Actionable insights — The Second Mouse Is About to Get the Cheese

1. Decode an ETF's share count against price and short interest

1. Pull three series for the fund: price, shares outstanding (creations/redemptions) and short interest (plus borrow rate / locate availability).
2. Classify the regime: price up + shares up = new longs piling in; price down + shares up = units created to feed short sellers (ARKK 2022); price up + shares down = short covering or long redemptions.
3. Break the ambiguous case with short interest: if shares fall while SI rises, both longs are exiting and new shorts are borrowing from a shrinking float — the crowd is fading the move, and the vehicle is getting squeezable.
4. Confirm with borrow cost: rising fees and scarcer locates mean the release valve (fresh creation) is not absorbing the demand.
5. Mind reporting lags — SI prints twice monthly and trails shares outstanding.

• The mid-September SI print: a drop means covering already happened; a rise with shares still falling sharpens the squeeze.
• Borrow fee and hard-to-borrow status; shares outstanding turning back up (longs returning).

2. Use retail's leveraged inverse product as a consensus gauge

1. Track flows and AUM of the popular 2× inverse product on the asset (products with heavy roll yield and volatility drag that only retail and hot money hold).
2. Inflows while the product makes new lows = the crowd is adding to a losing short view.
3. Compare its AUM to the long-side vehicle's AUM; a bigger inverse book means the ETF crowd is net fading the move.

• SCO outflows / capitulation, or its AUM falling below BNO's — the consensus turning.

3. Check futures positioning, open interest and vol against the move you expect

1. Compare managed-money net length now to its earlier peak at a similar price — lighter length near the highs = a "second mouse" setup.
2. Look at managed money as a % of open interest versus prior bull-market peaks (here 2017-18).
3. Check open interest itself: if it has shrunk (here Brent OI ~-25%), re-entry by speculators needs higher prices to coax hedgers to write new contracts.
4. Widen to all speculators (Other Reportables + Non-Reportables) in notional terms.
5. Compare implied vs realized vol: low implied into known upside catalysts (Bab-el-Mandeb, depleted storage) = no fear of an upside skid, so upside convexity is cheap.
6. Add anecdote: respected longs lightening up is confirmation, not a warning.

• Managed money % of OI rising alongside expanding OI (fresh money arriving); implied vol bid on upside.

4. Handicap the central bank by committee arithmetic and the political calendar

1. Start from what the data says (2Y, inflation prints) and what the curve prices (here 88% for a hike this week, two by December).
2. Ask what the chair needs to act: a controversial move close to an election needs near-unanimity. Count the likely dissents.
3. If the committee is split, fade the priced move: receive the front end (2Y, SOFR) and expect a stop-out of the trades built on it (flatteners) — which hits the long end.
4. Carry the read-through to other assets: a long-bond blowout is hard for equities to ignore.

• The September FOMC decision and dissent count; flattener unwinds showing up as a long-end selloff.

5. Read inflation from the input the whole economy pays for

1. Track ISM Services Prices Paid as the lead on CPI.
2. Watch diesel, not just gasoline: diesel feeds freight, retail and food, so a diesel record passes through broadly.
3. Pair it with growth leads (stimulus roll-off, housing, large-employer openings, new orders/inventories) to judge stagflation vs reflation.

• Next CPI and payrolls; large-employer openings; PMI new orders slipping below inventories.