The US Treasuries market, an idol with feet of clay: Oiling the wheels

Since 2023, the US authorities have taken various measures to support the liquidity and stability of the Treasuries market. Transaction transparency has been enhanced, the use of centralised clearing for repurchase agreements has been increased, and a programme to buy back the least traded securities (Treasury buybacks) has been put in place.

However, the balance sheet constraints faced by the banks responsible for intermediating this market remain an aggravating factor in times of stress. To remedy this, on 25 June, regulators proposed relaxing the leverage requirements imposed on systemically important banking groups and their deposit-taking subsidiaries. This measure should restore the leverage ratio's role as a safety net and reassure investors about the ability of banks to fully perform their role as intermediaries. However, the respite offered may be short-lived given the projected evolution of federal debt. The relaxation of the leverage ratio could even inadvertently support the strategies of leveraged funds, thereby reinforcing some of the vulnerabilities that regulators are specifically seeking to mitigate.

Restoring the leverage ratio's role as a safety net

Primary dealers are responsible for intermediating Treasury markets (cash and repo) by participating in Treasury auctions, brokering outright purchases and sales on the secondary market, and facilitating the circulation of cash and collateral on the securities repurchase (repo) markets^[1]. However, the regulatory framework put in place in the wake of the 2008 financial crisis has worsened their operating conditions (Li, Petrasek and Tian, 2024). Basel III^[2] has, in particular, tightened capital requirements linked to the size of bank balance sheets through the leverage ratio (see box). This has increased the balance sheet costs associated with the activities of primary dealers, most of which are subsidiaries of very large banks. However, as federal debt increases, so does the balance sheet space needed to intermediate it.

Distortions caused by regulation

These new banking regulations have not only changed the positioning of primary dealers, but have also had a significant impact on the levels and volatility of returns on the financial markets on which they operate (Du, Hébert and Huber, 2019; Duffie et al., 2023; Favara, Infante and Rezende, 2024; Braüning and Stein, 2024). Thus, whereas until 2008 they favoured borrowing Treasuries (net short position), primary dealers have since become firm holders of Treasuries (net long position). They have also been encouraged to arbitrage between supporting Treasury market activity and dollar supply on FX swap markets, and to demand higher risk premiums (Du, Tepper and Verdelhan, 2018). Banking regulations and technological innovations have also led to the shift of some of the market intermediation activities carried out by dealers to trading firms (Principal Trading Firms), which have very little tolerance for stress situations^[3]. Finally, the deterioration in the absorption capacity of Treasuries by primary dealers is thought to have contributed to the erosion, both nationally (negative swap spreads^[4], even on very long maturities) and internationally (major deviations from covered interest parity^[5]), of the ‘convenience yield’^[6] associated with holding the asset considered to be the safest and most liquid^[7] (Jermann, 2019; He, Nagel and Song, 2022; Du, Hébert and Li, 2023).

Biased portfolio choices

Regularly discussed by regulators since March 2021^[8] and advocated by many researchers (Duffie, 2020; Liang and Parkinson, 2020; Chen, Liu, Rubio, Sarkar and Song, 2021), the question of a permanent revision of the leverage ratio had not, until now, been resolved. As a reminder, the leverage ratio is a non-risk-weighted capital requirement. It aims to ensure that a bank's assets or commitments, regardless of the risks associated with them, do not exceed a certain multiple of its capital. As a result, although considered ’safe’ assets with zero risk weighting for the calculation of risk-weighted capital ratios, US Treasuries or reserves held with the Fed are fully included in the calculation of leverage exposure (SLR denominator) like any other asset, even highly risky ones.

The standard must be calibrated in such a way that it acts as an extreme limit rather than a permanent constraint. Otherwise, it would encourage banks to arbitrage in favour of riskier and more capital-intensive assets that are more profitable. However, for some very large banks and their main depository institutions (Charts 1 and 2), the SLR leverage ratio^[9] is about to become more restrictive than the risk-weighted Tier 1 capital requirement and therefore more decisive in portfolio choices^[10]. Without an adjustment to the standard, some institutions will soon be unable to act as intermediaries in the Treasury market without raising additional capital. The need to recalibrate the leverage ratio so that it does not become more restrictive than the risk-weighted requirement and can fulfil its role as a safety net has therefore become a priority for regulators.

