The Fed meets Wednesday. While most people are focused on whether we get a 25bp cut (spoiler: we probably do), there's a different game in town - betting on the actual words that come out of Powell's mouth.
Kalshi has markets on whether Powell will say specific words during his press conference. "Pandemic," "Trump," "Beige Book" - you name it, there's a contract for it.
So we did what any reasonable person would do: pulled the last 9 FOMC press conference transcripts and counted every word.
The Methodology
Simple stuff. Take historical frequency, compare to current market price, find the gap. That gap is edge - or at least, that's the theory.
Obviously past performance doesn't guarantee future results. Powell could wake up Wednesday and decide he's never saying "pandemic" again. But the data is the data (assuming we didn’t make any mistakes - which is always possible).
The Full Breakdown
Word | Meetings | Hist% | Ask | Edge | Recommendation |
|---|---|---|---|---|---|
Unchanged | 13/18 | 72% | 29c | +43c | STRONG BUY YES |
Trump | 9/18 | 50% | 11c | +39c | STRONG BUY YES |
Pandemic | 17/18 | 94% | 59c | +35c | STRONG BUY YES |
QT/Quant Tightening | 10/18 | 56% | 34c | +22c | STRONG BUY YES |
Uncertainty (3+ times) | 11/18 | 61% | 57c | +4c | Fair / Skip |
Tax (exact) | 3/18 | 17% | 18c | -1c | Fair / Skip |
Volatility | 5/18 | 28% | 31c | -3c | Fair / Skip |
Consumer Confidence | 3/18 | 17% | 20c | -3c | Fair / Skip |
National Debt | 0/18 | 0% | 7c | -7c | BUY NO |
Dot Plot | 5/18 | 28% | 38c | -10c | BUY NO |
Anchor (exact) | 1/18 | 6% | 18c | -12c | BUY NO |
Credit | 9/18 | 50% | 65c | -15c | BUY NO |
QE/Quant Easing | 1/18 | 6% | 25c | -19c | STRONG BUY NO |
Probability | 6/18 | 33% | 53c | -20c | STRONG BUY NO |
Dissent (exact) | 7/18 | 39% | 71c | -32c | STRONG BUY NO |
Beige Book | 6/18 | 33% | 68c | -35c | STRONG BUY NO |
Tariff Inflation | 3/18 | 17% | 55c | -38c | STRONG BUY NO |
AI/Artif Intelligence | 5/18 | 28% | 77c | -49c | STRONG BUY NO |
Shutdown | 4/18 | 22% | 85c | -63c | STRONG BUY NO |
The Highlights
Pandemic (Yes at 59c) - Powell's said this in every single presser we looked at except the most recent one. 94% hit rate. The market's pricing it at 59% - that's a decent gap if history holds.
Trump (Yes at 11c) - Shows up 50% of the time historically, market says 11%. With tariffs in the news and reporters itching to ask about political pressure, this feels underpriced.
Dissent (No at 29c) - 71% implied probability, 39% historical frequency. Powell doesn't love talking about internal disagreements.
The Traps
Some of these are priced pretty fairly:
Volatility, QT, Uncertainty - all within a few cents of historical rates
No edge means no trade
Risks
Powell could surprise us (he's done it before)
Sample size is 18 pressers - not huge
Market context matters - if something's in the news, he might bring it up
Reporters ask the questions, Powell answers them
Bottom Line
The word markets seem to be pricing on vibes. We're pricing on transcripts.
If the historical patterns hold, there's edge on both sides of the book. If they don't - well, that's why it's anyone’s guess.
Stay Positive,
The Edge Desk
Nothing here is trading advice — just data, probabilities, and our best read of the markets. Event-market trading involves risk and you should make your own decisions. We may hold positions in the markets we discuss and receive compensation through affiliate links.