LIQUIDITY TRAJECTORY

CFTC Report Date: 2026-09-08 | Generated: 2026-09-11 15:50 ET

EXECUTIVE SUMMARY

• Duration is the board’s stress center and it deepened. UST 10Y dealers pushed further into EXTREME SHORT (z -2.03 to -2.29), the 0.0th percentile of the two-year range, on a fourth straight week of net decline (WoW -47,349, net -634,446). This is the one dealer extreme on the board carrying options corroboration rather than a caveat (OptZ=-1.07), so the options-book read has support this week. Measured realized vol in this regime runs 1.00x the all-week average (n=57 overlapping weeks), so no vol amplification is evident in this market’s history. Measured asset managers hold the other side at the 94th percentile (AM z +1.65).
• SOFR 3M remains the board’s only both-sides extreme and both sides widened. Dealers sit at the 99.0th percentile on an outright +3,693,025 contract long (WoW +313,707, flow z +1.59^) against leveraged funds at the 1.9th percentile, whose regime slipped a tier from ELEVATED to EXTREME SHORT (z -1.97 to -2.16). Measured asset managers are washed out at the 0th percentile (AM z -2.23), 39% of leveraged gross exposure is spread structure, and the options-implied dealer delta carries a caveat (OptZ=-1.77), so read the futures extreme as a basis or directional footprint, not gamma.
• Copper produced the week’s cleanest commodity development: a regime step to ELEVATED SHORT for Commercials (z -1.23 to -1.78, 1.9th percentile) on the third-largest relative flow on the board (flow z -2.29^), against Managed Money stepping up to ELEVATED LONG (z +1.41 to +1.77). That is a full opposed-extremes book with a concentration flag (28L/35S #), and it is positioning state only.
• The yen book reset both cohorts on the board’s largest flow (flow z -6.88^, WoW -118,798). Dealers crossed the zero line into net short yen (-39,477) and leveraged funds released an ELEVATED SHORT back to NEUTRAL (z -1.73 to -0.29). Elsewhere in FX the dollar story held its shape: USD Index dealers stayed ELEVATED SHORT (z -1.90, 5.8th percentile) on a one-sided, concentration-flagged ledger (top-4 64.7%, 7L/0S #) with measured realized vol at 1.18x the all-week average (n=35), the highest measured realized-vol ratio on the board, and the buy-side long DXY.
• Catalyst line: August CPI printed Wednesday Sep 9 and inflation failed to cool, and the FOMC decision lands Sep 17, six days out, directly on the SOFR, 2Y and 5Y books that already carry the board’s crowding. PCE follows Sep 25, fourteen days out, on the same front-end complex. NFP sits outside the two-week window on Oct 2.

SINCE THE SNAPSHOT

This brief reflects positioning as of the 2026-09-08 Tuesday snapshot; the price moves below are current through Friday September 11 and run on a different clock. Price action does not confirm any repositioning.

August CPI was released Wednesday and inflation did not cool, and the tape since the snapshot has traded a rate-hike repricing on top of an oil spike: the whole curve sold off with the long end worst, small caps slumped hardest of the majors, the dollar firmed and crude ran through 100. All of this is price action after the Tuesday snapshot; none of it confirms dealer, leveraged fund, asset manager, Commercial or Managed Money repositioning. The next positioning data is not visible until next week’s report.

POSITIONING TABLE

POSITIONING DYNAMICS

Rates (UST 2Y, UST 5Y, UST 10Y, UST 30Y, SOFR 3M)

The curve is the only place on the board where both sides of the book are stretched at once, and this week the stress migrated outward rather than releasing. The front end is where the crowding lives: leveraged funds hold the top of their own range across 2Y, 5Y and 30Y while dealers sit below theirs, and the SOFR contract carries both cohorts at outright extremes simultaneously. Read that as maximum counterparty tension into the Fed rather than as conviction about the level of rates; in every one of those books the leveraged position is still net short in absolute terms, so a “crowded long” here means crowded relative to their own history, not an outright bullish duration bet.

Two caveats travel with the rates complex unchanged from last week and they still bind. First, the leveraged spread share in SOFR is large enough that the directional read overstates conviction; the same basis and calendar-spread flow that produces a dealer footprint at the extremes produces no directional view at all. Second, the options-implied dealer delta in SOFR is not stretched, so the futures extreme there reads as a basis or directional footprint and the options-hedging hypothesis has no support. The 10Y is the exception on both counts: its options-implied delta is stretched the same way the futures book is, which is the only place this week where the gamma hypothesis has corroboration rather than a caveat, and even there the measured realized-vol ratio shows no amplification. The 5Y is the tenor to watch for the standoff resolving, because dealers and leveraged funds are both extending in the same direction there, which compresses the counterparty tension that normally drives sharp unwinds.

FX (USD Index, EUR, JPY, GBP, AUD, CAD, CHF, MXN)

The dollar story is coherent and it has not changed shape. The buy-side holds the Dollar Index long while running short the euro through the leveraged book, and dealers absorb the other side at a stretched short in DXY; those three facts tell one consistent bullish-dollar positioning story rather than three separate ones. The Dollar Index is also the board’s highest measured realized-vol ratio, and the ledger behind it is thin and one-sided enough that the concentration flag matters more than the z-score: a handful of traders carry the dominant side. Prior ELEVATED SHORT episodes in the Dollar Index ran a median 4-week return of -0.3% with 0 of 5 bullish, against an unconditional baseline of +0.1% median and 52% bullish across all weeks (n=885). Five episodes is not a sample to lean on; treat the skew as historical context, not a forecast.

