Relaxing the exclusion restriction in shift-share instrumental variable estimation

Abstract

Abstract Many economic studies use shift-share instruments to estimate causal effects. Often, all shares need to fulfil an exclusion restriction, making the identifying assumption strict. This paper proposes to use methods that relax the exclusion restriction by selecting valid shares. I apply the methods to estimate the effect of immigration on wages. The coefficient becomes much lower and often changes sign, which is in line with arguments made in the literature.