3.1. Robustness testing
Before concluding that entry barriers reduce productivity causally, I need to rule out potential confounding influences. In this setting the main concern regarding omitted variables is that the removal of entry barriers coincides with unknown contemporaneous improvements in the local business environment. To rule out this possibility I exploit the panel structure of the data to purge omitted variable bias. The results reported in column 2 are estimated conditional on county–year fixed effects that eliminate the confounding influence of time-varying, county-specific productivity factors common to both industries. An attractive property of this specification is that the treatment effect is estimated through between-industry variation within the county–year dimension of the data which provides exceptionally clean identification