We test the theoretical claim that coordination and centralisation in wage setting reduce strike activity by estimating nonlinear regression models using a dataset of 17 OECD countries for the period 1972-2000. We find moderating effects of coordination on strike activity but the effects are not stable over time. Several means are employed to check for the robustness of our results. We compute bootstrapped standard errors, conduct nonlinear median regressions. A remaining problem of the study is that we cannot control completely for country-specific heterogeneity in the estimates of coordination and centralisation effects.