Koalitionsbildung und insbesondere die Verteilung der Ministersessel unter den Koalitionspartnern ist das Ergebnis eines komplexen kollektiven Entscheidungsprozesses. Der vorliegende Beitrag untersucht dieses 'bargaining' von einem spieltheoretischen Ansatz her. Als Testfall dient die belgische Mehrparteienregierung während der Zwischenkriegszeit. Die tatsächliche Verteilung der Ressorts wird mit der aus dem Modell abgeleiteten Prognose verglichen. (pmb) ; 'The distribution of minister portfolios in government coalitions is the result of a complex process of bargaining. In that way power relations among political parties are reflected in the distribution of cabinet posts. In this paper the predictions of game theoretical concepts (bargaining set, -core) and the Gamson hypothesis are compared with the actual portfolio distribution. In most Belgian interwar governments the relatively small liberal party was able to obtain a share of ministeries substantially larger than their share of parliamentary seats in coalition. Game theoretical models proved superior in capturing this effect. They also accentuate the disruptive effect of the 1936 elections. The success of extreme parties changed the simple bargaining structure of the 1919-1936 period into a complicated one. Moreover, the 1936 game is as far removed as possible from a Pareto optimal solution, which explains the labourious formation of short-lived governments.' (author's abstract)