This article presents the first nonparametric test whether German works councils go hand-in-hand with higher labour productivity or not. It distinguishes between establishments that are covered by collective bargaining or not. Results from a Kolmogorov–Smirnov test for first-order stochastic dominance tend to indicate that pro-productive effects are found in firms with collective bargaining only. However, the significance level of the test statistic is higher than a usually applied critical level. This somewhat weak evidence casts doubts on the validity of results from recent parametric approaches using a regression framework that point to high positive effects of works councils on productivity.