NextFin News - The enhanced supplementary leverage ratio, a post-crisis capital backstop that U.S. regulators spent more than a decade defending, is being recast as the problem rather than the solution. In her first public assessment of the overhaul, Federal Reserve Vice Chair for Supervision Michelle Bowman said the recalibrated eSLR freed nearly $5 trillion of additional balance-sheet headroom for the six largest dealer parents in the first quarter of 2026 — and that dealers have already begun deploying it into the Treasury market the rule had quietly constrained.
The stakes sit in one sentence from Bowman's October 1 remarks at the Atlantic Council 2026 CEO & Senior Management Summit in Washington: leverage ratio reforms have "improved Treasury market functioning and strengthened its resilience to stress by relaxing regulatory balance sheet constraints." That is a consequential admission from the Fed's top supervisor — that a rule written to make the system safer had been making its most important market less safe, and that undoing part of it has worked.
The Rule That Bit Back
The eSLR was adopted in 2014 as a simple, risk-blind backstop to risk-based capital requirements. The initial framework required each U.S. global systemically important bank to maintain a supplementary leverage ratio of at least 3 percent plus a 2 percent leverage buffer. Their insured depository subsidiaries needed 6 percent to be considered "well capitalized." The design assumed a floor, not a ceiling.
It became a ceiling instead. Because the leverage ratio does not weight assets by risk, it treated U.S. Treasury securities — the safest, most liquid instrument in the financial system and the collateral backbone of the repo market — the same as far riskier holdings. When the ratio binds, Bowman said, banks rationally retreat from low-risk, low-return activities such as Treasury intermediation and chase higher returns elsewhere without facing higher capital charges. "Dealers have also noted their reluctance to expand Treasury holdings simply out of concern that it could become binding," she said.
The distortion runs deeper than reduced Treasury inventories. A binding leverage ratio penalizes the denominator — total leverage exposure — regardless of asset quality, so it pushes banks toward assets that generate more return per unit of leverage exposure, even when those assets are riskier. The rule meant to protect market stability was quietly driving inefficient capital allocation and undermining the very stability it was designed to protect.
Fed-cited research supports that reading. Work by Falk Bräuning and Hillary Stein found the eSLR was one of the main constraints on dealers' capacity to intermediate in Treasury markets precisely because it is "risk blind." Darrell Duffie and co-authors at the Federal Reserve Bank of New York showed that measures of Treasury market liquidity appeared to worsen as primary dealers faced capacity constraints. And Giovanni Favara, Sebastian Infante and Marcelo Rezende found that following shocks to bank balance sheets, the reduction in Treasury market participation was largest at banks with lower SLRs. The spring 2020 temporary eSLR relief — enacted when the Treasury market seized at the onset of COVID-era tightening — was the first official admission that the rule could bind at the worst possible moment.
That episode is the hinge of the whole story. In March 2020, the Fed and other agencies temporarily excluded Treasuries and reserves from the leverage exposure denominator, and dealers' balance sheets swelled as they absorbed selling pressure. The intervention stabilized the market, but it also created an awkward precedent: the rule worked only when regulators suspended it. Bowman's recalibration is, in effect, an attempt to make the 2020 fix permanent without calling it an emergency measure.
The Recalibration
In November 2025, the Federal Reserve Board, the FDIC and the OCC finalized the fix. The fixed 2 percent buffer was replaced with a buffer equal to 50 percent of each GSIB's method-1 surcharge under the Fed's GSIB surcharge framework. For depository-institution subsidiaries, the buffer is 50 percent of the parent's method-1 surcharge, capped at 1 percent — an explicit acknowledgment that a parent's systemic footprint is not wholly generated inside the bank subsidiary.
The arithmetic is the point. With GSIB method-1 surcharges running roughly between 1 percent and 2.5 percent, the effective minimum leverage requirement falls from a uniform 5 percent to a range of about 3.5 percent to 4.25 percent, according to analysis of the final rule published during the comment period. The most constrained dealers — those that had maintained the thinnest eSLR buffers for 24 consecutive months — received the most relief, which is exactly how a targeted recalibration is supposed to behave.
