Treasury market resilience: How big is US banks’ “leverage” now?

It was the 2nd anniversary of Lehman’s demise. The Basel mountain had been slogging heavily and finally produced a mouse; the banking lobby insisted the mouse was a tiger about to gobble up the economy. Alas, the mouse didn’t roar! So remarked a well-respected commentator, only half-jokingly, in 2010.

Yes, I am referring to the leverage ratio that set a minimum level of 3% for equity relative to total assets. If this number looked embarrassingly low, it was because the number was embarrassingly low! Indeed, the rule tripled the earlier requirement but then tripling almost nothing is still almost nothing!

The rewrite of Supplementary Leverage Ratio (SLR) rules marks the first major win for US banks under Trump’s deregulatory agenda. Banks are poised to win back leverage capacity. The trillion-dollar question is how much! According to risk.net estimates, the revised buffer would fundamentally reset and lift the 8 US GSIBs’ aggregate leverage capacity by 30%. Goldman’s balance sheet, which was 91% deployed under the current SLR, would drop to 55%. Wells Fargo would fall from 64% usage to a mere 6%. Top US dealers could gain $6.62 trillion in extra balance sheet capacity. Bank of America would likely gain $1.48 trillion and JPMC $1.06 trillion (relatively less due to its high GSIB surcharge). However, in most cases the binding capital constraint remains risk-based, so Tier 1 capital itself wouldn’t be freed up.

The need for leverage ratio

Regulators have long wished to limit banks’ reckless use of leverage by requiring them to maintain minimum levels of their own equity capital. They apply:

• “risk-weighted” approach with variable capital requirement based on the perceived riskiness of its various assets; AND
• a “leverage ratio”, an overarching constraint on its use of borrowed money.

The risk-based standards give the impression of being scientific; the risk of each of a bank’s assets is measured “scientifically,” and capital determined according to these measurements. One may imagine that a rule based on science trumps a crude one. Such logic has dominated the work of BCBS. Since 1988, regulators have been groping in the dark for the holy grail of the “most appropriate” risk weights. In 2004, Basel II was thought to have got it right, but the 2008 crisis proved otherwise.

The risk-weighted approach has many harmful unintended consequences. It allows banks to reduce their capital by focusing on investments that the regulation considers safe. And then there are derivatives that banks use to shift their risks to others, thus increasing interconnectedness. How can we forget the role played by CDSs in the government’s decision to bail out AIG?

In theory, risk weights are supposed to adapt capital requirements to the risks of the banks’ investments; in reality, the weights are a combination of “politics, tradition, genuine and make-believe science”, and the banks’ self-interest. This lethal cocktail completely ignored some critical but real risks.

Enter SLR

SLR is blunt instrument requiring banks to maintain capital against all assets, irrespective of their riskiness, as a backstop to risk-based requirements.

SLR is defined as Tier 1 capital divided by total exposure, that includes both on- and off-balance sheet items. The broadness of the measure makes it difficult to fiddle with. SLR treated German Bunds and US Treasuries the same as junk bonds and sub-prime mortgages. It was an easy-to-calculate and hard-to-game capital buffer in addition to the normal risk-weighted capital requirements.

US banks with more than $250bn in assets must maintain a minimum SLR of 3%. The 8 US GSIBs are subject to enhanced SLR (eSLR) of 5%, rather than the 3% applicable to other large banks.

Why were banks complaining?

Big banks were upset that though the leverage ratio is intended to be a backstop, it often requires more equity capital than the risk-based rules, thus becoming a binding constraint.

But isn’t the main goal of the leverage ratio to create a minimum capital base as an antidote to imperfect-and-easy-to-rig risk-based standards? They are meant to limit banks’ vulnerabilities to future, unexpected decline in the value of assets that have till now been treated safe and thus have low weights in risk-based measures. Also, if it is too often “binding”, shouldn’t the risk-based rules be strengthened rather than diluting the leverage ratio?

Banks argue that the leverage ratio is counterproductive as it punishes them for holding zero and low-risk assets like US Treasuries and requires unnecessary capital against these investments.

The assumption that there is “no risk” with Treasuries is wrong. Actually, there is no credit risk, however with caveats. There is surely market risk in holding even the safest of assets, as SVB painfully learned in 2023. Also, the Treasuries continuing as a superlative safe asset into the indefinite future is uncertain. The recent Moody’s downgrade of US sovereign debt indicates that massive, growing fiscal deficits and fading attractiveness of the dollar as a reserve asset offer a plausible scenario of a future disruption in timely payments to US debt-holders.

Now that we know why banks dislike the SLR, it’s worth noting that even neutral observers have concerns about its unintended consequences.

The leverage ratio often becomes a constraint on banks during market stress when deposits flow into the banks, making it more expensive for their desks to facilitate trading in the Treasury markets. In the dash for cash in 2020, some banks had to turn away deposits because of the leverage ratio constraints. By disincentivising banks from absorbing some of the selling pressure, the SLR can be irksome when volatility spikes. Even regulators acknowledged this when they temporarily exempted Treasuries from SLR during Covid in 2020.

