Following last year's bank failures, U.S. regulators have indicated they will propose targeted changes to liquidity requirements, particularly for large banks. Depending on the details, those changes could potentially, in our view, lead to more robust liquidity risk management over time and lower the odds that large banks will fail for reasons similar to those that caused the demise of Silicon Valley Bank (SVB).

Among other things, regulators have signaled they could require banks of a certain size to maintain a minimum amount of readily available liquidity at the Federal Reserve's discount window; restrict the degree to which the liquidity buffers at large banks rely on held-to-maturity (HTM) securities; and update the treatment of certain types of deposits in the regulatory liquidity framework.

Even without those potential enhancements, U.S. banks have largely navigated through threats to their liquidity following last year's bank failures, a sharp rise in interest rates, and a decline in deposits--albeit with help from the Fed's emergency measures. They've also seen some lessening of liquidity pressure in recent quarters, and they could see further relief as the Fed slows quantitative tightening and potentially cuts rates later this year.

However, even if the worst liquidity challenges are behind them, many U.S. banks are operating with some key funding and liquidity metrics that have weakened materially in the last two years from strong levels. Many banks, particularly regional and community banks, have high unrealized losses on their securities; limited cash; and already significant usage of noncore deposit funding in an environment of low deposit growth.

We believe those factors, combined, have eaten into (or could eat into) some banks' appetites for growing loans, particularly beyond what they view as their core customers. Until liquidity strains ease further, we expect some banks to limit their loan growth, partly because of a conservative bias with respect to liquidity risk management. Should liquidity strains unexpectedly intensify, we could take negative rating actions on banks that are poorly positioned.

Potential Changes In Regulation Could Bolster Liquidity In The U.S. Banking System

Following the bank failures of 2023, U.S. regulators have indicated that they'll propose adjustments to the regulatory liquidity framework (as well as changes to capital, resolution, and other areas of regulation and supervision). Michael Barr, the Fed's vice chair of supervision, said in May 2024 that those liquidity-related changes could include:

• Requiring banks of a certain size to maintain a minimum amount of readily available liquidity at the Fed's discount window, calculated relative to their uninsured deposits;
• Restricting the degree to which the liquidity buffers at large banks rely on HTM securities, such as HTM securities counted in the liquidity coverage ratio (LCR) and internal liquidity stress tests (ILSTs); and
• Updating the treatment of certain types of deposits in the regulatory liquidity framework (presumably including the LCR and ILSTs) to reflect the potential for rapid outflows during times of stress.

The impact of these potential changes would depend on the details, but our expectations are that the changes would most likely be manageable for rated banks and that they could potentially result in better liquidity risk management.

A new requirement to maintain a certain amount of readily available liquidity from the Fed's discount window would aim to ensure that banks would be better positioned to meet especially large liquidity outflows occurring in a short period of time--the kind that helped lead to the failure of SVB.  According to Barr's 2023 report on SVB's failure, that bank wasn't able to quickly reallocate collateral from the Federal Home Loan Bank (FHLB) system and its custody bank to the discount window; that step was part of its attempt to quickly obtain liquidity and meet a rapid outflow of deposits.

We assume banks under the new requirement would have sufficient unencumbered assets to pledge at the discount window without trouble, though that would also depend on the amount of their uninsured deposits and the details of the requirement.

A new limitation on the inclusion of HTM securities in liquidity buffers may change how large banks classify their securities under accounting standards--with more securities being counted as available for sale (AFS) rather than HTM.  We would expect that such a change, which could effectively improve the overall liquidity of their securities portfolios, would be manageable for most large banks, though it may force them to more closely manage the interest rate risk on their securities.

When interest rates rose in 2022, many banks counted a large portion of their securities as HTM, presumably to avoid the impact that unrealized losses--resulting from rising rates--can have on balance sheet equity. (Since HTM securities are held at their cost basis, changes in their fair value do not affect balance sheet equity. AFS securities, by contrast, are held at fair value, and unrealized losses thereby weigh on equity.)

Some Category I and II banks (the eight global systemically important banks and Northern Trust Corp.) may also have counted more securities as HTM to avoid the impact that unrealized losses on AFS securities have on regulatory capital. All other U.S. banks have the option of excluding unrealized losses from regulatory capital. However, regulators have proposed requiring all large banks (those generally with at least $100 billion in assets) to count unrealized losses on AFS securities in regulatory capital.

