Geographic Spillovers of Wind Energy Development on Wages and Employment

The goals of this paper are twofold: we first aim to quantify the impact of US wind development on local communities in terms of earnings and employment for workers and establishments. Second, we examine and then illustrate how the use of data aggregated to arbitrary spatial units (i.e., counties) can lead to biased economic impact estimates. We accomplish these goals using disaggregated, geocoded data on the universe of workers and establishments from 23 states who participated in their state’s unemployment insurance program. We compare estimates of regional economic spillovers from aggregating this data in two ways. First, we aggregate to increasing 20-mile rings around the location of wind projects (and matched control locations) and estimate impacts in a difference-in-differences framework. Second, we aggregate to the county level and then use county centroids to further aggregate county-level data to increasing 20-mile rings before re-estimating the same specifications. We find that the average wind project causes employment at establishments within 20 miles to increase by between 3.5 to 4.5 percent, whereas we find no discernible average effect on employment at greater distances, or for earnings at any distance. We find that the employment effect persists for as many as four years after the project arrives. Using worker data, we find employment and earnings effects of similar aggregate magnitude, but spread across further distances from the wind project. This is consistent with wind development spurring economic activity at nearby establishments, which improves employment prospects for workers who may commute to those establishments. Results using county-aggregated data are much smaller and mostly statistically insignificant. This has implications not just for policymakers, but for researchers who aim to continue to understand how burgeoning energy sectors such as wind impact local areas.