
A link to a recording of the symposium is available upon request.
Panel 1: The outlook for the Fed balance sheet and money market conditions
The panelists and participants identified several reasons why banks’ demand for reserve balances (deposits at a Federal Reserve Bank) could be higher than had been anticipated. The rapid deposit outflows experienced by some banks this spring, as well as a recognition that high-quality liquid assets (HQLA) held in held-to-maturity accounts are less liquid than had been thought, has likely increased banks’ demand for cash to cover contingencies. Those calculations have been amplified by the potential that large increases in the overnight reverse repurchase agreement (ON RRP) facility or Treasury general account (TGA) could reduce the aggregate supply of reserves, leading each bank to want to hold a higher level. Moreover, banks have likely been encouraged by examiners to hold more cash, and there is an expectation that coming adjustments to regulatory requirements will do the same. Responses to the Fed’s Senior Financial Officer Survey have similarly indicated that banks for the most part either do not expect their reserve balances to decline further or expect them to go up. Adding it all up, one panelist judged that the minimum ample level of reserves could be a bit higher than $3 trillion, which is where they are now.
The panelists differed, however, over whether reserve balances were already becoming scarce. On the one hand, there were some signs that banks were already straining to maintain liquidity. Despite the decline in the Fed’s balance sheet, banks have been maintaining their level of reserve balances with the result that the decline in the Fed’s securities holdings under its quantitative tightening (QT) program has been matched for the most part with declines in use of the ON RRP facility. While there has been a reduction in lower-cost retail deposits, banks have replaced them with large time deposits and other borrowing. Although Federal Home Loan Bank (FHLB) advances have not been rising, panelists speculated that this phenomenon may be due to bank examiners providing guidance that banks should refrain from borrowing from the FHLBs.
On the other hand, money market spreads have remained steady, showing no signs of building scarcity. In particular, the GCF-triparty repo spread has not widened nor has the SOFR – fed funds rate spread narrowed.
A critical question for assessing the overall availability of liquidity for commercial banks as well as the outlook for QT is the substitutability between ON RRP and reserve balances. If they are highly substitutable, the relevant quantity of liquidity can be thought of as the sum of the two Fed liabilities, but if there are frictions hindering the flow between the two, then the level of reserve balances by itself is more relevant.
There are some reasons to judge that the substitutability is relatively high. In particular, in the spring, when bank deposits shifted to government-only money funds, banks took down FHLB advances, and money funds invested in FHLB discount notes. Moreover, the level of ON RRP had varied across a wide range this year, suggesting it was flexibly determined.
There were also reasons to judge the substitutability to be low. The shift from ON RRP to reserve balances via FHLBs can take days or weeks to play out, whereas liquidity can be needed on the same day or even intraday. With short-term interest rates expected to stay higher for longer, corporations could continue to keep their cash in money funds, especially because banks are not eager to attract nonoperational corporate deposits, money not needed for the corporation’s day-to-day operations, because the liquidity coverage ratio (LCR) requirement assigns a high outflow rate to such deposits.
Nevertheless, the panelists all agreed that QT was likely to end in about the middle of next year, about a year earlier than the Fed seems to have expected. The earlier expected end to QT reflected the significantly higher demand by banks for reserve balances. Panelists judged that that demand could only be reduced materially if the Fed were to make banks more comfortable using the discount window or standing repo facility (SRF) to meet same-day contingency funding needs, which they determined to be unlikely. That prospect seemed particularly dim as long as examiners forbid banks from pointing in their internal liquidity stress tests to the discount window or SRF as the means by which they would monetize assets. Panelists expected QT going forward to continue to reduce the size of the ON RRP facility, although they differed on how low facility usage would drop. One panelist estimated that use of the facility could decline organically to about $750 billion to $1 trillion, and, if the Fed intervenes, to about $500-600 billion. Another panelist judged that the facility would be about $100 billion when QT ends.