A limitation on the inclusion of HTM securities also might, as Barr said in May 2024, "address the known challenges with their monetization in stress conditions." A security classified as HTM arguably can be less liquid than one classified as AFS. That's because by classifying a security as HTM, a bank indicates it has no intention of selling the security. A sale of an HTM security can raise questions about a bank's accounting for its securities in general, and it can potentially force the bank to reclassify other HTM securities as AFS--with implications for balance sheet equity and, potentially, regulatory capital.

For that reason, banks often may borrow against their HTM securities, rather than selling them, though that typically can only be done with a haircut against the collateral. They may borrow from the FHLB system, from the discount window, in the repo market, or from other third parties.

The third potential change Barr spoke about--an update of deposit outflow assumptions on "a handful of types of deposits" in the LCR and ILSTs--presumably could weigh on those ratios and thereby increase liquidity requirements.  Barr said the outflow assumptions for high-net-worth individuals and for companies associated with venture capital or businesses related to crypto assets need to be recalibrated. We're not sure whether outflow assumptions could rise on uninsured deposits, or at least on nonoperational corporate deposits, in general.

Barr didn't clarify in his recent remarks whether the LCR requirement in general would apply to all large banks. However, in his 2023 report on SVB's failure, he indicated that the LCR could be applied to more firms--presumably Category IV banks (generally those with assets between $100 billion and $250 billion). Regulators eliminated the LCR requirement for Category IV banks as part of the tailoring rule in 2018.

We believe that more stringent liquidity requirements, applied to more firms, could support the stability of the U.S. banking system and could potentially help reduce risk at the individual banks we rate.

U.S. Banks Have Employed A Variety Of Strategies To Protect Liquidity

Since the failures of SVB, Signature Bank, and First Republic Bank in 2023, U.S. banks have improved their liquidity risk management and have largely met deposit outflows, avoiding the stress that plagued the industry in the first part of last year. Since those sizable failures, only three banks--all unrated, with assets of about $6 billion, $66 million, and $139 million--have failed, out of the roughly 4,600 institutions insured by the Federal Deposit Insurance Corp. (FDIC).

The Fed's quantitative tightening helped drive total industry deposits down in both 2022 and 2023. Total deposits and uninsured deposits fell about 5% and 19%, respectively, between year-end 2021 and year-end 2023. FDIC data shows that 2022 and 2023 were the first years since 1946 where deposits fell. Positively, deposits rose in both the fourth quarter of 2023 and the first quarter of this year, climbing about 2% in those six months.

Banks have met deposit outflows without depleting their cash (reported as cash and balances due from depository institutions and the Fed), and they've better positioned themselves to handle any further outflows through a variety of strategies. Cash held by FDIC-insured banks rose between year-end 2022 and March 2024 (see chart 1).

Banks maintained cash in part by simply paying more to retain and build deposits, including by sourcing brokered deposits. An almost $450 billion rise in brokered deposits partly offset a $660 billion drop in nonbrokered deposits in 2023 and the first quarter of 2024.

During this period, a portion of non-interest-bearing deposits shifted into interest-bearing accounts, and time deposits rose while transaction and savings accounts fell (see chart 2). Banks also sourced "reciprocal" deposits, where two banks essentially exchange deposits to gain insurance on them. Reciprocal deposits rose to about $380 billion in March 2024, more than doubling from year-end 2022. While the shift in the deposit mix came with higher costs and hurt banks' net interest margins, it helped them avoid liquidity outflows. For rated U.S. banks, the median ratio of uninsured deposits (adjusted for collateralized and intercompany deposits) to total deposits has fallen below 40%, from almost 50% two years ago.

Banks also supported their cash balances through run-off and sales of securities and an increase in borrowings, including from the Fed's Bank Term Funding Program (BTFP), the emergency program launched after the failure of SVB. Borrowings from the BTFP peaked at $168 billion earlier this year but fell to about $107 billion in early July, with the deadline to request new advances from the program having passed in March. Banks will have to repay all BTFP borrowings by the time the program expires in March 2025.

Chart 1

Chart 2

Chart 3

A Recent Uptick In Deposits Eased Liquidity Strains, And Banks Have Increased Their Contingent Liquidity

While deposits at U.S. banks fell 2% in full-year 2023, they rose 1% in the fourth quarter of that year and 1% in the first quarter of this year, indicating easing liquidity pressures.

The recent slowdown of the Fed's quantitative tightening may also provide relief for deposits. The central bank previously rolled over any principal payments it received on its Treasury securities each month that exceeded the $60 billion cap. That cap came down to $25 billion on June 1, 2024, thereby slowing the decline in its securities.

