Drawing on a large-scale German Linked Employer- Employee data set spanning the time period 1995-2008, this paper provides new evidence on the collective bargaining wage premium in western Germany. By using longitudinal data, we seek to improve on recent evidence which relies on cross-sectional data. Unlike the previous literature, we assess the extent to which differences in wages between workers in covered and uncovered firms result from a non-random selection of workers and firms upon time-invariant unobservables into the different regimes. The fact that we observe employers changing their contract status over time provides us with the opportunity to measure the relative wage gains or losses of workers employed in firms that change their contract status.
Taken together, our analysis of separate transitions suggests that workers in firms leaving industry-level bargaining may incur wage losses relative to those workers who are employed by "stable" firms. However, adjusting the estimates for differential time trends supports the notion that plants changing from industry-level contracts to no coverage experience more negative time-specific shocks than stable plants, thereby confirming the result that there is no "true" wage effect of leaving wage bargaining.
For firm-level contracts, the analysis of separate transitions shows that joining firm-level bargaining from no-coverage may be associated with a positive wage premium, whereas the transitions between firm and industry-level contracts tend to give rise to negative wage premiums of firm-level contracts. This finding is consistent with firm-level bargaining being initiated by employers who were formerly covered by an industry-level contract and argues against the view that unions tend to enforce such contracts in order to secure above average wage gains in highly successful firms.
Gürtzgen, Nicole (2012), Estimating the Wage Premium of Collective Wage Contracts - Evidence from Longitudinal Linked Employer-Employee Data, ZEW Discussion Paper No. 12-073, Mannheim.