Tariff Reflexivity
The ISM's surge to 52.6 was driven by tariff front-running, not organic demand. Prediction markets are pricing the Fed's response off contaminated data about the economy tariffs are distorting.
On February 3, the ISM Manufacturing PMI came in at 52.6—up from 47.9 in December, a 4.7-point jump that crushed the consensus forecast of 48.5. The highest reading since August 2022. The first expansion in 12 months, ending a contraction streak that, combined with the prior 26-month run, meant manufacturing had contracted in 36 of the previous 38 months.
New orders surged 9.7 points to 57.1, the largest monthly gain since early 2022. Backlogs of orders rose 5.8 points to 51.6, ending 39 consecutive months of contraction. Commerce Secretary Howard Lutnick declared the number vindication: “For the first time in over two years the United States has delivered manufacturing expansion, all thanks to President Trump's trade policies.”
Employment fell for the 28th consecutive month.
That single data point breaks the narrative. If the orders were real—if manufacturers believed this demand was durable—they would be hiring. Two-thirds of ISM panelists reported that “managing head counts” remains the norm. This is the fingerprint of anticipatory purchasing, not organic recovery. And that contaminated signal is now feeding a triple feedback loop: tariff front-running inflates the ISM, the ISM reprices prediction markets on the Fed, and those prediction market prices feed back into the financial conditions the Fed monitors.
The contaminated signal
The ISM PMI is a diffusion index. It measures how many firms report increases versus decreases in orders, production, employment, and deliveries. It cannot distinguish between a new order driven by customer demand and a new order driven by a purchasing manager accelerating shipments before the next tariff round. Both register identically: expansion.
ISM Chair Susan Spence said it directly: “Some buying appears to be to get ahead of expected price increases due to ongoing tariff issues.” Approximately 40 percent of survey responses focused on tariff concerns.
This is not theoretical. The exact pattern played out in Q2 2025. Manufacturers built inventories to front-run the April 2 reciprocal tariffs. PMI readings inflated. By July, the S&P Global PMI fell back to 49.8—contraction—under the headline: “Manufacturing PMI back into contraction as boost from tariff front-running fades.” One quarter of inflated signal, then reversal.
January 2026 looks like the same cycle restarting. The pending Supreme Court ruling on IEEPA tariff authority is the new uncertainty catalyst. The Court heard V.O.S. Selections v. Trump on November 5 and is expected to rule around February 20. Both lower courts held that IEEPA does not authorize tariff imposition. If the Supreme Court agrees, the tariff landscape changes overnight. If it upholds the authority, the current regime hardens. Either outcome rewards stockpiling now.
Spence acknowledged the dynamic: “I do believe the Supreme Court ruling, whenever that is, is going to have a big impact on it, but we will take the win this month.”
The data vacuum
The contamination would matter less if markets could cross-reference it against other signals. They cannot.
The Bureau of Labor Statistics announced on February 2 that the January nonfarm payrolls report—scheduled for February 6—will not be released due to the partial government shutdown that began January 31. December JOLTS data is also delayed. The consensus forecast had expected 55,000 new jobs with unemployment steady at 4.4 percent.
The jobs report is the most important macro data release in the country. Its absence means the ISM is the dominant signal in a vacuum. Markets cannot check whether the “expansion” story is corroborated by actual hiring. At the exact moment the ISM is most contaminated by tariff front-running, it is the most influential data point in the system.
The prediction market divergence
The FOMC held rates at 3.50–3.75 percent on January 28, following three 25-basis-point cuts in late 2025. The March 17–18 meeting is next. The prediction market landscape on the March decision is fractured:
Kalshi prices the probability of a March rate cut at roughly 60–64 percent. CME FedWatch, derived from Fed Funds futures, shows approximately 48 percent. Polymarket shows roughly 10 percent—a 90 percent probability of no change. Goldman Sachs forecasts 50 basis points of total easing in 2026 with no cut until June. J.P. Morgan also forecasts a hold through Q2.
A 50-point spread between Kalshi and Polymarket on the same binary event is extraordinary. A 12-to-16-point spread between Kalshi and CME FedWatch is unusual. These platforms are pricing fundamentally different realities.
