July 2015

IZA DP No. 9221: Sources of the Union Wage Gap: Results from High-Dimensional Fixed Effects Regression Models

revised version forthcoming as 'Union Membership Density and Wages: The Role of Worker, Firm, and Job-Title Heterogeneity' in: Journal of Econometrics

This paper provides estimates of the union wage gap in Portugal, a nation until recently lacking independent data on union density at firm level. Having estimated nonlinear and linear estimates of the effect of union density on the wage gap, the next stage of the analysis seeks to account for the influence of worker, firm, and job-title permanent heterogeneity, using a three high dimensional fixed effects strategy. Gelbach's decomposition is used to determine the role of each as sources of the union wage gap. A generalization of this technique is applied for the nonlinear case. We find evidence of a substantial union wage gap in circumstances where the majority of the workforce is organized. There is also some clear indication of substitution effects among different types of compensation that favor of wage supplements (likely reflecting tax considerations) as bargaining power increases. The principal result of the decomposition exercise at the most aggregate level is that the union wage gap is mainly manifested through a firm fixed effect, suggesting that unions may force firms to reposition themselves as far as their wage compensation policies are concerned. A subsidiary result is that matching plays almost no role. More importantly, however, are the results obtained when a distinction is drawn between the (estimated) bargained wage and total earnings. Rising union density has a much greater influence upon the former and is negatively related to the difference between the two. This result is consistent with a wider literature indicating that firms have a 'wage cushion' enabling them to avoid some of the strictures of unionism bargaining power.