Bowman on Treasury dealers: distinguish capacity, positions and participation

The question is how additional room becomes actual intermediation

In her October 1 speech, Federal Reserve Vice Chair for Supervision Michelle Bowman offered a favorable initial assessment of changes to the enhanced supplementary leverage ratio, or eSLR. She reported that dealer Treasury positions rose from roughly $600 billion at the start of the modification period to more than $700 billion by the end of April. She also said only a few institutions used additional capacity for Treasury activities.

That combination makes the distribution of the response as important as its aggregate size. A research note should ask three separate questions: how much room was available, how much activity actually took place, and how widely the response was shared. Answering one does not settle the other two.

This article compares the speech with the bank survey it cites and earlier research on dealer constraints. It does not estimate the effect on any individual bank’s earnings, predict Treasury yields or treat a regulatory assessment as a new interest-rate decision.

Keep the stock of holdings separate from available capacity

Bowman described the eSLR as a leverage-based capital backstop to risk-based requirements. Her speech links the recalibration to greater dealer capacity, but the supporting staff note is identified as forthcoming. The underlying supervisory analysis has therefore not been independently reproduced here.

For a working research sheet, place capacity and holdings on different lines. Capacity describes room under a stated constraint; holdings describe securities actually carried at a point in time. A change in the former is not a purchase order. A change in the latter does not reveal, by itself, why those securities were acquired or how much unused room remains.

Consider a deliberately simplified hypothetical dealer with capacity of 100 units and holdings of 60 units. If capacity rises to 140 while holdings remain 60, unused room rises from 40 to 80 units, but holdings have not increased. If holdings subsequently reach 75, the increase in holdings is 15 units and unused room is 65. The numbers illustrate bookkeeping only; they are not eSLR calculations or estimates for a real institution.

The example prevents a common modeling error: inserting a reported capacity increase directly into an asset-growth or revenue forecast. Such a forecast needs evidence about deployment and the economics of the activity, not merely evidence that one constraint became less restrictive.

Use the March survey to test how broad the response was

The March 2026 Senior Financial Officer Survey supplies a useful qualification. A majority of responding U.S. globally systemically important banks reported that the eSLR changes had, or were expected to have, no effect on holdings of reserves, non-reserve Level 1 securities or Level 2A securities. A majority also reported no change in their institution’s or affiliated dealer’s repo activity that year.

The survey was distributed on March 20 with responses due March 30. Those responses mix observed and expected effects, while the speech’s position comparison extends through April. They are not interchangeable observation windows. Nor does a majority measured by respondent count tell us the dollar-weighted change across the entire group.

Suppose, hypothetically, six institutions report no change and two expand substantially. The majority response can remain unchanged even when the aggregate dollar total rises. This is a possible reconciliation, not a claim about the survey’s individual respondents. Their undisclosed behavior cannot be reconstructed from the summary.

Record both findings rather than choosing the one that supports a preferred conclusion. The survey is useful evidence about breadth and reported attribution. The aggregate position figure addresses scale. Neither should erase the other, and the difference is a reason to investigate timing and coverage rather than declare an immediate contradiction.

Historical evidence supports a mechanism, not a completed causal test

The New York Fed’s abstract for Staff Report 1070, published in 2023 and revised that October, reports that Treasury liquidity was unusually weak when dealer balance-sheet utilization became sufficiently high, beyond what yield volatility alone would predict. Its findings are consistent with constraints that become binding under pressure.

That is relevant background for asking whether spare capacity helps during demanding conditions. It does not independently evaluate the 2026 recalibration. An earlier study can support the plausibility of a mechanism while leaving the size and timing of a later policy effect unresolved.

For a stronger test, specify the outcome before examining the results. Are we looking for larger positions, narrower trading costs, greater transaction volume or better performance during stressed sessions? These are different measures. A research design that switches between them whenever one looks favorable will be difficult to evaluate.

Also distinguish a before-and-after comparison from evidence that isolates a cause. Changes in demand, volatility and the composition of market participants could accompany the same period. Listing those alternatives does not disprove a regulatory effect; it makes clear what an empirical assessment would still need to separate.

What to add to the next evidence update

Maintain a compact ledger with an observation date, the institution group, the measure, its unit and its source. Flag whether each entry is an observed result, a respondent expectation or an official’s interpretation. Keep missing details visible instead of supplying assumptions that appear to be measured facts.

The most valuable follow-up would connect the same institutions across the same period: their available capacity, actual positions and measurable intermediation outcomes. That would help test whether the response is concentrated, whether it persists and whether it appears when market conditions become more demanding.

Until that evidence is available, the useful conclusion is a research distinction rather than a trading instruction. Additional room, greater use of that room and improved market functioning are successive propositions. Each deserves its own evidence, especially before a broad policy narrative becomes a bank-specific financial forecast.