This paper studies how climate change affects unemployment dynamics by focusing on the often-overlooked non-seasonally adjusted (NSA) unemployment rate. While standard practice emphasizes seasonally adjusted (SA) data, we show that seasonal adjustment removes roughly 80% of within-year variation in unemployment, potentially masking important climate-related effects .

Using a newly constructed panel dataset of U.S. counties from 1990 to 2019, the paper links local temperature fluctuations to monthly unemployment outcomes. The empirical strategy exploits plausibly exogenous variation in temperature across time and space within a two-way fixed effects framework. The key finding is that extreme temperatures—both hot (≥75°F) and cold (<50°F)—significantly increase unemployment, with 10 additional extreme-temperature days raising the unemployment rate by about 0.2–0.3 percentage points . In contrast, these effects are largely absent in SA unemployment data, indicating that climate impacts are concentrated in the seasonal component.

Quantitatively, the authors estimate that, absent extreme temperature days, the average NSA unemployment rate would have been about 11% lower in the 2000–2019 period . However, the corresponding effect on SA unemployment is negligible, reinforcing the idea that conventional macroeconomic measures understate climate impacts.

To understand mechanisms, the paper analyzes job and worker flows using QWI and JOLTS data. The results show that extreme temperatures primarily reduce labor demand, operating through fewer job openings and lower hiring, with additional contributions from increased separations, especially layoffs. This leads to a “slacker” labor market and higher unemployment insurance recipiency.

The paper also evaluates long-run climate change effects. Warming has produced milder winters, which reduce winter unemployment, but also hotter summers, which increase unemployment more strongly. As a result, the net effect of climate change is slightly positive, increasing overall unemployment. Moreover, climate change explains about 5.6% of the decline in seasonal volatility of unemployment by lowering winter peaks and raising summer troughs.

Overall, the paper contributes to macro-labor and climate economics by showing that climate change is an important but hidden driver of unemployment dynamics, one that is largely missed when using standard seasonally adjusted data. This paper is currently under submission and is scheduled to be presented at the 2026 Spring Meeting of the Japanese Economic Association.