AbstractWe examine the effect of unionization on firm innovation, using a regression discontinuity design that relies on “locally” exogenous variation generated by elections that pass or fail by a small margin of votes. Passing a union election results in an 8.7% (12.5%) decline in patent quantity (quality) three years post-election. A reduction in R&D expenditures, reduced productivity of inventors, and departures of innovative inventors appear to be plausible underlying mechanisms through which unionization impedes firm innovation. In response to unionization, firms move their innovation activities away from states where union elections win. Our paper provides new insights into the real effects of unionization.