A relaxed rule but little balance sheet space freed up

The proposal

The long-awaited reform is now underway. On 25 June, the three banking regulators (Fed, FDIC, OCC) proposed lowering the level of the enhanced leverage ratio (eSLR). The rule is open for comment for 60 days. The final rule, which will be announced by the end of September at the earliest, will specify the date of entry into force of the new requirement. The regulators plan to reduce the current eSLR requirement (5% for G-SIBs, 6% for their deposit-taking subsidiaries) to the Basel recommended level (3% plus a buffer set at 50% of the G-SIB surcharge calculated using method 1^[11]). On this basis, eSLR requirements would range from 3.5% to 4.25% (Table 1). In the case of the eight G-SIBs, this revision would restore the leverage ratio’s role as a safety net (Chart 3). However, for most of the largest US depository institutions, it would remain as restrictive as, or even more restrictive than, the risk-weighted capital ratio, with the notable exception of Bank of America and Wells Fargo subsidiaries (Chart 4).

A ‘freed up’ balance sheet space that is smaller than it appears. According to regulators, this relaxation would offer the eight G-SIBs Tier 1 capital savings of USD 13 bn in total (1.4% of their Tier 1 capital stock) and USD 213 bn for their main deposit-taking subsidiaries (27%)^[12]. Theoretically, taking into account their capital stock as at 31 March 2025, this relaxation would allow G-SIBs to collectively increase their exposure to risk-free assets by a maximum of approximately USD 6 trn^[13] (Chart 5). After taking into account all capital constraints, regulators estimate that this capacity would not exceed USD 1,100 bn. In our view, these margins for manoeuvre could be even narrower (around USD 300 bn).

The G-SIB surcharges currently in force (under Method 2) were defined on bank balance sheets for 2022. However, these balance sheets have since increased and become more complex, meaning that the surcharges that will apply in the coming years are also likely to increase. This is the case for JP Morgan, whose systemically important score in Q1 2025 would imply a surcharge of 5.5% (compared to 4.5% currently), for Bank of America (3.5% compared to 3%), for Citigroup (4% versus 3.5%), for Goldman Sachs (4% versus 3%) and for Morgan Stanley (3.5% versus 3%). With the given balance sheet structure, only Bank of New York Mellon, Wells Fargo and State Street could increase their exposure to risk-free assets without increasing their current systemically important scores (Chart 5).

The alternatives

Other relaxation measures could be considered. Regulators have submitted four for comment.

- Under alternative 1, the revision of the eSLR requirement (as proposed) would be supplemented by a reduction in the SLR ratio denominator through the exclusion of Treasuries held for trading purposes by broker-dealer subsidiaries of the large banks. We estimate that this exclusion would improve the average SLR ratio of the 20 banking groups subject to the SLR requirement by just 15 basis points^[14].

- Alternative 2 would be to deduct reserves held with the Federal Reserve and the entire Treasuries portfolio from the SLR ratio denominator. On average, this exclusion would reduce the leverage exposure (ratio denominator) of the 20 banking groups subject to the SLR requirement by 14%^[15]. It would improve their average SLR ratio by 100 basis points (Chart 6, Table 2) and that of their largest deposit-taking subsidiaries by 116 basis points^[16] (Chart 7, Table 3).

- Alternative 3 would reduce the current eSLR requirement to 3% plus a buffer set at 50% of the G-SIB surcharge in force (i.e. the surcharge calculated according to method 2). In this scenario, the eSLR requirements would range from 3.5% to 5.25%^[17] (Table 1).

- Alternative 4 would combine the proposed reduction in the eSLR requirement and Alternative 2.

Reassuring investors

Barriers to large-scale purchases of Treasuries by banks

On its own, the relief proposed in June (or the exclusion of Treasuries from the ratio denominator) should not encourage banks subject to the SLR to make massive purchases of Treasuries. In Q1 2025, the weight of Treasuries in the balance sheets of banking groups subject to the SLR (10%) and in the inventories of primary dealers had already reached historically high levels. While Treasuries are viewed very favourably in banking regulations (zero credit risk weighting, eligibility for the range of high-quality liquid assets), their exclusion from the SLR denominator calculation would not make holding Treasuries entirely painless.

