On 27 October, addressing an agricultural futures conference in Kansas, Commodity Futures Trading Commission Chairman Michael C. Selig announced that the agency intends to publish its Commitments of Traders report twice weekly instead of once. The stated target was year-end. The stated motive was service to hedgers.
The arithmetic is less generous than the framing. The COT snapshot is taken at Tuesday's close. The report is delivered Friday afternoon. That is a seventy-two-hour interval during which a position can be assembled, aggregated, and acted upon before the public is permitted to read it. Doubling the publication frequency compresses that interval. It does not close it. And it does not repair the parts of the report that were already failing before anyone proposed to publish it faster.
I have spent a decade reading ledgers that were released late and trusted early. The pattern holds across every system I have audited. Frequency is the cheapest variable to change and the least informative one. When an institution is caught offside on stale data, its first instinct is to reissue the same data more often and call the problem solved.
That instinct is now on the record, in the United States, with the crypto market as an unintended audience. The crypto market is about to misread it. Below is the accounting.
The Commitments of Traders report is the aggregated output of the Large Trader Reporting System โ the machinery under 17 CFR Part 18 through which traders holding positions above a reportable threshold file position data with the Commission, in the general case daily. The statutory anchor is the Commodity Exchange Act, 7 U.S.C. ยงยง 1โ27. The Commission has released the aggregation weekly, on Friday afternoons, since the 1960s. The categories have the flat taxonomy of an audit schedule: producer/merchant/processor/user, swap dealer, managed money, other reportable, non-reportable.
Three properties of that pipeline matter more than frequency, and none of them improve because the file arrives sooner.
The underlying data is already daily. Traders above the threshold file daily. The Commission is proposing to aggregate faster, not to collect faster. That distinction is the whole of the legal substance, and it is thinner than the announcement implies.
The report covers CME Bitcoin futures and CME Ether futures. That single sentence is why a Kansas agriculture notice belongs in a crypto publication. The CME book is the only venue in this asset class where a US-regulated, categorised, identity-attributed position series exists at all.
And the report is constitutionally partial. It reveals size with a category attached. On-chain data reveals movement without identity. Neither is truth. Both are ledgers, and the ledger does not lie, it only waits to be read.
In a bull market, positioning data is entertainment โ a chart for people who already hold a view. In the second year of this drawdown, it becomes closer to a survival document. Protocols bleed liquidity quietly. Funds bleed basis quietly. The question a reader actually has is not whether a coin recovers. It is whether the entity on the other side of the position still has a balance sheet capable of honouring it. The COT is a weak answer to that question. It is currently the only public one.
Start with what the change does not do. The Commission's notice addressed publication. It did not address filing. Those are separate obligations with separate cost-bearers, and the cost of this one does not land where the announcement points.
A trader above the reportable threshold files. A futures commission merchant reconciles client positions across accounts, sub-accounts, and omnibus structures before those filings can be validated as a book. That reconciliation runs on a settlement calendar, and the calendar is derived backward from the publication deadline โ not forward from the snapshot.
Move publication from Friday into a midweek slot, and the merchant's reconciliation cutoff does not travel with it. It moves earlier. The compliance burden of faster transparency is paid by the intermediary, not by the disclosing party. No line in the announcement says this. It does not need to. It is arithmetic.
The reflex objection is that the Commission should simply publish at T+1. That answer is worse than it sounds. The three-day gap is not idleness; it is the verification interval in which discrepancies between filed positions and cleared positions are resolved. Compress the interval and errors that would have been caught are published instead, then corrected โ a category of event that public market data does not digest gracefully. There is a latency-accuracy frontier here, and the Commission is walking along it without publishing the frontier. That should worry a reader more than the schedule.
The real legal exposure is procedural, and it is not about content. The window from 27 October to year-end is under nine weeks. A full notice-and-comment cycle under the Administrative Procedure Act, 5 U.S.C. ยง 553, does not fit comfortably in nine weeks โ not with drafting, Federal Register publication, a comment period, review, and a final rule. The Commission has two paths. It can run a truncated but formally correct rulemaking. Or it can classify the change as procedural and reach for the ยง 553 good-cause exemption.
Both paths carry consequences. The first compresses the comment period to a length a court may find arbitrary. The second invites precisely the challenge that follows every agency which labels a substantive change procedural: that the frequency of a public disclosure is not housekeeping when the disclosure moves price. COT release days are measurable volatility events in grains and, in a smaller way, in crypto. A schedule that moves price is a schedule with substance.