The Kalshi March contract alone has traded over $120 million in volume, with $450 million in open interest across the full Fed rate curve. A January 2026 NBER working paper found that Kalshi's modal forecast had a perfect track record predicting Fed rate decisions from 2022 through June 2025, with a 40.1 percent lower mean absolute error than consensus forecasts. That validated accuracy gives Kalshi's prices institutional credibility. When Kalshi prices a 60 percent probability of a March cut, institutions do not treat it as noise.
But that signal is downstream of the ISM print. And the ISM print is downstream of tariff front-running.
The triple loop
The reflexive mechanism operates in three stages.
Stage one: tariffs distort the ISM. Companies front-run tariffs by accelerating purchases, inflating new orders by 9.7 points and backlogs by 5.8 points. Chris Williamson of S&P Global warns that “over the past three months, factories have typically produced more goods than they have sold to a degree we have not previously seen since the global financial crisis back in early 2009.” The headline PMI reads “expansion.” The production-order gap reads 2008–09.
Stage two: the contaminated ISM feeds prediction markets. The 52.6 print, released into a data vacuum, shifts probability distributions on Kalshi and CME FedWatch. A strong PMI is traditionally read as reducing the urgency for rate cuts. But this PMI is inflated by the tariff policy that creates the inflationary pressure the Fed monitors. The Prices Index stands at 59.0—16 consecutive months of increases driven by steel, aluminum, and tariff pass-through.
Stage three: prediction market prices feed back into financial conditions. AI trading systems and institutional desks now treat prediction market probability shifts as actionable signals, executing corresponding trades in Treasuries and equity futures. On February 1, a 4 percent shift in a shutdown probability on Kalshi was reflected in market prices within 400 milliseconds. Traditional news wires took nearly three minutes. As examined in The Feedback Loop, the infrastructure connecting prediction market prices to traditional markets has become structural—but here the mechanism is machine-speed arbitrage between probability shifts and rate-sensitive instruments, not narrative formation through media distribution.
When Kalshi's March cut probability moves, Treasury yields move. When Treasury yields move, financial conditions change. The Fed monitors financial conditions when setting rate policy. The 10-year rose approximately 4 basis points to 4.283 percent on the ISM release—its largest one-day move of the year.
The result: tariff policy distorts the data that markets use to price the Fed's response to tariff policy.
The Warsh variable
The loop has a new input. President Trump nominated Kevin Warsh as Fed Chair on January 30, four days before the ISM release. Warsh, a former Fed governor who served as Bernanke's primary Wall Street liaison during the financial crisis, has a historically hawkish record but has recently shifted toward favoring more easing, arguing productivity gains can boost growth without driving inflation.
His confirmation is not assured. Senator Thom Tillis has refused to back Warsh until a DOJ probe of Powell is resolved. With the Banking Committee at 13–11, Tillis alone can block the vote. Majority Leader Thune acknowledged Warsh “probably” cannot win confirmation without him.
Markets must now price not only the probability of a March cut under the current Fed, but the probability that the current framework will be replaced by Warsh's—and when. Powell's term ends May 15. The nomination is being read as reducing the likelihood of an aggressive cutting cycle, pushing Treasury yields higher, tightening financial conditions, changing the backdrop the Fed evaluates when deciding whether to cut.
The structural tension
The ISM print of 52.6 will be cited for weeks as evidence that manufacturing has turned a corner. Lutnick has already claimed it. Markets have already moved on it. The 10-year yield has already repriced.
The sub-index data tells a different story. Employment contracting for 28 months. Prices rising for 16. A production-order gap not seen since the financial crisis. The Fed's own 2019 research on the first Trump tariff cycle found that tariffs were “a drag on employment and failed to increase output”—230,000 manufacturing jobs lost through the input costs channel alone. The current regime, at an effective rate of roughly 16.8 percent, is broader than 2018–19. Factory input prices are increasing at the sharpest rate since August 2022. A third of the reasons producers cite for price increases now mention tariffs—already exceeding the 2018–19 average.
The ISM is the most important macro signal in the system right now, and the most contaminated. Prediction markets are pricing the Fed off that signal. The prediction market prices are feeding back into financial conditions the Fed monitors. The government shutdown has removed the one release—the jobs report—that could discipline the ISM's story. And the Supreme Court is about to rule on whether the tariff regime that contaminated the signal has legal authority to exist.
Dawson Smith writes Reflexivity, a newsletter on prediction markets as reflexive systems.