Quant Memo
Core

Month-End and Quarter-End Repo Pressure

Repo rates reliably jump on the last day of the month, and especially the last day of the quarter, because banks temporarily shrink their balance sheets to look better on the regulatory reports that snapshot exactly that date.

Prerequisites: How SOFR Is Calculated From Repo Transactions, Procyclical Haircuts and the Collateral Multiplier

Repo rates in the US have a strange, predictable pattern: they are usually well-behaved for weeks at a time, then spike — sometimes sharply — on the very last business day of a quarter, before dropping right back the next morning. Nothing about the demand for cash or Treasuries has fundamentally changed overnight. What changed is that banks' balance sheets get photographed on that specific date.

Bank capital and leverage-ratio requirements are measured as of a single reporting date, so banks temporarily shrink low-margin, balance-sheet-heavy businesses like repo intermediation right around quarter-end to look better in that snapshot — pulling supply out of the market exactly when demand for financing hasn't gone anywhere.

Why the balance sheet snapshot matters

The Supplementary Leverage Ratio and similar capital rules don't care about a bank's average balance-sheet size over the quarter — many are measured, or at least highlighted, based on the level on the last day. Repo intermediation is a classic balance-sheet-heavy, low-margin business: a dealer borrows cash from one client and lends it to another, holding both a repo and a reverse repo on its books, which counts as leverage-ratio exposure even though the position is close to riskless. Right before the reporting date, banks pull back from this business to shrink their measured balance sheet, reducing the amount of cash they're willing to intermediate in repo. Cash-rich lenders — money-market funds especially — still need somewhere to park cash, but the usual channel to a dealer's balance sheet has narrowed, so rates on the remaining capacity spike.

Q-end spike Q-end spike
The rate reverts almost immediately once the reporting date has passed, which is the tell that the move is a balance-sheet-reporting artifact, not a shift in underlying funding demand.

Worked example

GC repo trades at 5.30% for most of a quarter. On the last business day, dealer capacity to intermediate repo shrinks sharply as balance sheets pull in for the leverage-ratio snapshot; the same volume of cash chasing less available intermediation pushes the overnight rate to 5.75%, a jump of 45 basis points for one night. A fund with $1 billion in overnight repo earns an extra 1,000,000,000×0.0045/360=12,5001{,}000{,}000{,}000 \times 0.0045/360 = 12{,}500, i.e. $12,500 that single night versus a normal day — and the rate is back at 5.30% the next morning once the snapshot has passed and dealers resume normal intermediation.

What this means in practice

Money-market desks that manage cash around these dates try to time maturities to either avoid needing to reinvest exactly on quarter-end, or deliberately seek it out for the extra yield if their cash isn't rate-sensitive that day. The Fed's own facilities — the Standing Repo Facility and the overnight reverse repo facility — were designed partly with these predictable pressure points in mind, giving the market a backstop precisely when private balance-sheet capacity is most likely to disappear.

Don't mistake a quarter-end repo spike for a sign of systemic funding stress the way a spike driven by, say, a bank failure would be. The two can look similar on a chart, but the quarter-end version is calendar-driven, known in advance, and self-reverses within a day.

Related concepts

Practice in interviews

Further reading

  • Federal Reserve Bank of New York Liberty Street Economics, 'Reserves and Repo Market Pressure'
  • Correa, Du & Liao, 'U.S. Banks and Global Liquidity'
ShareTwitterLinkedIn