Unionization and Productivity: Microeconometric Evidence

It is widely agreed that unionization affects the rules and procedures governing the employment relation in organized establishments. The effect of these changes on establishment productivity, however, is unclear. Existing evidence is based on a comparison of union/non-union differences in value added per hour worked. Although positive union effects have been estimated, possible differences in prices and technology in the union and non-union sectors render the results inconclusive. The effect of unions on productivity is examined in the present paper using establishment level data from the U.S. cement industry. The cement industry provides a useful empirical framework. Output is easily measured in physical terms, and data on both union and non-union establishments permit estimation of the union effect controlling for differences in technology. The results suggest that unionized establishments are 6-8 percent more productive than their non-union counterparts. This conclusion is supported in time series data, where a comparison of productivity before and after unionization reveals a positive union effect of similar magnitude. Since the statistical analysis controls for capital-labor substitution, scale effects and technological change, the evidence suggests that unionization leads to productive changes in the operation of the enterprise. The results are relatively robust. Specification changes and adjustments for omitted variables leave the basic findings intact.