It’s no secret that the United States has a lot of debt to finance. If you’re on Truth Social or listen to long-only equity managers, you also know that interest rates are high – relative to the last decade. Issuing on the long end of the curve right now would lock in a higher cost of borrowing for the maturity of the issuance. Instead, the Treasury can tilt issuance toward short-term bills. The government can continue to pay its liabilities then finance the current debt when interest rates are appropriately low. This maneuver, however, carries a rollover risk if long-end bond yields remain stubbornly elevated.

At the same time, the Treasury Secretary needs to find buyers for all this debt to finance. As national external wholesaler in chief, Secretary Bessent’s job is to find someone willing to hold US debt with the looming promise of supply increases. The two answers he’s come up with are stablecoins and banks. To encourage large banks to hold more Treasury securities on balance sheets, the Fed has proposed changes to the supplementary leverage ratio (SLR). For further details, please refer to my writing from April 22.

In that piece, I mentioned the risk of increasing systemic vulnerability and leverage in money markets. To understand where that risk may come from, we need some context. Before the Great Financial Crisis (GFC), banks served a larger role in financial intermediation. Starting in the 1980s, banks started underwriting the rapidly growing derivatives market. Each trade earned a small marginal fee, so banks held increasingly large levels of leverage on their balance sheets. By 2008, the notional exposure to the derivatives market was $176 trillion while maintaining 48x balance-sheet leverage. After the GFC, new banking regulations – including the SLR – increased the cost of expanding balance sheets and rendered market-making activity unprofitable for the large banks. Hedge funds have stepped in to fill the role of highly levered relative value arbitrage traders.

In today’s basis trade ecosystem, many hedge funds buy Treasuries, finance them in bilateral repo, and short the matched futures or pay-fixed swaps, locking in the cash–derivative spread. Banks supply the leverage and collateral transformation but leave the market risk with the fund. In May of this year, Lina Lu and Jonathan Wallen published a great paper analyzing the practices of banks in the repo market. They found that banks are frequently lending against US Treasury collateral at a negative haircut. This means that – for example – they’ll accept $100 of collateral for $101 in cash. For the banks, this pricing allows them to effectively bundle two loans together: a secured leg equal to the value of the collateral and an unsecured leg that represents the remainder. This regulatory arbitrage lets banks execute riskier loans that are nominally collateralized by safe Treasury assets without having to allocate expensive equity on the balance sheet. However, the banks are taking on credit counterparty risk. If the non-bank financial institutional borrower defaults, the unsecured portion of the loan enters the general pool of ordinary, unsecured claims.

This practice is correlated with a bank’s ability to function as a security warehouse. ‘Warehouseness’ can be thought of as the degree to which a bank is a net lender in the repo market; an adaptive balance sheet lets a bank hold Treasuries through volatility and fire sales. By selling futures or paying the fixed leg of a swap, warehouse banks can hedge the interest rate risk of holding Treasury securities and earn positive arbitrage carry by renting balance sheet capacity.

The two recent policy initiatives are going to push haircuts further negative thus increasing leverage in the non-bank financial intermediaries (NBFI) sector. First, the deluge of bill issuance is going to cheapen the cost of collateral in the market. This cheapening occurs because of both Econ 101 and financial processes. The easy answer is price goes down when supply goes up, so repo lenders can’t demand as relatively high a haircut when the collateral is less expensive to obtain. Financially, the value-at-risk model used to calculate haircuts are a function of dollar losses. As yields increase – and mathematically the price decreases – the dollar duration of a bond decreases. A less expensive bond has less to fall than a more expensive bond.

A counter to this vector of haircut decrease would be the loss in repo liquidity from money market funds (MMF). MMFs tend to be more active participants in the repo market when bills are scarce. The rate of banks’ lending, however, is a function of balance sheet capacity and liquidity of the underlying collateral. The cost of allocating scarce balance sheet equity toward lending is more expensive than the changes in supply dynamics from MMFs.

This leads us to the second change in governance: the lifting of the SLR. At the current 5% level for the largest banks, each $100 billion lent in the repo market consumes $5 billion in balance sheet equity. Changing the SLR gives banks more room on their balance sheet to extend leverage to NBFIs. In recent history, the Fed issued an exemption from the SLR limits during COVID. As the exemption took effect in the market, the Fed’s FR 2052a filings show the share of negative-haircut reverse-repo trades jumping from ~10 % (Mar-2020) to >60 % (Dec-2020). Lu and Wallen estimate that a 1% decrease in the SLR leads to an 18 bp decrease in haircuts.

Systemically, the move in haircuts increases the leverage in the NBFI system and has secondary impacts on the liquidity of the underlying market for US Treasuries. Just as an increased capacity for warehousing is associated with extending leverage, warehouse dealers are more prone to cyclicality in lending practices. If yields fall too far or balance-sheet costs tighten, the warehouse dealers pull back funding and revert to positive haircuts, triggering a pro-cyclical unwind in leveraged positions. The fire sale dynamics can dry up liquidity and force the Fed to re-engage in market supporting activities.

As regulatory policy makes balance sheet space cheaper and haircuts continue sliding below zero, NBFIs can leverage their activity through the repo market. The extra liquidity doesn’t just sit in the Treasury market. Risk assets – especially tech stocks and crypto – can get a price lift as buyers bid faster in larger lots. However, this trade is dependent on the rising tide in the repo market, but when liquidity dries up, the positions are quickly unwound. We see a short-term boost for risk assets as leverage percolates through the system.

Lu, L. & Wallen, J. (2025). Negative Treasury Haircuts. Available at SSRN: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5239611

Steven J. Wagner, Investment Adviser
Bray Farm Income Advisory LLC
3375 Brookdale Drive, Pittsburgh PA 15241
412.504.9412
412.848.2410 (cell)

Download PDF