Coalitional Politics and Logrolling in Legislative Institutions

Abstract

We examine how a foresighted legislative chamber will design its institutions in response to ex ante incentives for and ex post incentives for minimum winning coalitions and what coalitions will form as a result. To do so, we develop a model of vote trading with an endogenous voting rule and coalition formation process. We find that in equilibrium, legislative chambers will almost never choose institutions that guarantee either simple majorities or universalistic outcomes. Rather, coalition sizes from winning to universalistic will be possible under certain conditions given the choice of voting rule. Further, these coalitions will be minimal necessary, just large enough to sustain cooperation. olitical scientists have long been aware that politicians have an incentive to construct minimum winning coalitions (MWC) for distributive policies (Riker 1962; Koehler 1972; Shepsle 1974; Uslaner 1975; Koford 1982; Denzau and Munger 1986; Baron and Ferejohn 1989, but see Baron 1989; Groseclose and Snyder 1996). Including additional members tends to raise costs without adding benefits to the coalition. However, starting in the late 1970s, and at least partly in response to empirical evidence (Ferejohn 1974; Arnold 1979; Wilson 1986; and Collie 1988, among others'), political scientists also started to generate theories that predicted universal or oversized coalitions. For example, Weingast (1979) finds that, although games of distributive politics resulted in minimum winning coalitions, the electoral need to bring benefits back to the district and fear of being left out of the coalition could cause politicians ex ante to prefer universalistic coalitions. As such, politicians would have an incentive to set up norms of universalism enforced, perhaps, by punishment for not going along with the all-inclusive logroll. Later work found that such would be preferred even for inefficient pork barrel legislation (Shepsle and Weingast 1981; Weingast, Shepsle, and Johnsen 1981), as well as in political systems with committee structures (Fiorina 1981).2