Extending a Dynamic Model of Protocoalition Formation

Abstract

With some loss of historical accuracy, we may think of there having been two generations of research on formal models of coalition formation. The first generation (see, e.g., Browne 1971; Koehler 1972; Rohde 1972a, 1972b; Leiserson 1968; Dodd 1976) was inspired by Riker’s seminal statement of the minimal winning coalition hypothesis (Riker 1962; Gamson 1961).² The second generation of models has focused on policy-driven motivations. The best known of these models is Axelrod’s (1970) notion of undimensionally “connected” coalitions (cf. Leiserson 1970a, 1970b; Hinckley 1972; Rohde 1972c). DeSwaan (1970, 1973) proposed a closely related notion of “policy distance minimizing” coalitions. Similar approaches are also found in Morgan (1976) and in Browne, Gleiber and Mashoba (1984). The early emphasis on minimum winning coalitions has been found inadequate both empirically (Browne 1971; Taylor and Laver 1973)³ and theoretically. For example, Grofman (1984) has argued that real world coalition politics is very unlikely to be zero sum in nature, and thus a key assumption of Riker’s (1962) work is inappropriate to the cabinet coalition context. More generally, several authors (see Luebbert 1986; Franklin and Mackie, 1984; Schofield 1984) have made very similar points about the inability of most coalitional models of cabinet formation to account for the substantial number of minority and supra-minimal coalitions or to accommodate the dramatic differences in coalitional types across countries.

This article is a revision and extension of Grofman (1982). Figures 1–4 are reproductions or adaptations of those presented in that article.