Stepping Down or Getting Pushed? A Competing Risks Analysis of Ministerial Turnover in the German Länder (1990-2010)

The tenure of ministers may end because of very different terminal events. Some ministers may step down voluntarily because they find other career options more intriguing, others will have to leave office because their party was voted out of government and again others may be forced by the prime minister to leave their ministry because their enmeshment in a scandal would otherwise seriously affect the government's popularity. These three examples show that ministers face different risks which could all finally lead to losing their job in the cabinet. Yet, most students of ministerial careers do not distinguish between these diverse hazards and therefore treat all terminal events the same way. In my paper I want to follow a different path. I distinguish between twelve specific terminal events and show how often they empirically account for a minister's end in cabinets of the German Länder. These twelve exit-types can then be aggregated to three broader types of terminal events: voluntary exits, forced exits and collective exits. I assume that there are distinctive factors at work for each of these terminal events. My study adds to the existing literature in several ways: 1) Based on a principal agent framework I develop a general theoretical model of a continuum of restrictions the prime minister is facing when it comes to the firing of ministers; 2) I test the hypotheses derived from that model using a newly compiled original dataset on the German Länder which includes all ministerial spells between 1990 and 2010. These subnational data are also one building block for the understanding of more complex multi-level career paths. 3) Applying a statistically more elaborate competing risks Cox-model I can estimate the different determinants of ministerial turnover according to the type of terminal event in a better way than previous studies