The Effect of Unionization on Productivity: Evidence from a Long Panel of Coal Mines

Abstract

The author measures the effect of unionization on productivity based on a panel of West Virginia coal mines from 1897 to 1928. Output and inputs are measured in physical terms, and most of the mines in the panel changed union status at least once, though not simultaneously, so the panel is close to ideal for measuring the effect of unionization on productivity. Fixed-effects estimates show that the union had little effect on productivity before 1914, but thereafter it had a negative effect of 5 to 10%. This negative effect was not reversed when mines were later deunionized. The author evaluates a variety of possible explanations for these results. Some evidence points to declining investment at union mines relative to nonunion mines, but the evidence is circumstantial and the direction of causality is unclear. The most plausible explanation is a sharp deterioration in labor relations at union mines after the violent Paint Creek-Cabin Creek strike of 1912–1913.