Kalshi and Polymarket put the chance of a Fed hike on Wednesday at roughly 80 percent.

That sounds decisive. And has been a widely quoted number across financial media.

But that only tells a small sliver of how the market is really positioning for this FOMC decision.

This week, I pointed Claude Code to dig behind this implied rate move. We’ll dig into the prediction market odds, what the options market says, and study context around Kevin Warsh’s past decisions.

If you want the full research, download the full report here: Pricing_the_Pivot_Sept2026_FOMC_v5.pdf​

Here are the key takeaways on how the market is positioning around Wednesday’s FOMC:

1. The hike odds jumped after two public releases

Two events drove the biggest changes in Polymarket’s hike odds:

• Warsh’s Jackson Hole speech: roughly 30 to 50 percent within half an hour.
• The August inflation report: roughly 58 to 79 percent within ten minutes.

Across the five scheduled releases I examined, the jumps added up to more than the overall increase. Some of those gains reversed between releases.

The odds were reacting quickly to the same public information available to other investors. To judge how much weight to give those moves, we also need to look at the market behind the quote.

2. It takes surprisingly little money to move the headline odds

The short answer is I wouldn’t trust prediction market odds.

In the Polymarket snapshot I examined, roughly $560,000 of buying could have lifted the quoted hike probability by five percentage points.

That estimate comes from the sell orders visible at the time. It shows how little capital stood between the current quote and a meaningfully different headline probability.

Ownership was concentrated too. The largest wallet held about 20 percent of the sampled hike position, and the top ten held 55 percent.

A five-point change in the odds doesn’t necessarily mean a broad change in investor opinion. One determined buyer can have a large influence on a thin market.

When that price becomes a headline about what “the market” expects, the concentration and limited liquidity are easy to miss.

Kalshi’s similar odds provide a useful cross-check. Even so, I wouldn’t treat the 80 percent quote as settled consensus: thin betting markets can misprice an event.

The next check is whether bond and options trading supports the same view of the risks.

3. Options traders favor medium-term bonds and higher-rated companies

Not all types of bonds / rates are being positioned equally.

Four funds accounted for 98% of the money traded in options across the fourteen bond funds I examined.

During the September 8-11 repricing, the average daily flows were:

These are estimated net options flows, which can include protection for investments traders already own.

The split has a useful portfolio implication. Long-term government bonds and weaker corporate borrowers were attracting downside exposure. Medium-term Treasuries and stronger borrowers were attracting upside exposure.

The government-bond flows suggest caution about the bonds most sensitive to changes in rates. The company-bond flows suggest a preference for borrowers more likely to repay their debts.

So the hike odds aren’t a reason to treat every bond fund the same way. The options activity favors specific maturities and credit quality, rather than a broad bet on all bond prices falling.

These flows don’t translate into an alternative hike probability. They show which exposures traders prefer around the decision, which is more useful for positioning than repeating the 80 percent headline.

4. We will have amplified volatility on long bonds vs typical FOMCs

Options markets say we should get amplified volatility vs other rate decisions.

TLT holds long-term US government bonds. If you own it, the firms handling its options trades may add to the swings in your position this week.

The main positioning estimate says they are short gamma:

• If TLT falls, dealers may need to sell more, adding to the decline.
• If TLT rises, they may need to buy more, adding to the rally.

A first drop may attract another round of selling rather than buyers who cushion it. That matters if you’re expecting long-term government bonds to provide stability around the decision.

Fifteen of the previous sixteen Fed meetings began with the opposite estimate, where dealer adjustments tended to soften price moves.

IEF, the medium-term Treasury fund, still showed that more stabilizing setup. This is another reason to distinguish medium-term from long-term bond exposure going into Wednesday.

5. Options sellers face risk beyond whether the Fed hikes

TLT options implied a move of about 0.8% between September 11 and the September 16 expiry. The event’s pricing looked fairly ordinary compared with recent Fed meetings.

Yet the Fed could be starting a hiking cycle, with long-bond dealers positioned to reinforce the reaction.

A widely expected quarter-point hike can still be followed by a larger-than-priced bond move. Warsh could signal further hikes, causing investors to raise their expectations for rates well beyond Wednesday.

For someone selling options, getting the rate decision right wouldn’t protect them from that larger move. For someone buying protection, the guidance and the dealer setup matter alongside the headline odds.

The potential mismatch is an ordinary price for movement around a meeting that could change the expected path of rates. That’s the risk to examine when deciding how much protection to carry.

6. Warsh may wait to hike, even as pressure inside the Fed builds

Warsh’s past decisions show that he can support easier policy when the economy calls for it. In 2010, he voted for QE2, the Fed’s bond-buying program, while unemployment was high and underlying inflation was low.

His recent record is more relevant for Wednesday. As chair, he held rates unchanged in both June and July.

June was unanimous. In July, three of the twelve voters wanted an immediate hike, but Warsh still supported holding.

He has been willing to wait even when colleagues see enough reason to raise rates. That makes the latest vote and his explanation for it more informative than labeling him simply hawkish or dovish.

His stated approach puts weight on underlying inflation when the job market is balanced. Today, headline and core PCE inflation are rising, at annual rates of 3.7% and 3.3%.

The Dallas Fed’s trimmed mean, which removes the biggest price jumps and drops, is closer to 2.3% and broadly flat. That gives him a different reading of the inflation trend.

If the Fed holds again but more members vote for a hike, the pressure to raise rates is growing. A “rates unchanged” headline would miss that signal.

The vote will show how much support there is for the decision. Warsh’s explanation will help show whether it is a pause or a change in direction.

Where does that leave us on Wednesday?

A quarter-point hike is the leading bet. For portfolios, the question is also how much more tightening the Fed signals after that.

• Treat the 80% odds as a rough guide. Polymarket’s thin order book makes smaller moves easy to overread. I'd take those odds with a grain of salt as the options market is showing bit more tame expectations. A few large traders can shift the quote without a broad change in expectations.
• Bond positioning is selective. The options flows favor medium-term government bonds and higher-rated companies, with bearish flows in long-term Treasuries and weaker borrowers. Investors aren’t expressing the same view across all bonds.
• Long-term bonds could have the rougher reaction. TLT dealers may need to sell into a drop or buy into a rally, extending the move. Medium-term government bonds have a more stabilizing dealer setup.
• An expected hike can still surprise options sellers. TLT options priced about a 0.8% move through Wednesday’s expiry. Guidance toward further hikes could matter more for bond prices than the widely expected first hike.
• Read the vote alongside Warsh’s explanation. If he holds again but more members want a hike, pressure for higher rates is building. A hold alone wouldn’t tell us that the risk has passed.