Weather-Adjusted Employment Change

The Weather-Adjusted Employment Change data page provides estimates of monthly weather-adjusted employment changes in the United States. Starting with the official Bureau of Labor Statistics (BLS) series on the monthly change in total nonfarm payroll employment, we adjust for the employment effects of atypical seasonal weather, defined as deviations from historical seasonal norms following the methodology described in Wilson (2019).

The approach proceeds in four steps. First, we estimate the short-run effects of unusual weather on employment growth at the county level using historical data from January 1990 through December 2015. Second, we feed county-level weather data up to the latest month available into this estimated statistical model to measure the effect of unusual weather on each county’s employment growth in the latest month. Third, we aggregate these county-level effects to the national level, weighting counties by employment levels, to yield estimates of the effect of unusual weather around the country on national employment growth. Fourth, we translate these growth effects into level effects using the level of employment in November 2015 as an initial base.

The table displays the resulting estimates of weather-adjusted employment change and the official BLS payroll employment series for the past six months. We provide two alternative sets of estimates reflecting differences in step 1 of our process above. For the series labeled “Regional Heterogeneity,” the empirical model estimated in step 1 allows the effects of each weather variable to vary by Census region. For the series labeled “No Regional Heterogeneity,” the empirical model in step 1 constrains the effects of each weather variable to be constant across regions.

The bars in Charts 1 and 2 depict the most recent 18 months of data for the same three series in Table 1, along with lines showing their six-month moving averages. Chart 1 uses the Regional Heterogeneity model, while Chart 2 uses the No Regional Heterogeneity model. In addition, an Excel file below contains the full times series of the data underlying these figures.

Chart 1: Weather-Adjusted Employment Change (with Regional Heterogeneity)

Monthly change in 6-month moving averages, seasonally adjusted

Weather-Adjusted Employment Change (with Regional Heterogeneity)

Source: Bureau of Labor Statistics and Author’s calculations.

Chart 2: Weather-Adjusted Employment Change (with No Regional Heterogeneity)

Monthly change in 6-month moving averages, seasonally adjusted

Chart 2: Weather-Adjusted Employment Change (with No Regional Heterogeneity)

Source: Bureau of Labor Statistics and Author’s calculations.

Weather-Adjusted Change in Total Nonfarm Employment (monthly change, seasonally adjusted)

Month Official BLS
(Not Weather-Adjusted)
Weather-Adjusted
(No Regional Heterogeneity)
Weather-Adjusted
(Regional Heterogeneity)
Sep 2023246259269
Oct 2023165154154
Nov 2023182180176
Dec 2023290208204
Jan 2024229264288
Feb 2024275173186

Sources

Official Bureau of Labor Statistics series on the monthly change in total nonfarm payroll employment.

Data on monthly weather by county are constructed from daily weather station measurements provided by the National Oceanic and Atmospheric Administration (NOAA). See Wilson (2019) for details.

References

Wilson, Daniel J. 2019. “Clearing the Fog: The Predictive Power of Weather for Employment Reports and their Asset Price Responses.” American Economic Review: Insights 1(3, December), pp. 373–388.

Download Data

Weather adjustment time series (Excel document, 41 kb)


Contact Daniel.Wilson (at) sf.frb.org