Significant purchases of securities would risk:

1/ increasing G-SIB surcharges by inflating the size indicator,

2/ degrading simple leverage ratios (restrictive for deposit-taking subsidiaries),

3/ increasing banks’ exposure to transformation and interest rate risks^[18] and conflicting with internal exposure limits to market risks, Value at Risk^[19] (thereby increasing the value of risk-weighted assets) and

4/ worsening the liquidity position of certain banks^[20].

It therefore seems unlikely that the relaxation of the standard alone will stimulate increased demand from banks and lead to a sharp and rapid decline in Treasury yields. Any comparison with the period of temporary relaxation of the SLR standard in 2020-2021 would be misleading: the decline in yields at that time was the result of the Fed's unlimited QE.

The benefits of confidence

In our opinion, the aim of this relaxation is rather to reassure investors about the ability of banks to fully play their role as intermediaries and, ultimately, to increase their appetite for Treasuries. However, relaxing the leverage ratio can only achieve this objective if it is permanent. Otherwise, in the event of a shock, some investors, fearing a future drop in the price of securities, could be tempted to sell their holdings. According to Eisenbach and Phelan (2022), in March 2020, before the Fed announced its ‘unlimited’ government securities purchase programme, uncertainty about the ability of dealers to absorb net sales of securities prompted some financial institutions, without pressing liquidity or financing constraints, to sell their portfolios prematurely, thereby making their expectations self-fulfilling. Last April, the announcement of higher tariffs did not raise the same fears of market disruption. Moreover, the leverage and liquidity constraints of the largest banks still offered room for manoeuvre. As a result, they were able to continue to act as intermediaries on the repo market and provide hedge funds specialising in cash-futures basis trade strategies^[21] with the resources they needed to maintain their positions. Otherwise, these players would have sold off part of their portfolios, which would have amplified the stress (Perli, 2025).

The risk of unforeseen adverse effects

Bank solvency is preserved

In principle, the measures to relax the SLR (revision of the requirement level, or even exclusion of ‘safe’ assets from the ratio denominator) do not pose any risks to financial stability. The fact that risk-weighted solvency requirements remain more restrictive than the leverage ratio for most large banks (and only slightly less restrictive for others) rules out the risk that they will increase their exposure to risky assets. This is evidenced by the low capital savings expected from the relaxation of the eSLR at the consolidated level (USD 13 bn according to regulators' estimates).

An incentive for short-term strategies

In our opinion, beyond the risks mentioned by regulators in their impact assessment (increased leverage and exposure to interest rate risk), efforts to facilitate the role of dealers could nevertheless have unintended adverse effects. Just like the rise of centralised clearing for Treasuries repos (and more specifically ‘sponsored’ transactions^[22], Chart 8), the relaxation of the leverage ratio could expand the scale of hedge funds' cash-futures basis trade strategies.

Hedge funds have become major intermediaries of interest rate risk in the Treasury market by taking positions in the futures markets against asset managers. Increased availability of repo loans would support the profitability of their strategies. Admittedly, these strategies enable the Treasury to place some of its debt and support market liquidity. However, given the strong leverage effect inherent in these transactions, in the event of a volatility shock on the Treasury market, a rapid unwinding of their positions would increase yield volatility and reduce market liquidity.

Ultimately, relaxing the leverage ratio could have a paradoxical effect. While it aims to strengthen the stability of the Treasuries market, it could, incidentally, support leveraged fund strategies, thereby reinforcing some of the vulnerabilities that regulators are specifically seeking to mitigate.

