Works Councils in the Production Process

This paper uses data from a nationally representative panel of establishments to estimate the effects of German works councils on firm performance, 1997-2000. We analyze the impact of this institution on sales and sales growth using OLS and fixed effect estimates of a translog production function as well as by employing a model in first differences. With cross section and pooled data, the strong pro-productivity effects of works councils noted in the recent literature prove sensitive to disaggregation – most notably for plants with 21 to 100 employees, where the powers of the council are a datum – even if the coefficient estimates for the works council variable are often substantive. However, the fixed effects estimator yields much smaller works council effects that are (weakly) statistically significant in only one instance, while productivity changes do not differ between plants with and without a works council in the first differences specification. We conclude that reports of positive works council effects on productivity have been much exaggerated. That said, there is no evidence that works councils adversely affect firm performance, as suggested by an earlier empirical literature based on small samples of firms.