The Freeman-Lazear works council/worker involvement model is assessed over two distinct industrial relations regimes. In non-union British establishments our measures of employee involvement are associated with improved economic performance, whereas for unionized plants negative results are detected. The suggestion is that local distributive bargaining can cause the wrong level of worker involvement to be chosen. Also consistent with the model is our finding that mandatory works councils do not impair, and may even improve, the performance of larger German establishments. Yet smaller plants with works councils under-perform, illustrating the problem of tailoring mandates to fit heterogeneous populations. Copyright Blackwell Publishers Ltd/London School of Economics 2000.