Return to Janda, "Comparative Political Parties: Research and Theory," home page
Concepts for Analyzing Political Parties (cont)

This is p. 176

"benefits"--which may or may not come from winning (Schlesinger 1991: 138). For Schlesinger, who views party leaders as pure office seekers, Wright has it all wrong: there are not "different kinds of parties but the same kinds of parties subject internally to the tensions provoked by the conflicting goals of office and benefits" differentially shared by leaders and activists (p. 145). May (1973) applied the term "curvilinear disparity" to the condition arising when top party leaders and party voters were congruent in their goal orientations while middle-level leaders differed. Kitschelt's analysis (1989) of "the curvilinear disparity of political incentives between voters, party activists and party leaders" (p.401) contended that evidence of such disparity comes mainly from two-party systems and that "curvilinearity is a much less plausible trait of the micropolitics of party organization" in continental Europe (p.421).

Nevertheless, Schlesinger uncovered a confusing question in the discussion of incentives: are incentives properties of individuals or of groups? He regretted that Downs, who defined a party as "a team" seeking office through election, assigned ambition to the party collectively rather than to party activists individually. To Schlesinger, Downs' approach led away from understanding party organization (Schlesinger 1991, 36). By shifting discussion from the motivational bases of party activists (incentives) to the nature of party goals, Wright moved to the concept of party strategy.

Strategy and Tactics

Earlier, I favored defining "party" broadly to include restrictive and subversive parties as well as competitive parties. Instead of categorizing parties into different strategies, one can regard a party's strategy for obtaining its goal as a variable. For example, the French Communist Party, which competed in elections and won about a quarter of the vote during the 1950s, also operated as an anti-system party, using strikes and demonstrations to destabilize the government. Thus the PCF could be scored as following a mixed strategy: mostly competitive but somewhat subversive. The dominant Mexican PRI, on the other hand, followed a mostly competitive but somewhat restrictive strategy, hampering and even controlling its opposition. One study of a representative sample of 150 parties across the world in 1960 found that only about 50 percent followed a pure competitive strategy, 11 percent were pure restrictive, 3 percent were pure subversive, and the other third employed some mix of strategies (Janda 1980b, 78-90). Mixed strategies are consistent with Wright's view (1971) of competitive parties as combining elements of the Rational-Efficient and the Party Democracy models. As we will see, the concept of party strategy as a variable opens new avenues in rational choice analyses of the behavior of competitive parties. Not only is the theory of party strategies best developed for competitive parties, but most of the world's parties (68 percent in 1960) have been purely or mostly competitive. So further discussion will deal only with competitive party strategies.

Competitive strategies: Downs' seminal book (1957) is often viewed as belonging to the spatial theory of voting, but it is equally pertinent to the theory of parties. Downs assumed not only that parties seek to win office but that they seek to maximize votes (1957, 11,31). Arguing that parties formulate policies to win elections, rather than win elections to formulate policies, he theorized that parties in a two-party system with a unimodal distribution of voter preferences would locate their policies vaguely at the middle of the distribution. In so doing, Downs neatly accounted for the centrist two-party system in the United States. Downs' model of the vote-maximizing party has been especially influential for American party theorists, such as Schlesinger (1991, 143). In comparative research, however, Downs' work has been less important than that of Riker (1962). Riker also assumed that parties seek to win office, but they do not maximize their votes (see also Wellhofer 1990). Instead, they seek to win by the smallest possible margin, called the "minimum winning coalition" (1962, 32-33).[24]

Riker's work was important to comparative parties theory because it applied to the creation of parliamentary coalitions to form a government. Coalition theory assumed an "office-seeking" rather than "vote-seeking" strategy.[25] Early tests of coalition theory by De Swaan (1973) and Dodd (1976), and articles in Browne and Dreijmanis (1982) typically found minimum winning coalitions (measured by size of the legislative majority or by number of parties) forming less than half the time. Dodd also showed that minimum winning coalitions lasted longer than non-minimum winning coalitions, but his finding was challenged by Grofman (1989), who explained it by features of the party system. Laver and Schofield (1990) provided the most thorough nonmathematical discussion of coalition theory, comparing minimal winning coalitions (using various criteria) with minimal winning coalitions that are also ideologically "connected," with other versions that incorporate policy considerations, and with the game-theoretic concept of the "core point" in parties' preferred policy positions. They noted that while the best "policy blind" theory "is more often wrong than right about which coaltion will form (making correct predictions in about 40 percent of all situations in which no party has a legislative majority), it does very much better than we

go to page 177