October 2024 – The Slippery Business of Fixing US Bank Liquidity and What Could Help

During the pandemic, lockdowns and supply chain challenges shutdown much of the economy. Significant monetary and fiscal policy interventions came to the rescue. However, for US banks large scale quantitative easing created trillions of uninsured deposits. Pandemic era waivers of US banks’ leverage ratios also encouraged some banks to invest in “risk-less” long-dated government securities as quantitative easing created trillions in zero yielding bank reserves at the Fed that depressed US banks’ net interest margins. Some US banks invested new deposits in long-dated fixed rate assets at super low interest rates. Highly stimulative monetary and fiscal policy was maintained too long in the face of supply disruptions and resulted in inflation. Significant monetary policy tightening resulted.

All but the largest US banks had no meaningful guardrails on interest rate and liquidity risk. After being slow to recognize inflation risks, in the fall of 2022, the Federal Reserve embarked on a significant tightening cycle. Quantitative tightening (QT) and growth in the Fed’s ON RRP facility reduced US banking sector liquidity and pressured banks’ deposits. Significant rate hikes increased US banks’ unrealized securities losses to over $600 billion. High losses posed risks to confidence in some US banks’ regulatory capital metrics. These low yielding assets and greater deposit competition also began to erode US bank earnings, a key first buffer against loss. In the spring of 2023, three US banks with over $500 billion in combined assets failed. Plain and simple — Silicon Valley Bank (SVB), First Republic and Signature took on excessive interest rate and liquidity risk, but macro-financial policies and insufficient regulatory guardrails helped create the necessary conditions.

On Sept 26, 2024 Federal Reserve Vice Chair Barr outlined his thoughts on potential changes to US bank liquidity regulation in a speech titled “Supporting Market Resilience and Financial Stability.” He outlined five potential adjustments to the US bank regulatory framework for liquidity:

US banks subject to internal liquidity stress tests (ILSTs) can assume “some” official sector monetization in ILSTs and their contingent funding plans.[1]
Larger US banks will need to maintain a minimum amount of reserves and pre-positioned collateral at the discount window, based on a (yet to be unspecified) fraction of their uninsured deposits.
(Unspecified) limits on larger banks’ held-to-maturity (HTM) securities in their liquidity coverage ratio (LCR) and ILST liquidity buffers to reduce interest rate risk.
With regards to the LCR, changes to assumed runoff rates for certain types of deposits will be proposed.[2]
“Revisiting the scope of application of the current US liquidity framework for large banks,” was suggested. However, details remain unclear.

Through this speech and others, the Federal Reserve has created a significant policy buzz around liquidity regulation focused on uninsured deposits, pledging collateral at the discount window and destigmatizing and modernizing the discount window. It is true that SVB was unable to access the discount window and that the failed banks had elevated levels of uninsured deposits. Modernizing the technology that supports the Fed’s discount window is useful and encouraging all US banks to be set up to obtain a discount window loan is sensible.

However, proposed liquidity regulations related to requiring banks to preposition collateral at the Fed in some proportion to their uninsured deposits rest on three flawed premises:

Fed lending can cure a bank’s liquidity problems in all states of the world;
contingent liquidity is as good as on-balance sheet liquidity; and
collateralizing a portion of uninsured deposits at the discount window can reduce deposit run risk.

First, central bank liquidity provision to a bank with significant embedded interest rate losses simply does not work when inflation is a risk and interest rates are high. Unlike SVB, First Republic was able to successfully tap the Fed for ~$100 billion in borrowing. However, given First Republic’s low asset yield, borrowing from the Fed only increased the bank’s cost of funding, eroded its economic capital, and enabled deposit flight. Sound bank regulation should work in all states of the world, including when interest rates are high and inflation is elevated. Additionally, much of First Republic’s interest rate losses related to its 30-year fixed mortgage portfolio, not its securities holdings. Such risks can only be fixed through a regulatory standard limiting excessive interest rate risk. The Basel Committee adopted a standard on interest rate risk. The US never implemented it nor adopted any standard at all after the savings and loan crisis. The result is that the US banking sector faced far more challenges than any other advanced economy banking sector as monetary policy tightened globally.

Second, contingent liquidity is borrowed liquidity. Liquidity buffers are supposed to allow banks to meet their obligations timely without damaging capital. It increases a bank’s leverage. Contingent liquidity also may stress a bank’s earnings when monetary policy is tight. By contrast, when a bank already maintains adequate on-balance sheet liquidity that it can monetize, the bank can withstand a broad range of shocks, including interest rate shocks. Such a bank need only to utilize contingent liquidity when the shocks are unexpectedly large or market functioning is impaired.

In 2023, when confidence was lost in regulatory capital metrics failing to include interest rate risk, those US banks subject to the liquidity coverage ratio (LCR) were better able to maintain depositor confidence. Recent Fed research shows that no US banks with assets greater than $250 billion (all of which were subject to the LCR) were subject to a run in March 2023 even though some of these banks also had high unrealized securities losses. Recall that SVB and First Republic were $211 and $232 billion in assets, respectively, so a stabilizing LCR effect really seems more plausible than a too big to fail effect. After the 2018 Economic Growth, Regulatory Relief, and Consumer Protection Act (EGRRCPA), US banks with assets above $50 billion but below $250 billion were no longer subject to the LCR. Instead these banks were subject only to internal liquidity stress tests (ILSTs). Unlike the LCR, banks subject only to ILSTs do not have their liquidity buffers (including HTM) marked to market and publicly disclosed. From the failed US banks’ ex-post material loss reviews, it is clear that supervisors did not adequately enforce ILSTs. Why should depositors and bank creditors have more confidence in banks’ ILSTs in the future?

