Hot CPI Months: S&P 500 Falls 49.5% of the Time Afterward

The S&P 500 fell 49.5% of the time in the month after a hot CPI print, against 34.8% after all other months. That's the core finding of a backtest by SignalPilotLabs covering every month since April 1970 — 676 months in total — published ahead of October's US CPI release on 10 November at 13:30 UTC.

September's CPI rose 0.66% on the month. The analysis defines a "hot" month as one where consumer prices climbed 0.4% or more. By that line, 214 of the 676 months since 1970 qualify. For each one, the analysts checked whether the S&P 500 fell that same month and in the month after, comparing those results against all remaining months. Both the tests and the thresholds were written down before any market data was matched to any CPI month.

What the numbers show

The headline result is the follow-through month. After a hot CPI reading, the index dropped 106 times out of 214 — the 49.5% figure. After every other month, it fell only 34.8% of the time. The average return tells a similar story: the month after a hot print averaged +0.01%, essentially flat, versus +1.30% after ordinary months.

The hot month itself told a weaker story. The S&P 500 fell 45.3% of the time during hot-CPI months versus 37.0% otherwise, but that edge didn't survive the 1998 divide. Since 1998, the pattern actually reversed: 29.6% versus 40.1%.

Could it be luck?

SignalPilotLabs stress-tested the result. They drew 214 random months 10,000 times; only two draws produced 106 or more declines. The effect also held in both halves of the sample: 51.0% versus 33.7% from 1970 to 1997, and 46.5% versus 35.7% from 1998 to 2026. That met the statistical bar the analysts set in advance.

The record's hottest month, August 1973 at +1.81% CPI, shows the pattern's messiness. The S&P 500 fell 3.67% that month, then rose 4.01% the next.

The 2026 tape gives the finding fresh relevance. Four months ran hot this year: March (+0.85%), April (+0.64%), May (+0.48%) and September (+0.66%). After September's print, October fell 0.93%. August's +0.39% sat just under the threshold. October and November 2025 are excluded from the sample entirely, because October 2025's CPI was never published during the government shutdown.

On the accompanying chart — the S&P 500 monthly on a log scale from January 2020 to October 2026 — arrows mark each hot CPI month, orange when the index fell and grey when it rose. Only 9 of 28 are orange. A blue line below tracks year-on-year US inflation, peaking through the hot run of 2021 and 2022.

What does this change for traders?

For practitioners, the practical takeaway is narrow but usable. As SignalPilotLabs put it: "When a CPI print comes in at 0.4% or more, the record says the S&P 500's next month has been down about half the time, against about a third after other months." They add the necessary caveat: "It does not say how far, and a down month is not a crash."

The study is also explicit about its limits. It doesn't capture what the market expected — a hot print can still beat the forecast, and expectations often matter more than the headline. It says nothing about the day of the release, about rates, or about other markets. It rests on monthly closes and past data only.

The authors close with a one-line disclaimer practitioners should take literally: "Observations, not recommendations." Still, with September's 0.66% print already in the books and October's number landing 10 November, the base rate now slightly favors the bears for the month ahead — roughly a coin flip versus the usual one-in-three.