Misery in America
SightSpinner Team ·
Misery in America
Americans are not happy. Some 77% of Americans say they are dissatisfied with the direction the country is going. Politics, war, and AI anxiety have all contributed to keeping that number elevated for the past year, but what about economics? By most measures, the economy has been thriving: the S&P 500 is up over 11% since this time last year; despite AI jobpocalypse fears, the unemployment rate sits at a healthy 4.2%; and GDP growth is still running at about 2.2% annualized. It's not all vibes, though. This bull market run has been accompanied by another less desirable surge: inflation. September's numbers have CPI YoY at 3.7%, and it shows little sign of slowing. Tariffs and war have been pushing prices up faster than a somewhat reluctant Fed has been able to tame. Gas prices are creeping up on their record highs. Diesel is currently sitting at around $6.35 a gallon, which is upstream of almost everything, causing cost-push inflation. And this, to put it succinctly, has made people miserable, in a very quantitative sense.
The Misery Index, as defined by Arthur Okun in the 1970s, is a relatively simple measure: the unemployment rate + inflation rate. Despite this, it's been remarkably successful in quantifying the real "mood" of the people, doing everything from capturing true public sentiment to helping predict elections.
The index itself might be simple; its components are anything but. They are highly aggregate quantities across a variety of industries, each with their own drivers, which are themselves dependent on other drivers. This makes forecasting quite tricky. Here we take a relatively simplistic approach in an attempt to keep the model somewhat digestible, keeping the width and depth of the driver chain small. We then compare to Polymarket data as a judge of at least the near-term forecast.
The actual Misery Index numbers for August and September were 7.5 and 7.94, respectively, well above the pre-pandemic range of 5.3–5.7 and almost double the low. The forecast says it's going to get worse before it gets better, predicting a January 2027 peak at about 8.86 before a decline into the following year.
As with the historical data, the forecasted index is expected to be mostly inflation-driven. Barring some large unforeseen shock, the unemployment rate is not expected to change much, hovering around 4.2% for the coming months. CPI YoY is expected to keep pushing up to a 4.66% peak in January before steadily falling to 2% by this time next year. Mid- to long-term forecasts, of course, depend on the moves by a less predictable than usual Fed. Interest rate forecasts have been left out of this model in the interest of simplicity.
---This model's data is limited to monthly publicly available data, so its near-term prediction is limited. Prediction markets are the gold standard of near term forecasts due to their real time crowd-sourced nature and to the fact that they frequently contain proprietary insider data. As you can see the forecasts largely agree with each other, differing by a percent or two.
The model doesn't try to predict world events, which could change any one of these forecasts. However, the current trajectory suggests a still-miserable winter, but spring and summer are something to look forward to.
Method / sourcing box (short)
- Misery Index = U-3 unemployment (FRED UNRATE, SA) + headline CPI YoY (computed from FRED CPIAUCNS, NSA index).
- Monthly, 2019-01 → 2027-10; the CPI leg forecasts from its data-end boundary (2026-10), the labor leg from the graph boundary (2026-11); single US aggregate, no dimensions.
- All inputs are real public data: FRED, BLS, EIA, Zillow ZORI rents, EPU, ISM-sourced via public/Philly-Fed substitute (ISM series not on FRED since 2016, substitution documented in-graph).
- Benchmarked against Polymarket implied forecasts (Gamma + CLOB midpoints), bucket parsing is dynamic per market.