Sources of the union wage gap : results from high-dimensional fixed effects regression models

We estimate the impact of union density on wages using Portuguese matched employer-employee-contract data, extending Gelbach's (2016) omitted variable bias decomposition procedure to obtain the contribution of worker, firm, and job-title heterogeneity to the union wage premium. The principal result is the dominance of the firm fixed effect: the allocation of workers among firms with different wage policies. For their part, the unobserved skills of union workers have only a modest impact on wages. In turn, job titles reflect the average contract in the collective agreement, while the wage cushion offers firms a margin of flexibility, partially undoing increases in the bargained wage. Finally, there is little to suggest that the union wage gap is influenced by improved match quality.