Invest1 publisher2 min readPublished
Treasury basis trade's 20% shrink to $1.2 trillion fits a quiet exit or a forced one
Morgan Stanley estimates hedge funds' Treasury cash-futures basis trade has fallen 20% this year to about $1.2 trillion. Funds on rolling repo can shrink by letting positions lapse or by being made to sell, and the position figure looks the same either way.
The Investor · Invest desk
Drafted by a language model from the sources cited here and checked against its claim ledger before publication. How we use AISend a correction
What happened
- Hedge funds in the trade buy Treasury securities, sell futures against them to collect a small pricing gap, and borrow most of the purchase money through repo.
- Funding through overnight repo means a fund has to renew or replace its loan every time it wants to keep a position open.
- In CryptoSlate's illustration, a $100 million position earning 0.2% a year makes $200,000, a 4% return when the fund puts up only $5 million of its own.
- A repo lender advancing $98 against $100 of bonds sets a 2% haircut, and raising it to 4% doubles the cash the fund must supply against the same collateral.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- cost A 0.2-point rise in repo cost on $95 million of borrowing hands $190,000 of a $200,000 profit to the lender, cutting the fund's return on its own $5 million to 0.2%.
- constraint A fully hedged fund can still run short of cash, since a futures loss is called as variation margin while the matching gain on the bond stays locked in the security.
- exposure About $1.2 trillion of Treasury holdings sits with buyers who stay only while the trade pays, and whether it pays depends on repo terms the government does not set.
If Morgan Stanley's estimate is right, a 20% drop that lands at $1.2 trillion means the book started the year near $1.5 trillion. Roughly $300 billion of positions have already come off [1][1]. Three readings fit that figure. In the first, the spread narrowed or funding got dearer, and funds simply stopped replacing positions as they expired, something CryptoSlate notes needs no crisis [8]. In the second, repo lenders raised haircuts or futures losses were called in cash, and some positions were sold on a lender's schedule [9][10]. The third is a mix: voluntary lapses so far, forced sales still ahead.
I think the first reading fits the evidence better for now. Morgan Stanley had not found evidence of broad basis-related market stress when the estimate was reported [2]. At the illustration's 20 times leverage [2], a small rise in funding cost can erase most of the return without any margin call, and a fund in that position has little reason to renew. The counter-thesis is that a forced unwind would produce the same position figure. CryptoSlate is plain that a decline alone cannot show whether funds are exiting calmly or being forced to sell [3].
Haircuts are where the two readings part. Suppose the whole $1.2 trillion were financed at the illustration's 2% haircut. A lender move to 4% would then ask funds for about $24 billion more of their own cash against the same bonds, before any futures margin [4][10]. The $24 billion applies the illustration to the whole book; CryptoSlate does not report the haircuts funds actually pay.
The test is in prices. Funds closing together sell bonds to repay loans and buy futures to close shorts, and those trades can push bond prices down relative to futures, raising exit costs for whoever is still in [11]. If bonds cheapen against futures while repo costs climb, I am wrong, and the $300 billion was the first leg of a forced unwind [1]. If the basis holds steady while positions keep shrinking, the funds were moving their money to something that paid better [8].
What to watch
- Repo haircuts and overnight funding costs on Treasury collateral: a lender move like the illustration's 2% to 4% would turn lapsing positions into forced sales.
- Whether cash Treasuries cheapen against futures as funds sell bonds and buy back their shorts at the same time.
- Morgan Stanley's next estimate of basis positions, and whether it then finds basis-related market stress.