Identifying latent conflict in collective bargaining

The purpose of this article is to find indicators of the substantive incentives for industrial conflict. We argue that collective decision-making models can be helpful in developing such a measurement tool. These indicators will enable scholars in industrial relations to distinguish the substantive incentives from other reasons for calling out strikes, e.g. union internal political incentives or trade union competition for members. The models we use are the expected utility model of Bruce Bueno de Mesquita and the exchange model of Frans Stokman and Reinier van Oosten. These models differ in their assumptions on how the collective bargaining process proceeds. The expected utility model is based on the assumption that actors reach collective decisions after threats and conflict. The exchange model is based on a cooperative strategy. Our indicators of industrial conflict are tested on a data set with information on 29 collective agreement negotiations in which 90 trade unions are nested. The data set contains instances of both industrial conflict and peaceful settlements of collective agreements. The results show that the indicator based on the expected utility model is a good indicator of industrial conflict, whereas that based on the exchange model is a poor one. This seems to contradict findings in previous applications of both models of collective bargaining, in which the exchange model provided the most accurate predictions of the bargaining outcomes. We conclude that although the exchange model provides better predictions at the aggregate level (the collective outcome), the expected utility model generates superior predictions of the individual positions after bargaining.