Per an August 2024 Federal Reserve Reg YY FAQ banks can assume monetization of “some” of highly liquidity assets at the Fed. How much is “some”? In an idiosyncratic stress, the only reason for a bank to need to monetize buffer assets at the Fed is if those assets are not trading near par. This FAQ inadvertently could increase some US banks’ willingness to take interest rate risk. The LCR both through marking large banks’ liquidity buffers to market, including HTM, and disclosing bank liquidity publicly helps support confidence and reduce run risk. Most other countries apply the LCR to all banks. It is time to recognize that tailoring, which narrowed the LCR’s scope of application to US banks in 2018 was a cause of the 2023 regional bank runs.

Third, placing requirements on banks to collateralize uninsured deposits at the Fed will not reduce run risk. The FDIC Improvement Act of 1991 imposes limits on Federal Reserve discount window lending to an undercapitalized bank. Thus, an ex-ante Fed commitment to fund discount window loans to all covered banks at all times may not be fully consistent with current US law. Perhaps Fed policymakers believe — like some other Fed crisis facilities – a central bank balance sheet commitment is sufficient to forestall uninsured deposit runs without draws from the discount window being needed – a mighty assumption. Additionally, requirements to partially collateralize uninsured deposits may not ameliorate bank creditor concerns if a bank’s on-balance sheet liquidity is inadequate. At the same time, a requirement that banks collateralize all uninsured deposits at the Fed implies discount window collateral values and haircuts become shadow capital standards. Finally, as the 2008 crisis showed, banks also can experience liquidity strains from a broad range of factors unrelated to uninsured deposits, including unfunded commitments, mortgage warehouse lines and derivatives. Unlike the Fed’s discount window proposal which focuses on uninsured deposits, the LCR instead considers a broader range of liquidity and funding risks.

To sum up, it is tempting to think that the debate on US bank liquidity regulation is merely re-litigating the events of 2023. However, in my first BTRM column I expressed concern shortly after the 50 bps cut at the September FOMC meeting that because fiscal and monetary policy are both stimulative the US economy is stronger than the Fed may appreciate and that the bond market is signaling that inflation risks again are rising. It is plausible that US interest rates may not come down as much or as quickly as many banks and market participants anticipated. Pressure in US funding markets related both to QT and ongoing rapid growth in the Treasury and repo market is also evident. In short, there remains a risk of further monetary and fiscal policy errors which could spillover to US banks. For this reason, US bank liquidity regulation remains an important topic.

Steps that could help strengthen US banks going forward include:

More fully implementing Basel III in the US — undo “tailoring” of the LCR and other prudential standards and adopt quantitative regulation to guard against excessive interest rate – not roll back post crisis banking reforms late in the economic cycle, as some have suggested. Other countries’ banking systems did not see significant interest rate losses mount relative to bank capital or deposit runs as monetary policy tightened globally.
Over time shifting the culture in banks and supervision towards bankers’ skills, culture, and executive accountability for bank outcomes. Over decades, the increase in bank regulation has created more of a compliance culture as opposed to a risk culture within some banks. There are opportunities to adopt a different supervisory approach. However, for now, the macro environment argues for better regulatory guardrails on liquidity and interest rate risk.
Stopping QT at the November FOMC meeting as evidence of growing pressure in US funding markets is abundant.
Ensuring US banks better understand how unconventional monetary policy impacts US bank liquidity and funding through the cycle.
Separating monetary/fiscal policy from prudential regulation of US banks more cleanly. This will improve monetary policy, fiscal policy and bank regulation and ensure that regulatory policies are more effective even when inflation risks are elevated.

Actions that are unlikely to help strengthen US banks include:

Adopting quantitative discount window requirements on G-SIBs and other large US banks tied to some portion of uninsured deposits.
Over-relying on monetary policy to address banking sector liquidity risks and not addressing interest rate risk comprehensively. Implicitly this approach reflects the belief that when US banks are stressed the Fed will be able to cut the policy rate and inflation risks will not constrain the Fed. Sound bank regulation should work in all states of the world, including when interest rates are high and inflation risk is elevated.
Creating policy confusion around US banks’ use of Federal Home Loan Bank (FHLBs) advances as a tool to term out and diversify their sources of funding. FHLBs are particularly important sources of funding to US banks during monetary policy tightening cycles.
Recalibrating the LCR deposit runoff rates to the experience of three failed US banks to whom the LCR standard was not even applied and that manifestly had inadequate on-balance sheet liquidity. Again, Fed research suggests that LCR banks did not experience deposit runs.
Continuing to argue about implementing Basel End game. Basel III is sufficient.
Proposing further waivers of “risk-free” government securities from US banks’ leverage ratios as opposed to running less accommodative fiscal policy.

[1] Category I-Category IV US bank holding companies (BHCs with assets > $100 billion) conduct internal liquidity stress tests (ILSTs) which are shared with bank supervisors at least monthly. These ILSTs span several horizons: overnight, 30-day, 90-day, and one-year. Banks need to hold highly liquid assets (HLAs), such as reserves at the Fed, Treasuries or agency MBS sufficient to meet estimated stressed net cash outflows. Banks also need to show supervisors that they can convert these HLAs to cash. US banks’ ILST assumptions and results are not publicly disclosed.

[2]In the US Category IV banks and below (banks with assets < $250 billion) are not subject to the liquidity coverage ratio (LCR) which makes them more susceptible to excessive interest rate risk (IRR) in their HLA in light of the new FAQ. Category I-III banks publicly must disclose quarterly their LCR ratio as part of their Pillar III disclosures.