The panelists were all somewhat concerned that the end of QT could be accompanied by repo-market volatility. In March 2019, the federal funds rate rose above the IOER (interest on excess reserves) rate, providing a signal that reserves were becoming scarce well in advance of the September 2019 episode of repo volatility. However, that signal resulted from banks borrowing from FHLBs in the fed funds market to boost their LCRs, and banks’ LCRs are currently elevated, so there might not be a similar signal this time around. The potential for volatility will depend on the distribution of reserves as well as the total level because liquidity shocks will hit each bank differently, and some will have insufficient reserves at that moment while others will have surplus.
The Fed may be looking for an increase in SRF usage as a signal that reserves were becoming scarce, but there was some skepticism that such an uptick should be expected. Panelists and participants suggested that the Fed consider widening the counterparties for the SRF from the current list of primary dealers and very large banks to include more banks as well as other large borrowers in the repo market, presumably hedge funds.
The sensitivity of money market rates to the level of reserves has been high when the ON RRP facility is low, so when reserves approach scarcity levels any shocks to reserve demand or supply, or repo demand or supply, could generate considerable spikes in repo rates. Moreover, such volatility could reduce the willingness of hedge funds to continue their high levels of intermediation between interest rate futures and off-the-run Treasuries, creating turmoil in Treasury markets.
Lastly, the panelists all expressed the view that the Fed was unlikely to end QT in reaction to the recent sharp rise in Treasury yields. While noting that the Fed could conceivably end QT for market functioning reasons, the Treasury market was functioning reasonably well. The rise in rates was the result of an upward revision in the outlook for Fed policy and heavy borrowing by Treasury.
Panel 2: The condition of and outlook for the banking system
The second panel started off by discussing the key trends on which bank investors are focusing, which suggest challenges for banks on the horizon. The number one issue is deposit flows and funding pressures. While there has been some stabilization of total deposit flows, there has been a shift toward more expensive types of deposits. Investors are focused on how much of a bank’s deposits are core rather than brokered and how much must a bank pay up to retain deposits. On the asset side, loan growth is slowing, especially commercial loans. Banks are tightening lending standards in response to concerns about the economic outlook and selectively reducing risk-weighted assets in anticipation of Basel III Endgame. Taken together, net interest income is declining, although possibly at a slower rate than earlier in the year. With a challenging set up for revenues heading into 2024, banks are focusing on cost reduction plans.
In addition, loan credit performance is beginning to deteriorate. Delinquency rates on consumer loans are increasing, and a rising share of commercial real estate loans are nonperforming.
Lastly, capital ratios are feeling tight. Markets are focusing on capital measured after taking into account losses on investment account securities, both available-for-sale and held-to-maturity. Capital pressure from securities losses and pending Basel Endgame reforms is pressuring banks to slim down their balance sheets by optimizing loan mix, letting assets with higher risk weights run off and testing various credit risk transfer options. Banks and their equity investors are hoping the companies will not be forced to sell entire lines of business or resort to dilutive equity issuances.
The second panelist discussed lessons for regulation and supervision from the bank turmoil this spring. The panelist observed that the banks that failed had not managed their interest rate risk, their supervisors had failed to require necessary changes, markets had not imposed discipline increasing risk premiums on the bank’s liabilities, accounting rules had inflated regulatory capital measures, and authorities treated banks as systemic in death but not in life. He argued that regulations need to acknowledge weaknesses in supervision and the inability of government agencies to credibly commit to limit deposit insurance, limit official lending or implement resolution regimes.
The panelist proposed that regulations should be simple and transparent and set stringent standards that focus on prevention. He argued that the regulatory changes that would satisfy these principles and address the deficiencies observed this spring are to raise capital and liquidity requirements, improve stress tests, and modify accounting standards, while expanding subordinated debt requirements would be much less effective. A discussant observed, however, that all of the long-term debt holders of the U.S. banks that failed this spring, as well as the bondholder in Credit Suisse, took losses, imposing market discipline and supporting effective resolution. The panelist also suggested that liquidity requirements be recalibrated to account for interest rate risk—for example, by making assumed run-off rates sensitive to capital ratios—and argued that the Fed’s current proposal for raising capital requirements is too modest.