For mortgage-backed securities, the Fed has kept the cap at $35 billion, but it's now reinvesting principal payments on those securities into Treasury securities.

Banks have bolstered their contingent liquidity, most notably by pledging more assets that they could borrow against at the Fed discount window and the FHLB system. As of March 2024, FDIC-insured banks pledged more than $8 trillion of assets, up more than 25% from 2022 (see chart 4). We believe pledged collateral would allow banks to borrow significantly more, if needed.

Chart 4

Despite Improvement In Contingent Liquidity, On-Balance-Sheet Liquidity And Unrealized Losses Will Continue To Test Many Banks

The Fed's quantitative easing in 2020 and 2021 led to a surge in deposits, leaving the U.S. banking system with a historically high level of liquidity (as measured by cash and securities). Cash and securities still well exceed pre-pandemic levels (see chart 5).

Chart 5

However, when the Fed raised rates, the value of securities fell. Our view is that, for many banks, a combination of limited cash, unrealized losses on securities, and already high usage of noncore deposit funding amid low deposit growth presents risks to their liquidity. But beyond that, we also think these same factors may limit many banks' appetites for growing loans, especially outside of what they view as their noncore customers.

Moreover, these factors compressed the net interest margin and net interest income at many banks with the growth in funding costs outpacing the growth in asset yields.

FDIC-insured banks reported more than $500 billion of unrealized losses on securities as of March 31, 2024, which suggests that many banks couldn't sell material amounts of securities without realizing losses and potentially eating into their capital.

Cash also made up a small amount of assets at a significant number of U.S. banks (see table 1). At more than 20% of U.S. banks, cash made up less than 5% of assets, and unrealized losses on securities exceeded 20% of Tier 1 capital. (We chose 5% in this analysis because, since 1990, cash in the banking system only dipped below 5% of assets in the 2004-2007 period. The annual average is about 9%, and it was 12% at the end of 2023.)

That implies that a sale of securities resulting in realized losses could materially hurt capital at many banks. We believe that, all else being equal, banks with low cash have a weaker ability to meet deposit outflows with cash on hand or by selling securities than banks with higher cash balances and lower unrealized losses.

Table 1

Banks with less cash would probably be less willing to make new loans if that lending further depletes their cash. They could opt to fund new loans by sourcing more brokered deposits or secured funding, but the cost of that funding could limit profitability on new loan originations.

Furthermore, many banks have already used significant amounts of brokered deposits and nondeposit funding. Beyond potentially weighing on profitability, taking on even more brokered deposits or nondeposit funding could also cause a breach of a bank's internal limits pertaining to funding.

Rated U.S. Banks Have Good Contingent Liquidity, But Further Liquidity Pressure Could Contribute To Negative Rating Actions

The U.S. banks we rate have generally increased their access to contingent liquidity over the last year, particularly by pledging more assets to the Fed. The median ratio of liquidity sources--including cash, unpledged securities, and availability on funding lines from the FHLB system and Fed--to uninsured deposits was 158% among the U.S. banks that disclose that availability (see table 2). In other words, rated banks should have the capacity to cover a run-off of uninsured deposits, should that occur in a stress situation. That ratio also does not include unpledged loans, which also could be pledged to garner additional liquidity.

Rated banks' exposure to uninsured deposits has fallen. Some of that is due to the overall decline of uninsured deposits in the banking system. In addition, many banks have pledged collateral against an increasing portion of their uninsured deposits, making those depositors less likely to withdraw their funds should they have concerns about the bank.

We believe banks with ratios of cash to assets that are well below peer levels have mitigated the associated liquidity risk in a variety of ways. They may have below-peer exposure to uninsured deposits or good levels of securities and contingent liquidity access. For instance, uninsured deposits, after adjusting for collateralized and intercompany deposits, make up 20%-30% of the deposits at most of the banks where cash is less than 5% of assets. That compares with uninsured deposit ratios of 70%-90% for the largest banks that failed in 2023.

At the same time, while access to contingent liquidity can be crucial, it doesn't solve all problems. Drawing substantially on secured funding--whether from the Fed, FHLB, or other sources--is costly. It also would be impractical, and probably not feasible, to replace a large amount of deposits that run with a substantial increase in secured borrowings.

We could take negative rating actions on U.S. banks that are forced to draw materially further on contingent liquidity sources; we could also take negative rating actions on U.S. banks that see their on-balance-sheet liquidity drop further, by a meaningful amount. We've already indicated in our rating outlooks on certain banks, including some that have negative outlooks, that further deterioration in liquidity could contribute to downgrades.

Table 2