What the Commitments of Traders report is
The Commodity Futures Trading Commission (CFTC) requires large traders in US futures markets to report their open positions each week. Positions are snapshotted every Tuesday and published the following Friday afternoon as the Commitments of Traders report, or COT. The result is a weekly, standardized view of how different categories of market participants are positioned across the major US futures contracts. It is one of the few positioning datasets that is public, consistent, and long-running enough to support historical percentile analysis.
Since 2009 the CFTC has published the Disaggregated COT, which splits what used to be a single “non-commercial” bucket into more analytically useful categories. Four of them are surfaced in most published research:
- Producer / Merchant / Processor / User, or commercials: entities that produce, handle, or consume the physical commodity and use futures primarily to hedge price exposure.
- Swap Dealers: financial intermediaries that manage commodity swaps books and hedge them in futures.
- Managed Money: commodity trading advisors, hedge funds, and commodity pools that take systematic directional views.
- Other Reportables: large reportable traders not fitting the above categories.
This chart focuses on the Managed Money category, which is the standard shorthand for the speculative community’s directional posture.
Why managed money positioning matters
Managed money participants are not in the business of producing or consuming the underlying commodity. They are expressing a directional view, systematically, with size that is reportable to the CFTC. When managed money is heavily net long a commodity, the speculative community is collectively bullish. When heavily net short, collectively bearish. The size and direction of the aggregate position tell you what the speculative community is already doing.
Extreme readings are informative but ambiguous. A net long that reaches the 95th percentile over the trailing five years means the speculative community is as long as it has been in half a decade. That reading has been followed historically by both continued price strength on momentum extension and by exhaustion reversals. The academic and practitioner literature is mixed on which interpretation dominates, and it is very likely commodity-specific and regime-specific. MSCIP does not make a directional claim from an extreme reading. The chart shows where positioning sits; the interpretation is left to the reader.
Percentile ranks and how to read them
Rather than leaning on raw contract counts alone, the chart ranks current positioning against its own history. The headline percentile is computed over the full twenty-year weekly record for each contract, which makes readings comparable across commodities: a 92nd percentile long in copper and a 92nd percentile long in gold are directly comparable statements about how unusual current positioning is within each contract’s own long-run range. The chart draws the 90th and 10th percentile levels as threshold lines, so the current reading can be judged against the historical edges at a glance. A rolling five-year z-score is shown alongside, which answers a different question: how stretched is positioning relative to the recent regime, rather than the full record. A reading between the thresholds is unremarkable. A reading at or beyond them is at the historical edge for that commodity.
Commercials divergence
When managed money is at an extreme long and commercials are at an extreme short at the same time, the two sides of the market are aggressively opposed. This configuration is called a “bearish divergence” in the traditional COT literature because commercials are, on average, better informed about physical supply and demand than speculators. The label is convention, not causation: divergence does not force a reversal, and the timing of any resolution is not signalled by the divergence itself. What the divergence does is quantify the crowding of the speculative side against the commercial side, which is a useful state variable to know before interpreting price action. The precise scoring behind the divergence flag is documented on the methodology page.
What the chart shows and does not show
For each contract on the chart, the display carries the weekly managed money net position, the corresponding percentile rank in the trailing five years, the week-over-week change, and a flag for bearish divergence when both sides are simultaneously extreme.
What the chart does not do: it does not predict prices. Positioning is descriptive of the current state of the market, not predictive of the next move. Historical studies of positioning as a systematic signal produce mixed results and are highly regime- and horizon-dependent. Read the chart as context that informs interpretation of other data, not as a standalone entry or exit signal.