Understanding and predicting the effects of weather on macroeconomic data is critically important, but it is hampered by limited time series observations. Utilizing geographically granular panel data leverages greater observations but introduces a “missing intercept” problem: “global” (e.g., nationwide spillovers and GE) effects are absorbed by time fixed effects. Standard solutions are infeasible when the number of regressors is large. To overcome these problems and estimate both granular and global weather effects, we implement a two-step approach utilizing machine learning techniques. We apply this approach to estimate both historical backcasts and rolling out-of-sample nowcasts of weather’s effects on U.S. employment growth. We obtain several novel findings: (1) Contemporaneous and lagged weather anomalies explain 6-8% of the non-recessionary variation in employment growth, with substantially higher shares in certain industries. (2) Granular weather effects impact employment most in sectors such as construction and retail that reflect local labor demand, while global weather effects impact employment more in sectors such as manufacturing and trade/transportation that reflect national labor demand stemming from downstream weather-sensitive industries like retail and construction. (3) Employment weather effects also have substantial explanatory power for many other macroeconomic time series including retail sales, industrial production, PCE and housing starts. (4) Though weather data are available to financial market participants in real-time, the weather effects estimated in this paper are found to have statistically significant effects on payroll employment report surprises, and in turn, Treasury yield responses to employment reports.
Suggested citation:
Daniel J. Wilson. 2026. “Estimating National Weather Effects from the Ground Up.” Federal Reserve Bank of San Francisco Working Paper 2025-18. https://doi.org/10.24148/wp2025-18