Research in labour conflicts crosscuts disciplines. Economists assign the incidence of labour conflicts to the economic development. The institutionalist camp is divided upon whether the power of organized labour or the industrial relations system is the main determinant of conflicts. All these approaches have run into difficulties when it comes to explaining the conflict patterns since 1970. Although empirical studies have combined economic and institutional explanations, important problems remain: A conceptual framework which links economic to institutional factors is lacking; the focus on labour power ignores the relational character of power; endogeneity problems are neglected; quantitative studies covering a larger number of institutional variables are merely cross-sectional, whereas TSCS analyses examine only a few of them. This paper tries to overcome these shortcomings. Its theoretical approach interprets labour conflicts as a problem of union choice which is embedded in a specific economic and institutional context. The hypotheses on these contextual factors are tested, using time-series data for 19 countries. The findings show that the trend towards social peace primarily follows from changes in the economic context, along with a weakening of the unions relative to employers. Cross-national differences in the incidence of conflicts persist mainly due to 'sticky' corporatist institutions.