Article completed on 3^rd July 2025

REFERENCES

Bräuning F. and Stein H. (2024), The Effect of Primary Dealer Constraints on Intermediation in the Treasury Market, Federal Reserve Bank of Boston Research Department Working Papers

Chen J., Liu H., Rubio D., Sarkar A. and Song Z. (2021), Did dealers fail to make markets during the pandemic?, FRBNY, Liberty Street Economics, March 2021

Du W., Tepper A. and Verdelhan (2018), Deviations from covered interest rate parity, Journal of Finance, vol 73

Du W., Hébert B. and Huber A. (2019), Are intermediary constraints priced?, NBER Working Paper Series, June 2019

Du W., Hébert B. and Li W. (2023), Intermediary Balance Sheets and the Treasury Yield Curve, NBER Working Paper Series, May 2023

Duffie D. (2020), Still the World’s Safe Haven? Redesigning the US Treasury market after the Covid-19 crisis, Hutchins Center WP n°62

Duffie D. (2025), How US Treasuries Can Remain the World’s Safe Haven, Journal of Economic Perspectives, Volume 39, Number 2

Duffie D., Fleming M., Keane F., Nelson C., Shachar O. and van Tessel P. (2023), Dealer Capacity and U.S. Treasury Market Functionality, Federal Reserve Bank of New York Staff Report, October 2023

Eisenbach T. and Phelan G. (2022), Fragility of Safe Asset Markets, FRBNY Staff Reports, n°1026, July 2022

Favara G., Infante S. and Rezende M. (2024), Leverage Regulations and Treasury Market Participation: Evidence from Credit Line Drawdowns, SSRN

He Z., Nagel S. and Song Z. (2022), Treasury inconvenience yields during the Covid-19 crisis, Journal of Finance Economics vol 143liJermann U. (2020), Negative swap spreads and limited arbitrage, The Review of Financial Studies, Volume 33, Issue 1, January 2020

Li D., Petrasek l. and Tian M. H. (2024), Risk-Averse Dealers in a Risk-Free Market – The Role of Internal Risk Limits, Finance and Economic Discussion Series

Liang N. (2025), What’s going on in the US Treasury market, and why does it matter?, Brookings Commentary, 14 April 2025

Liang N. and Parkinson P. (2020), Enhancing liquidity of the US Treasury market under stress, Hutchins Center WP n°72

Perli R. (2025), Recent Developments in Treasury Market Liquidity and Funding Conditions, FRBNY, Remarks at the 8th Short-Term Funding Markets Conference, Federal Reserve Board, Washington, DC

[1] A repurchase agreement – a form of temporary transfer of securities – can be likened, from an economic point of view, to a secured loan (cash against securities, the value of which is subject to a discount); from the point of view of the lender of the cash, it is a reverse repo; from the point of view of the borrower, it is a repo. A repurchase agreement on a security involves a commitment to repurchase it at a later date at an agreed price. The interest rate, or repo rate, corresponds to the difference between the sale price and the repurchase price.

[2] Various regulatory requirements constrain the activities of primary dealers: the SLR leverage ratio, the specific capital surcharge for G-SIBs (the size score includes the value of Treasury portfolios on the balance sheet, while the complexity score includes securities lending/borrowing transactions), the Stress Capital Buffer (the standardised measure of counterparty risk penalises large balance sheets) and risk exposure limits (via the calculation of Value At Risk). Internal requirements in terms of profitability and risk tolerance, as well as the highly procyclical nature of clearing houses’ margin calls, may also prompt them to limit their exposures.

[3] These companies are not affiliated with banks and do not have dealer status. They employ high-frequency, automated trading strategies on their own behalf. They place a large number of orders but only hold their positions temporarily (usually for a few hours).

[4] In theory, the swap spread is positive because a swap agreement includes a larger credit risk (pertaining to the bank serving as a counterparty to the investor) than the sovereign issuer’s credit risk.

[5] According to the hedged interest-rate parity theory, there is a relationship between the interest rate differential of risk-free assets denominated in two currencies, and the difference between forward and spot exchange rates. Since 2014, this parity has no longer held true due to increased demand for currency hedging and dealers’ reduced ability to supply dollars on foreign exchange swap markets.

[6] The ‘convenience yield’ is the value that investors attribute to the liquidity and security services offered by Treasuries.

[7] According to Duffie (2025), the fact that large institutional investors such as mutual funds, pension funds and insurance companies prefer to hedge their long liabilities through derivatives (which benefit from a more advantageous accounting treatment) rather than holding long assets such as Treasuries, also contributes to increasing long-term Treasury yields compared to the implied yields of interest-rate derivatives.

[8] The SLR standard was temporarily relaxed during the Covid-19 crisis, from 1 April 2020 to 31 March 2021.

