SOFR futures have that unique expiration feature where they settle four months after the headline contract date. Does this essentially mean that, on any given day, the price of these contracts is indicating the market's expectation for SOFR rates at the actual settlement date in the future? For instance, is the price of Z23 today telling us what the market believes the SOFR price will be when it settles on March 19, 2024?
That makes no sense. Barchart is either butchering their contract specs or we're not interpreting them properly. Chapter 460 Three-Month SOFR Futures 46002. TRADING SPECIFICATIONS 46002.G. Termination of Trading Trading in an expiring contract shall terminate at the close of trading on the Business Day immediately preceding the third Wednesday of the contract delivery month. 46003. SETTLEMENT PROCEDURES 46003.A. Final Settlement Price 1. Definition of Reference Quarter For a contract for a given delivery month, the Reference Quarter shall be the interval that ends on (and does not include) the third Wednesday of the contract delivery month, and that begins on (and includes) the third Wednesday of the third calendar month preceding the contract delivery month. Example: For a hypothetical contract for which the delivery month is March 2022, the contract Reference Quarter shall start on (and shall include) the third Wednesday of December 2021 (December 15, 2021) and shall end on (and shall not include) the third Wednesday of March 2022 (March 16, 2022). 2. Definition of Final Settlement Price For a contract for a given delivery month, the Final Settlement Price shall be 100 minus compounded daily SOFR during the contract Reference Quarter (“R”) https://www.cmegroup.com/markets/interest-rates/stirs/three-month-sofr.contractSpecs.html https://www.cmegroup.com/content/dam/cmegroup/rulebook/CME/IV/400/460.pdf
There is nothing wrong with the expiration date. Different Exchanges wrote their products' Last day in various fashion. SR3 U23-CME expires on 20 Dec 2023 - this contract volume is rather low. SR3 V23-CME expires on 17 Jan 2023 - ditto SR3 X23-CME expires on 21 Feb 2023 - ditto SR3 Z23-CME expires on 20 Mar 2024 - One of the most active contracts SR3 H24-CME expires on 20 Jun 2024 - ditto You can still trade SR3 U23-CME but the volume is rather low. For comparison NG Z23 expires on 28 Nov 2023
Thank you. You seem really knowledgeable about this topic. Just a quick question: Does this mean that, on any given day, the price of these contracts reflects the market's expectation for SOFR rates at the actual settlement date in the future? To give you a hypothetical example, if traders in December were confident that the Fed would cut rates on March 18, 2024, a day before the Z contract's settlement date, would this expectation be reflected in the Z contract during the month of December itself, four months prior to settlement?
It's not a unique feature, it's interesting terminology. The "September" 3-month future SR3 U23 is actually a future with a "Reference Quarter" from September to December. It's just like any "averaging" future (e.g., Fed Funds Futures) in that the price (including final settlement) is some combination of a) history accrued up to "today" and b) expectations of the daily amounts up to contract expiration. In that sense it's different from Eurodollar futures, which were a future on one single rate: the LIBOR rate at expiry. You seem to be interpreting SOFR futures as single-rate futures, like Eurodollar futures. I'd suspect the U future has low trading volume because there's only 1 Fed meeting between now and contract expiry, and that is at the tail end of the Reference Quarter. So not much chance to change b) above. The quarterly futures will always have more volume in these products, and not surprisingly "December" is high volume, having 2 meaningful Fed meetings between now and expiry (i.e., December and January meetings).
So does a Z SOFR contract in December reflect expectations of where rates will be by (future) mid-March?
No. It reflects what the "average" SOFR rate (via the compounding formula) will be from mid-December to mid-March. Meaning the December & January Fed meetings have more impact on it than anything that happens in mid-March.