The Cash Pool That Buys the Short Paper

At the end of September the Treasury had 7.12 trillion dollars of bills outstanding, which is 22.4 percent of all marketable federal debt. At the end of 2022 the same share was 15.4 percent. The Treasury's own borrowing advisory committee recommends the bill share stay between 15 and 20 percent. It has been above the top of that band for some time, and the government is funding a growing share of itself at the shortest maturity available.

Someone has to own that paper, and the answer is mostly money market funds, which now hold 7.89 trillion dollars by the industry's count and rather more on the Federal Reserve's own flow of funds measure. That pool has grown roughly in step with the bill supply. The arrangement has worked.

The thing to notice is what used to absorb the slack. The Federal Reserve's overnight reverse repurchase facility, where cash funds park money they cannot place elsewhere, held a record 2.554 trillion dollars on the thirtieth of December 2022. On the sixth of October it held 414 million. That is not a decline. That is a facility that has gone to zero, and it was the shock absorber.

The Historical Echo

The seventeenth of September 2019 is the reference event, and it lasted about one morning.

The overnight secured financing rate, the price at which the financial system lends against Treasury collateral, rose roughly 280 basis points to 5.25 percent. There was no credit event. A corporate tax date and a large Treasury settlement drained cash from the system on the same day, and the banks that would normally have lent into that gap either could not or would not. Bank reserves at the time had fallen below about nine percent of national output, and the reverse repo facility had been running near zero since May of that year, which meant there was no visible buffer and no early warning. The Federal Reserve was buying bills within weeks.

The configuration before that morning had three parts: a heavy bill calendar, reserves that had been drained to an unknown floor, and no cash sitting idle in the backstop facility. Every one of those three is present again.

Reserves are 2.948 trillion dollars, a far larger number than in 2019, though the economy and the Treasury market are much larger too. Quantitative tightening ended on the first of December last year, and the Fed bought roughly 120 billion dollars of bills through the middle of April. The system has already shown the stress twice in a mild form: overnight financing printed 22 basis points above the Fed's own standing facility rate on the thirty first of October last year and twelve basis points above it on the thirty first of December. Both were month ends. Neither has recurred this year.

Where Patient Capital Is Positioning

The near term calendar is the part that is knowable. The Treasury's general account held 948.7 billion dollars, and the government plans about 628 billion dollars of net borrowing between October and December with its coupon auction sizes effectively pinned. Pinned coupons and a large financing need mean the bills do the work, which means several hundred billion dollars of additional short paper has to find a home in a cash pool that no longer has a backstop balance to draw down.

For long horizon capital this is not a trade. It is a reason to be careful about the assumption that sits underneath nearly every other position: that cash is costless to hold and instantly available at par. Both episodes above were one day events in the plumbing, invisible to anyone not funding a balance sheet overnight, and both were resolved by the central bank within weeks.

What deserves attention is the structural drift rather than the next month end. The federal government has shifted a larger share of its obligations to the shortest maturity it issues, which means it reprices the whole of that share every few months at whatever rate prevails. That is the same choice a household makes in moving from a thirty year mortgage to a variable rate one: a lower known cost in exchange for taking the interest rate risk onto its own books. Hard assets do not care what the overnight rate does on any given morning. They care a great deal about a sovereign whose average maturity is shortening into a persistent deficit, because that is the configuration in which the pressure to accept a higher rate of inflation becomes an arithmetic convenience rather than a policy error.

The Capital Memo