LIBOR Transition Part 1.5: SOFR and the Debt Ceiling
Ksusha McCormick
October 27th, 2021
At the end of my last blog post, I promised to follow up with a discussion of potential solutions for those unhappy about the fact that switching from Libor to SOFR means swapping credit sensitivity for repo rate sensitivity. Part Two is still in the works, but the recent battle over the debt ceiling provides such a timely distillation of the issues at play that it merits its own comment.
Tempest in a Teapot?
Before we get into that, however, it might be helpful to start with a memory jogging exercise. Think back to the fall of 2019, specifically September 2019. Pre-covid. Pre-presidential election. Is there anything particularly interesting that you recall happening in the markets during that time? If not, you may be surprised to hear that seemingly out of the blue, the overnight repo rate experienced a massive spike higher. On September 17, 2019, one unlucky guy or gal paid an eye-watering 10% to borrow dollars (on a loan secured by AAA-rated Treasury bonds!). SOFR, which is a volume-weighted median derived from the general collateral repo rate, jumped to 5.25%, almost 300bps higher than the previous day.
Why did this happen? There are several reasons, some structural, some specific to the market dynamics on that particular day. Corporate taxes were due on September 16th. Corporates withdrew dollars from money market funds in order to make these payments. To stay ahead of redemptions, money market funds, in turn, became more reluctant to lend their normally plentiful dollars. At the same time, a relatively large amount of Treasury debt was due to settle on the 16 th as well. Primary dealers, who are obligated to buy Treasuries at auction, were as a result longer than normal and tried to lend those bonds / borrow dollars in larger-than-normal quantities. Regulation restricted banks’ ability to use excess reserves to satisfy the surge in borrowing demand. Clearing requirements made participants sensitive to counterparty risk and therefore more reluctant to offer term repo as an alternative to overnight.
The So-What
I wouldn’t blame anyone for thinking that this all sounds like a bunch of mumbo-jumbo, but that’s precisely the point. There are four main takeaways from this strange occurrence: It had nothing to do with movements in interest rate expectations. It had nothing to do with credit spreads. It was caused by a sudden need for approximately $150B of cash, which is only about 10% of the normal daily volume in the repo market. And the dislocation was only resolved by rapid and aggressive intervention by the Fed.
Was September 2019 really such a big deal, though? It was just a one-day event. And the Fed intervened! These are certainly reasonable points. However, it is worth considering that if these events were a mere curiosity, the upcoming debt ceiling showdown may result in all of these forces playing out on a much larger scale.
Debt Ceiling Duality
The debt ceiling resolution, if it really will be a resolution, is at present a few weeks away, which is quite a while in market terms. The SOFR-related news du jour is actually a small but unexpected drop in SOFR, due to (you guessed it!) repo market-specific reasons. GSEs are lending out more cash than usual. People are shorting 2-year Treasuries because they’re positioning for an earlier Fed hike. 2-year repo has gone “special,” which means borrowing rates have gone negative. Negative rates? Aren’t we worried about inflation?? Another one-off, you may say. And the Fed has been preventing negative rate prints with a massive reverse repo backstop!
Predicting what will happen when Congress re-starts the debt ceiling negotiation is not Decameron Technologies’ strong suit. However, if the US government really defaults, how will the repo market react? This is surprisingly tricky to answer! Absent Fed intervention, if people decide to hoard dollars, because their Treasury bonds have stopped making coupon payments and are therefore less attractive, repo rates will spike higher (if almost everyone prefers to hang on to dollars instead of lending them out, people who want to borrow cash / lend bonds will have to pay up). If people lose faith in US Treasuries and everyone shorts them, then repo may go deeply negative (in order to short bonds, you have to borrow them, and if almost everyone rushes to borrow bonds / lend dollars, they may become willing to lend cash at negative rates). Wait a minute. Is it really possible that the same macro story, i.e., people not wanting Treasuries, can have two diametrically opposed results? Welcome to the repo market!
But SOFR calculation excludes specials, there’s now a standing repo facility, and the Fed will intervene, so none of this is a problem, right? Sure, the Fed can do repo and reverse repo in potentially unlimited quantities, and for the sake of argument let’s assume that on the day the US defaults, the opaque, human-driven intervention machine will function perfectly, and that despite the busy day, the repo market’s needs will be top of mind for Chairman Powell. The problem as it relates to the derivation of SOFR from repo is that SOFR calculation methodology specifically excludes transactions in which the Fed is a counterparty. It has to, in order to comply with Iosco principles. In a truly cataclysmic scenario, where the market seizes up and the Fed has to step in and act as the only borrower or lender left standing, there may not be enough transactions to calculate SOFR at all, at which point the Fed will resort to either a dealer survey or, if that fails, too, the previous day’s rate. This may all sound far-fetched, but those of us who lived through 2008 recall that the largest and most liquid markets experienced some of the greatest dislocations.
Behind the Curve
At the end of the day, when the central bank exercises a lot of control over the yield curve, yields are less reflective of both rate and credit risk than they would be if driven purely by market forces. However, that doesn’t mean that rate and credit risk disappears from the real economy. To the contrary, it becomes that much more important for a market participant to thoughtfully choose an instrument with the right set of exposures, or the right hedge to protect against unwanted exposures. Hedging GC repo risk? That’s a rhetorical question, but for the bond geeks reading this, I’ll add that the thought of it reminds me of a story I heard in 2008, about an IO trader who, when asked how to hedge IOs, replied, “with Valium.”
Debt ceiling or debt floor, if you are a bank who just wants the floating rate on the loans you make to be reflective of the credit environment; if you are an insurance company that needs to match assets with long-dated liabilities; if you’re an asset manager who wants to hedge out the interest rate risk on part of your portfolio, you may be excused for wishing you could say “Who cares?” rather than consider the technical quirks in repo that spill over into SOFR.
"Antique ceiling fan, Tovrea Castle machine shop shed." by Eileen M. Kane is licensed under CC BY-SA 2.0
