Actionable insights — Calendar-read Fed, normalized margins, and where to shelter

2:44 1. Deduce the Fed's next move from its own words plus the political calendar

1. Take the chair's latest speech at face value — "listen to what they say and that tells you where the head is at" — rather than the reputation or who appointed them.
2. Extract the one sentence that assigns responsibility (here: 65 months of elevated inflation "sits squarely with the central bank"). Responsibility claimed implies action.
3. Lay out every remaining meeting date against political dates. Strike any meeting too close to an election for a politically costly move.
4. The move lands at the first meeting that survives the filter. Then apply "don't fight the Fed" to positioning.
5. Test the "one soft print changes it" objection against the length of the record: one month does not undo 65.

• The chair walking back the responsibility language; a data shock large enough to outweigh the multi-year record; other central banks turning at the same time.

10:49 2. Read a geopolitical standoff through the incumbent's pocketbook exposure

1. Ask who benefits from the problem persisting until a vote. Voters "vote on their pocketbook" — fuel and food.
2. Check polling: if the incumbent is already weak, the adversary's incentive to keep prices high rises.
3. Find the historical precedent of an adversary timing concessions to a political change (Iran released the hostages hours after Reagan's inauguration, after 444 days).
4. Assume the disruption persists at least until the vote, and carry its second-order effect (oil into inflation) into your rate path.

• Gasoline prices into the vote; a negotiated reopening before November (would falsify the thesis); inflation expectations surveys.

18:55 3. Replace an index target with a conditional base-rate drawdown

1. Decline to forecast a level. Instead fix a recurring calendar window (end of July → November 9th).
2. Measure the median peak-to-trough drawdown in that window across a long sample (1990–2025).
3. Split the sample by the condition that applies now (midterm vs non-midterm year) and compare: 10% vs 5%.
4. Note the outliers and any analog year inside the sample (1990 Gulf War) before relying on the median.
5. Use the conditional number to size exposure, not to time the exact low.

• How much drawdown has already happened from the window's peak — a 10% median partly spent changes the risk/reward; the November 9th end of the window.

34:50 4. Normalize margins — an earnings bubble hides behind a reasonable multiple

1. Don't stop at P/E. Compare valuations against GDP or normalized earnings, because a low multiple on peak earnings is not cheap.
2. For a cyclical/commodity industry, pull current operating margins and compare to the cycle average.
3. Identify the force that restores the average — here a state-backed entrant for whom supply is "a national defense issue," with a captive home market (~20% of world PCs and phones).
4. Remember the entrant doesn't need to be leading edge: Japan (1970s–80s) and Korea (1990s) weren't when they took share.
5. Value the incumbents on the reverted margin, not the current one.

• CXMT/YMTC capacity additions and HBM qualification; memory contract prices rolling over; incumbents' margin guidance.

33:53 5. Pick the shelter by asking what is causing the risk

1. List the sources of equity risk. If rising bond yields are one of them, bonds cannot be the hedge ("it's hard to say, hey, you should go sit in bonds").
2. Check whether cash is actually paid: money-market yields versus inflation, in each relevant country.
3. Recognize the global competition for capital — when safe home-market yields return (Japan after three decades), foreign flows into equities weaken.
4. Use the 2022 analog: when stocks and bonds fall together, liquid cash is the one asset that works.
5. Keep the long-term thesis separate — being in cash for a quarter is not a call that the secular trend is over.

• Money-market yields falling toward inflation; bond yields stabilizing (reopens bonds as a hedge); the post-midterm window.