[9] The simple leverage ratio is also binding for some large institutions. However, the Dodd Frank Act introduced a permanent floor for any new capital rule. Unless Congress votes otherwise, the leverage ratio in force in 2010 cannot be reduced.

[10] In Q1 2025, the SLR requirement (plus a safety margin of 25 bps) was less stringent than the risk-weighted T1 requirement (plus a safety margin of 50 bps) for all large non-G-SIB banks (except Charles Schwab) and for 3 out of the 8 G-SIBs, and slightly more stringent (from USD 1 to 4bn) for the other 5 G-SIBs. It was more restrictive for 8 of the 9 largest deposit-taking subsidiaries of the 8 G-SIBs and 3 of the 12 largest deposit-taking subsidiaries of large non-G-SIBs.

[11] In the United States, the capital overlay imposed on G-SIBs is determined using two methods: that of the Financial Stability Board (method 1) and that of the Fed (method 2). The more stringent of the two methods (systematically the second method) is used.

[12] The regulator’s impact study includes the main deposit-taking subsidiaries of each of the groups subject to the SLR (the largest and all those with total assets of more than USD 50 bn in 2024). The USD 200 bn of Tier 1 capital ’freed up’ for depository institution subsidiaries could, under certain conditions, be reallocated within groups and support broker-dealer activity.

[13] This balance-sheet space is calculated as the difference between the maximum permitted exposure under the current eSLR requirement and under the revised eSLR requirement (plus a safety margin of 25 bps), without increasing capital (assuming full hedging against interest rate risk).

[14] Data on primary dealers’ Treasury portfolios are confidential. However, according to regulators, on average, 92% of securities recorded as trading assets on G-SIBs’ balance sheets and not held by their deposit-taking subsidiaries are recorded on the balance sheets of their brokerage subsidiaries. We have applied this proportion to the 20 groups subject to the SLR.

[15] The SLR denominator is calculated as the average of the positions booked during the quarter (on a daily basis for balance sheet exposures). We therefore exclude the average between 31 December 2024 and 31 March 2025 of outstanding deposits with the Fed and Treasuries recorded on banks’ balance sheets.

[16] The impact is comparable because reserves with the Fed and investment portfolios in Treasuries (held to maturity and available for sale) are, for the most part, recorded on the balance sheets of deposit-taking subsidiaries.

[17] With a G-SIB surcharge currently set at 4.5%, JP Morgan would see its eSLR requirement increase if option 3 were chosen.

[18] On average, over 70% of the eight G-SIBs’ Treasury portfolios were booked at their market value as “Securities Available for Sale” or “Trading Assets” in Q1 2025. However, in the event of a rise in interest rates, the unrealised losses recorded on these portfolios would reduce the risk-weighted capital ratios of large banks.

[19] VaR quantifies a bank's exposure to potential extreme losses on its market and investment positions resulting from market risks. It is highly sensitive to sudden fluctuations in market volatility.

[20] An expansion of banks’ securities portfolios would have no impact on the aggregate stock of reserves held by the Fed (the reserves used to make these purchases would be replenished by public spending enabled by the raising of Treasury debt). However, it could lead to a redistribution of reserves among banks. While both reserves and Treasuries are considered high-quality liquid assets in banking regulations, the liquidity provided by reserves is unique. They are the only asset that does not need to be monetised, and US supervisors therefore view them more favourably in the context of liquidity requirements.

[21] This strategy consists of exploiting the price difference between the cash and futures markets (‘cash-futures basis trade’). It involves buying securities on the cash market, financing this purchase by placing the securities in a repurchase agreement (repo), taking a short position on the futures market, and finally delivering the securities when the futures contract expires. To maximise their gains, hedge funds use leverage by stacking positions. However, in the event of a sudden surge in volatility in one of the market segments, as was the case in March 2020, the simultaneous unwinding of positions can spread stress and disrupt all the markets on which Treasuries are traded.

[22] The Sponsored Service of the Fixed Income Corporation (FICC) allows dealers to sponsor the ‘indirect’ membership of some of their counterparties (money market funds, hedge funds) to the FICC and to submit their repo market transactions to centralised clearing. This allows them to reduce their balance sheets.

Completed on 3 July 2025