Bowman framed the change as a return to first principles rather than a loosening for its own sake.
A leverage capital requirement functions best when it serves as a backstop to risk-based capital requirements. Because leverage-based capital requirements do not account for the risk of the underlying activities, they can provide an important signal, particularly during times of stress, and help support market discipline.
The recalibration, in her telling, restores the leverage ratio to that signaling role instead of letting it dictate day-to-day balance-sheet decisions. The 3 percent SLR floor remains untouched, and risk-based capital requirements — which do weight assets by risk — continue to bind for most lending and trading activities. What changed is the margin at which a blunt instrument stops overriding a risk-sensitive one.
The Early Results
"The impact of this recalibration has been encouraging," Bowman said. "So far this year, evidence shows that leverage ratio reforms have improved Treasury market functioning and strengthened its resilience to stress by relaxing regulatory balance sheet constraints."
The numbers she cited are large, specific, and internally consistent. The parent bank holding companies of the six dealers gained nearly $5 trillion of additional eSLR headroom in aggregate in the first quarter of 2026, the first full quarter the modified rule applied. Supervisory data from the Fed's FR 2052a liquidity monitoring report show dealers' total Treasury positions rising from roughly $600 billion at the start of the modification period to more than $700 billion by the end of April — an increase concentrated precisely among the firms that had been most constrained. One dealer reported that leveraged capacity at GSIBs surged from $1.8 trillion in the fourth quarter of 2025 to $6.4 trillion after the rule took effect, and leveraged exposure across GSIBs rose by approximately $900 billion. Some institutions directly attributed higher Treasury holdings and repo market activity to the recalibration in the Fed's March 2026 Senior Financial Officer Survey.
The mechanism is exactly what the pre-reform theory predicted: relax the binding constraint, and capacity flows first to the firms that were most constrained, into the asset class the constraint had penalized most. That alignment between prediction and outcome is what makes the early results more than a correlation — it is the signature of a constraint actually binding and then actually releasing.
There is a subtler read in the data worth holding onto. The $5 trillion of headroom is a stock of unused capacity; the $900 billion of deployed exposure is the flow. Roughly 18 percent of the freed capacity has been put to work in the first quarter. That leaves a large reserve of latent intermediation capacity — which is reassuring if the next stress event requires dealers to absorb a sudden wave of Treasury selling, and concerning if the question regulators really want answered is whether the reform is already being fully utilized.
What the Market Was Pricing — and What It Wasn't
The market reaction on the day of the speech was measured. The 10-year Treasury yield hovered near 5.33 percent — its highest level since 2002, after a quarter that one market summary described as the worst for Treasuries since 1994 — while bank shares edged modestly higher. That backdrop is not incidental to Bowman's message; it is the complication inside it.
The first-order effect of the eSLR change is mechanical and, on the evidence so far, largely achieved: headroom was created, and dealers used it. The second-order effect is a harder test — whether regulatory capacity translates into market resilience under stress. Capacity that sits idle in calm markets does not prove liquidity will hold in a selloff. The March-through-April buildup in Treasury positions happened as yields were already climbing toward multi-decade highs, meaning dealers were accumulating inventory that was simultaneously being marked down by the bond market's repricing of term premium, inflation expectations, and fiscal supply.
Here is the gap between what the reform accomplishes and what the market still fears. The eSLR governs how much balance sheet a dealer can deploy; it does not govern whether deploying that balance sheet is profitable or safe when rates move violently. A dealer with more eSLR headroom can hold more Treasuries, but if a disorderly move forces inventory losses that erode capital faster than the thinner buffer anticipates, the headroom evaporates precisely when it is needed most. The real experiment begins if a disorderly move forces dealers to decide whether expanded balance-sheet space makes them buyers of last resort or simply larger sellers with more room to run.