What has changed in the 25 June 2025 proposal?

On 25 June 2025, the US regulators decided to discard the fixed 2% SLR buffer and replace with a variable add-on equal to half the firm’s GSIB surcharge under Basel’s Method 1 framework. That would bring it in line with the biggest European, Chinese, Canadian, and Japanese banks. The proposal would also shave 5% off the big banks’ requirements to have debt that can absorb losses in a crisis —Total Loss Absorbing Capacity (TLAC)— and cut their minimum long-term debt level by 16%.

Will the proposed rule achieve the goal of strengthening the resiliency of the Treasury market?

Retaining Treasuries in the denominator while reducing the minimum eSLR requirement will allow GSIBs to take far more Treasuries onto their balance sheets. Following the Basel Method-1 standard of an eSLR add-on would (rightly) mean that the more systemically important a bank is, the higher its capital requirement should be. Those rules allow for temporary exclusion of central bank reserves, but not sovereign obligations, from the denominator of the leverage ratio surcharge.

But the proposed rule has drawbacks:

• Prudent risk management would deter banks from increasing long-term bond holdings funded by flighty deposits as seen in the SVB collapse.
• Some large broker-dealers enable hedge funds to increase their leverage in Treasury during normal times by offering them negative haircuts. These dealers exacerbate the liquidity concerns for these hedge funds during a shock. With more regulatory leeway for Treasuries, these dealers might increase their repo activity in ways that contribute to, rather than alleviate, market mayhem.
• True, the proposal may provide a strong incentive for banks to buy more government debt (and lower yields) but it may also redirect capital away from the private sector into public coffers. Rather than stimulating credit to support economic growth, more bank assets would move into T-bills or sit lazily in Fed accounts offering easy gains.
• The SLRs of the 4 biggest US banks have held steady at >6% for the last few years, even as the volume of their outstanding preferred stock – the cheapest form of “tier 1” capital – has been falling. A stable SLR coupled with falling preferred stock indicates SLR isn’t a binding constraint for these banks.

Is there a magic bullet?

The quantum of outstanding Treasuries will continue to grow rapidly relative to the intermediation capacity of the market thanks to large, persistent fiscal deficits and the limited flexibility of the bank balance sheets.

• As interest rates rise, the current market value of a recently issued bond falls. Unlike SVB, Treasuries held by GSIBs as available-for-sale (AFS) are marked to market, but there is no market risk capital charge for the risk of future drop in value due to rising interest rates. This may pose a bigger concern if banks take advantage of relaxation in the eSLR to raise their AFS holdings significantly. The need for regulators to address unlikely but plausible tail risks is a key lesson of the 2008 crisis. Higher rates, particularly on longer-duration Treasuries, could emanate from continued fiscal deficits leading to the US debt/GDP ratio rising further. As Daniel Tarullo of Brookings argues, there is really no reason for not subjecting AFS Treasuries to market risk capital charge as part of relaxing leverage ratio. Also, the leverage ratio is currently the only capital requirement for held-to-maturity Treasuries (since they are not subject to the market risk capital requirements and are not even marked-to-market for capital purposes).
• The proposed central clearing of Treasuries will increase the amount of trade that can be effectively intermediated on existing dealer balance sheets. More client clearing will allow more netting of Treasuries positions at the biggest dealers and reduce the overall size of their leverage assets and thus improve their leverage ratios.

Conclusion

A resilient US Treasury market offers dollar dominance, effective monetary policy, and capital market efficiency – and also financial stability. The biggest risks arise from fiscal and other government policies that endanger the status of Treasuries as the safest asset. With the Fed unwinding its Treasury purchases and foreign central banks diversifying their reserves away from the dollar, the Treasury is increasingly reliant on private players buying and holding Treasuries.

The assumption of continuity in the “zero risk” attribute of Treasuries is at odds with the need for a leverage ratio as a complement to risk-weighted capital requirements. Any proposal must consider the large increase in the volume of outstanding Treasuries while staying loyal to the objective of leverage ratio. Hence, a combination of (1) an increase in risk-weighted capital requirements by subjecting Treasuries that banks classify as AFS to market risk capital charge and (2) recalibrating the minimum eSLR, as proposed currently, is perhaps the most sensible path.

Implementing (2) but not (1) is unlikely to create a safe system. This mouse will never roar loudly enough!

Explore the latest regulatory trends and updates with colleagues at RiskMinds International!

Krishnan Ranganathan, a former Executive Director at Nomura, is currently a Visiting Faculty at various B-Schools in India on topics covering Risk, Finance, and International Banking regulations. He is an alumnus of Harvard Business School and his contributions have been published by Financial Times as part of their initiative to connect classrooms to current events on similar themes and develop students’ critical thinking.