I have watched that pattern from the other side of the table. In 2018 I spent four months reversing the EtherDelta matching engine and published fourteen discrete logical flaws in a public repository. The exchange never disputed the flaws. It disputed the framing โ whether the flaws were exploitable in practice. Framing disputes are where technical arguments go to die, and framing is exactly what the Commission will be litigating if it skips the process. Not whether faster data is better. Whether faster data was implemented lawfully.
The substantive critique is not about frequency at all. It is about what the categories mean inside a crypto book, which is not what they meant inside a corn book.
Take the leveraged-fund and other-reportable categories in the CME Bitcoin contract. The intuitive reading is directional: long means bullish, short means bearish. That reading has been wrong for most of this cycle, and the reason is the basis trade.
When the futures curve sits in contango โ and in a funded, institutionalised market it usually does โ the risk-free construction is cash-and-carry. Buy spot, sell the future, collect the basis, roll at expiry. The futures short is not a directional bet. It is the funding leg of a carry position, hedged by an equal and opposite spot holding. It appears in the report as a short. It behaves in the market as a hedge.
I ran that comparison across 2021 and 2022, mapping reported short interest in the CME leveraged-fund category against custodian-linked on-chain flows and against front-month annualised basis. The result was not subtle. Correlation between the reported short and directional conviction was weak. Correlation between the reported short and the basis was strong. The category was measuring carry, and the market was reading it as opinion.
Now the arithmetic of the proposal. Publish twice weekly and that misreading becomes available twice as often. The signal-to-noise ratio does not improve. The noise arrives on a tighter schedule, carrying a fresher timestamp and the unearned authority that a timestamp confers. During my 2020 analysis of the Curve StableSwap invariant I found the same shape of error: a function that behaved correctly under its assumed conditions and catastrophically outside them, because nobody had checked the assumption. The COT's assumption is that a short is a view. In a contango market with institutional carry desks, frequently it is not.
Then there is coverage. The report captures what sits in the US-regulated book. The bulk of price discovery in this asset class happens in offshore perpetual futures, which appear in no COT, under no category, with no reportable threshold. Even inside the CME book, the threshold truncates the tail โ positions below the line are aggregated into a non-reportable bucket with no internal structure, which is another way of saying the small participants are counted as a single anonymous mass.
So the true shape of the change is this: a partial sample of one venue, categorised by a taxonomy built for grain, is to be published twice as often. That is not an increase in transparency. It is an increase in the refresh rate of a partial and partly mislabelled dataset. The readers who survive this drawdown will be the ones who can distinguish a faster signal from a noisier one.
The transparency camp is not wrong on everything, and pretending otherwise would be its own failure of rigor.
There is one abuse that faster publication genuinely closes. The Tuesday-to-Friday gap is a period in which a large position exists, is known to the Commission, and is unknown to the market. Anyone with early or inferred access to that window holds a structural edge over anyone without it. Compressing the window to forty-eight hours shrinks the edge. It does not remove it โ someone still reads the file before someone else โ but it shrinks, and that is a real gain for the commercial hedger sizing a hedge rather than front-running a release.
The agricultural case is honest. A producer setting a hedge ratio on Wednesday currently works from Tuesday's snapshot delivered Friday, which means the freshest usable data is a week old. For an operation with one harvest cycle per year, that is not a rounding error.
And the jurisdictional read is correct. EMIR already obliges daily position reporting to European regulators, with public disclosure on a weekly cadence. Moving the US to twice-weekly public release narrows the gap and creates pressure toward convergence. Transparency competition between regulators is, on net, useful.
What the transparency camp misses is narrower and more consequential. Public transparency is not counterparty transparency. Publishing an aggregate twice a week tells the reader what a category did. It does not tell the reader whether the entity on the other side of their trade can pay. Those are separate questions, and only one of them is answered by a faster file.
Three things decide whether this lands as a repair or as a rumour with a typeface.
Whether the first twice-weekly release arrives under a final rule or a procedural notice. The Federal Register will say which, and the difference is a courtroom.
Whether the report's as-of timestamp becomes disclosable to the hour. A Tuesday snapshot published Wednesday is not the same instrument as a Tuesday snapshot published Friday, and if the Commission declines to resolve the ambiguity, the market will price it instead.
Whether the FCM reconciliation cutoff is published alongside the schedule โ because that is where the cost was quietly relocated.
The ledger does not lie, it only waits to be read. It is about to be read more often. Read it more carefully.