This is where the cyclical-versus-structural call matters. The eSLR fix addresses a structural flaw in the capital framework — a risk-blind rule that penalized the safest assets — and on that dimension the change is durable; a recalibrated buffer does not revert on its own. But the improvement in Treasury market functioning that Bowman is measuring is, in the near term, riding a cyclical wave of capacity deployment that will mean-revert once the easiest balance-sheet adjustments are complete. The structural gain is the removal of a distortion; the cyclical gain is the one-time surge in positions, and the two should not be confused when judging the reform's success.
The Counter-Thesis: Did Regulators Trade One Distortion for Another?
The strongest argument against the recalibration is not that it failed — the data Bowman presented say it worked — but that it succeeded at the wrong task. Critics of looser leverage standards, including voices who warned during the 2025 rulemaking comment period, argue that a surcharge-linked buffer makes capital requirements pro-cyclical: when a bank's perceived systemic risk falls, its required leverage buffer falls with it, precisely when the financial system can least afford thinner loss-absorption. The 3 percent SLR floor remains untouched, and risk-based capital requirements still bind for most activities, but the leverage backstop is now softer at the margin and more sensitive to the very risk models it was meant to backstop.
There is also a distributional concern the speech does not fully resolve. The recalibration applies only to U.S. GSIBs, leaving the leverage framework for smaller institutions unchanged. If the largest dealers gain a comparative advantage in Treasury intermediation that smaller banks and non-bank traders cannot match, the market could become more concentrated, not less fragile — more dependent on a handful of balance sheets that all face the same incentive structure.
Bowman's answer is that a backstop that binds on safe assets is not a backstop — it is a tax on market-making, and it pushed banks toward riskier assets without charging for the risk. That is a coherent defense of the design. It does not fully answer the pro-cyclicality charge, and it leaves one vulnerability exposed: if a future stress event produces losses that a thinner leverage buffer cannot absorb, the reform will be judged not by Treasury market liquidity in calm quarters but by bank solvency in a bad one.
The falsifying signal is observable and near-term. If dealer Treasury positions stall or reverse despite continued eSLR headroom — or if the firms that gained the most headroom do not show up as the most active intermediaries in the next Senior Financial Officer Survey — the mechanism Bowman described is not durable, and the reform's benefits are cosmetic rather than structural. A second, harsher signal would be a widening in Treasury bid-ask spreads during a rate-driven selloff despite the added headroom, which would indicate that capital is not the binding constraint on liquidity after all.
What Comes Next
Short term, the reform is a market-functioning story, not a bank-earnings story. The beneficiaries are the dealers that were most constrained — the largest U.S. GSIBs with the thinnest pre-reform buffers — and the Treasury market's bid-ask depth; the exposed are holders of bank leverage who assumed the pre-2025 eSLR was a permanent floor. Medium term, the question is whether the approximately $900 billion of deployed leveraged exposure survives a genuine volatility episode without forcing a retreat. Long term, the eSLR recalibration is a structural signal that post-crisis leverage rules are being re-examined on functional grounds — the same logic Bowman has applied to stress testing and supervisory tailoring, where she has argued for less volatility and more risk sensitivity.
Base case: Treasury market functioning improves incrementally, with intermediation concentrated at the largest dealers, and the reform is judged a success if liquidity measures hold through the next rate-driven selloff. Upside case: expanded dealer capacity absorbs a disorderly move without Federal Reserve intervention, validating the "preventive recalibration" thesis and reducing the odds of emergency facilities like those deployed in 2020. Downside case: a sharp widening in yields forces inventory losses that overwhelm the thinner buffer, and regulators face the same choice they faced in March 2020 — step in, or watch the market they tried to fix seize again.
The eSLR was supposed to make the financial system safer by making leverage harder to grow. What Bowman's first report card shows is that it made Treasury market-making harder instead — and that the fix, so far, has worked. The harder test is whether it works when nobody is clapping.
Explore more exclusive insights at nextfin.ai.