The third panelist questioned whether banks needed more capital in aggregate and whether the Basel Endgame proposal would be helpful. He pointed out that estimates of the optimal level of capital in research using modern macro tools and published in peer-reviewed journals, ranged from 6 to 16 percent with a midpoint of 11 percent whereas banks’ CET1 ratios average about 13 percent. The Basel proposal would raise risk-weighted assets for global systemically important banks by 25 percent and for regional banks with more than $100 billion in assets by 10 percent. The proposal would impose charges for operational risk, especially punitive for banks that earn a substantial share of their revenue from fee income. Additionally, it would impose charges for market risk that were reasonably designed but somewhat duplicative of the market-risk shock in banks’ stress tests. Consequently, it would be beneficial to take a renewed look at the design of the global market shock in the stress tests. A number of participants observed that the changes would do nothing to address the problems that caused the bank failures this spring – interest rate risk and liquidity risk – and would move more intermediation into shadow banks with negative consequences for financial stability.
The moderator suggested several ways that stress tests could be improved including by using multiple scenarios, incorporating liquidity risk, and testing a wider set of institutions. It was noted such changes could be done by the Fed internally and used to identify vulnerabilities for supervisory follow-up rather than for the purpose of setting capital requirements.
Participants also expressed concern about the high level of uninsured deposits. Such deposits had grown substantially since the 1990s, rising from 20 percent of deposits to 40 percent currently. Moreover, the rarity of uninsured depositors taking losses in bank failures suggests they are implicitly guaranteed, resulting in an undesirable misalignment between official policy and reality. It was noted that not all uninsured deposits are the same, with some stickier and some more inclined to run, requiring a granular approach to the issue.
Lastly, many participants noted that the concept of a deposit beta – the ratio of the change in the deposit rate to the change in the federal funds rate over the same interval – was flawed. The deposit beta expresses a completely linear relationship whereas the relationship between deposit rates and market rates is highly nonlinear and cyclical. A focus on standard linear deposit beta could reduce the quality of risk management by missing the convexity in the behavior of deposit rates.
9:00 to 10:30 AM: Federal Reserve balance sheet and money market conditions
Moderator: Brian Sack (Formerly of DE Shaw and the New York Fed)
Kicking off the conversation: Mark Cabana (BofA), Michael Cloherty (UBS) and Teresa Ho (JPM)
Potential topics include the outlook for the Fed’s balance sheet and reserve scarcity, ON RRP, demand for reserves post-SVB, QT, ratchet effects from QE, Treasury market capacity
10:30 AM to 12 PM: Condition of and outlook for the banking sector
Moderator: Til Schuermann (Oliver Wyman)
Kicking off the conversation: Francisco Covas (Bank Policy Institute), John McDonald (Autonomous Research), Kim Schoenholz (NYU Stern)
Potential topics include the outlook for bank profits, deposits, credit and regulations as well as the lessons from the bank failures in March and April
|David Beckworth||Mercatus Center|
|Richard Berner||New York University|
|Ulrich Bindseil||European Central Bank|
|Seth Carpenter||Morgan Stanley|
|Darrell Duffie||Stanford University|
|Bill English||Yale University|
|Jason Granet||BNY Mellon|
|Refet Gurkaynak||Bilkent University|
|Beth Hammack||Goldman Sachs|
|David Mericle||Goldman Sachs|
|Ellen Meade||Duke University|
|Patricia Mosser||Columbia University|
|Pat Parkinson||Bank Policy Institute|
|Simon Potter||Millennium Management|
|Matthew Raskin||Deutsche Bank|
|Philipp Schnabl||New York University|
|Samuel Schulhofer-Wohl||FRB Dallas|
|Bruce Tuckman||New York University|
|Lawrence White||New York University|
|Jonathan Wright||Johns Hopkins University|
|Nathaniel Wuerffel||BNY Mellon|
|Patricia Zobel||Formerly FRBNY|