The yen was the week’s real FX event. The repositioning was the largest relative flow anywhere on the board and it moved both cohorts at once: dealers crossed the zero line and leveraged funds abandoned an elevated short back to flat. That is a structural reset of the book rather than a directional statement, and with the Bank of Japan on the calendar chatter alongside the Fed, it leaves the yen with the least stale positioning of any major. The Canadian dollar produced a smaller version of the same thing, a regime crossing on outsized flow into a fourth straight week of dealer net decline. The peso stays the second opposed-extremes book, with leveraged funds actively extending a crowded long against a dealer short that keeps eroding, and its prior episodes carry a mild positive skew versus baseline that the sample size does not support leaning on. Sterling and the franc both cleared the gate on cohort crossings without level: leveraged funds released a long in sterling and deepened a short in the franc, neither from a stretched base. The Australian dollar sits quiet with both cohorts leaning the same way.

Equities (S&P 500, Nasdaq, Russell 2000)

Equity positioning normalized in both directions this week. S&P dealers crossed up out of neutral on outsized flow with no calendar catalyst behind it, while Nasdaq dealers crossed back down into neutral from the short side; the two indices are now moving toward each other rather than apart, which drains the sector-rotation divergence that has been the equity read for several weeks. The one figure that still separates them is the Nasdaq options-implied dealer delta, which is the most stretched on the board and sits well short of where the futures book reads, so the tech extreme this week is an options-book fact rather than a futures-positioning fact.

The Russell remains the group’s genuine tension and it is the standing story rather than the new one. Leveraged funds hold an extreme short against an outright dealer long that carries the equity complex’s concentration flag, and they are still reducing into it, which is the textbook squeeze-fuel configuration: a crowded short that is being extended has more to unwind if a catalyst forces it. Prior MODERATE LONG DELTA episodes in the Russell returned a median +5.8% over 4 weeks with 4 of 5 bullish, against an unconditional baseline of +1.1% median and 60% bullish (n=475). That is a deviation worth noting, but five episodes is context, not a signal, and the forward window is anchored to Tuesday, so three sessions of it have already elapsed. Small traders sit at a crowded extreme in the Nasdaq book, the classic contrarian-watched posture, and that is convention rather than a forecast.

Commodities (Gold, Silver, WTI Crude, Copper, Natural Gas)

Copper is the only commodity that cleared the gate and it cleared it on all three counts: a regime step, the group’s largest flow, and a full opposed-extremes configuration with Commercials stepping deeper into their hedge while Managed Money extends a speculative long, on a concentrated book. The Commercials-versus-Managed-Money gap is the canonical hedger-versus-speculator read here, and describing the tension is where it ends. There is no measured forward-return edge in this market; prior ELEVATED SHORT episodes for Commercials ran a median 4-week return of -1.4% with 2 of 5 bullish against an unconditional baseline of +0.8% median and 55% bullish (n=885), which is well inside noise for a five-episode sample. Other Reportables also sit near the top of their range in copper, so the long side of the hedge is spread across more than just the speculative crowd.

The rest of the complex sat quiet. Gold and silver hedgers hold the same mid-to-high posture as last week with the speculative crowd flat to slightly short, crude’s Commercials remain unusually net long for that cohort with the deepest trader participation on the board, and natural gas is a Swap Dealer story rather than a hedger-versus-speculator one, with Commercials and Managed Money leaning the same way.

Crypto (Bitcoin, Ether)

The intra-crypto split widened and it is a divergence in posture, not in direction. Ether dealers crossed back toward the middle of their range while Bitcoin dealers held at the bottom of theirs, so the structural dealer long in Bitcoin sits at the low end of its two-year range while Ether’s sits comfortably mid-range. That is the rotation the board has been signaling for two weeks, and the standing caveat still applies unchanged: crypto dealers are structurally net long, so a negative z-score here is a compressed long, never a literal short. Leveraged funds run the counterparty side in both, crowded long in Bitcoin relative to their own range while still net short in absolute terms, which is the same arithmetic trap as the rates book. Ether’s flow cleared the notable threshold on a neutral base, so it stays a table row rather than a story, and both books are thin enough at the top that the concentration numbers deserve more weight than the z-scores. Bitcoin’s prior MODERATE SHORT DELTA episodes returned a median +8.5% over 4 weeks with 3 of 5 bullish against an unconditional baseline of +2.1% median and 57% bullish (n=622); with five episodes and a wide dispersion that is context, not conviction. Bitcoin positioning stress is also decoupled from the equity complex this week, so do not read it as part of a broad risk-off.

COST BASIS LEVELS

Rates are excluded from cost basis by construction: structural shorts that never cross zero make the estimate unreliable. Four books trade inside their dealer zone this week (Nasdaq, Russell, VIX, Bitcoin), meaning the aggregate epoch position sits near flat P&L; the Dollar Index and natural gas trade below both zones and the commodity and large-cap equity books trade above. This is a descriptive P&L fact about the aggregate book. Tested in 2026, no covering response follows from a basis crossing, so do not treat any of these as triggers.

BOTTOM LINE

The risk into the Fed is concentrated in the front end, not in equities: size positions against a rates complex where leveraged funds and dealers are simultaneously at the top and bottom of their own ranges, and discount the crowding half of that read by the spread-structure share and the options caveat before acting on it. Everything else on the board, equities included, is normalizing toward the middle.