Electoral Politics in the Zero-Sum Society

In most recent work on the theory of elections, parties are assumed to compete over a multidimensional space of issues or policy variables. Distributional considerations arise only indirectly in this structure, and candidates cannot appeal directly to particular constituents or groups by offering them specific targeted benefits or services. This theory of pure "issue" politics thus ignores the prevalent constituent-service aspects of contemporary electoral politics. The present paper develops a theory of electoral competition under an alternative structure, in which candidates compete by directly offering particular benefits and services to voters. The analysis presumes a symmetry in the roles of incumbent and challenger, in that the former necessarily commits himself to an allocation first, by his actions in office, thereby presenting the challenger with a fixed target to optimize against. Voters tend to discount the challenger's promises to some degree in comparing them to the benefits currently being received under the incumbent, and cast their votes so as to maximize the level of benefits received. The main results are as follows: 1. Optimal candidate strategies in this regime turn out to be rather different from those in the classical spatial modeling framework. Challengers pursue a "divide and conquer" strategy of bidding for a minimum winning coalition of voters. Incumbents, by contrast, pursue a more even-handed strategy, attempting to appeal to all their constituents. The model thus predicts distinctive differences in the behavior of challengers and incumbents, with no tendency for the candidates to converge on a common strategy or position, as in the classical Downsian case. 2. The discount factors voters use in assessing the challenger's promises—the "incumbency premia"—can be interpreted as a set of constituent demands. If these are treated as endogenous strategic variables which voters vary so as to maximize their long-run level of the benefits, there exists an equilibrium. In equilibrium, voters capture all the benefits from the parties. The degree of inequality in the equilibrium allocation is related to the degree of risk aversion with which the electorate views candidate behavior. 3. An issue is a measure or proposal which, if enacted, would generate a fixed distribution of benefits and costs, and on which each candidate must take a position. We obtain simple classification of issues according to their electoral consequences, and show that one important category of issues—which we label the "controversial" issues—is strategically important. The existence of a controversial issue invariably work to the disadvantage of the incumbent; hence he always has an incentive to suppress or remove it from the electoral arena altogether, if he can. If he cannot, it will then be optimal for the incumbent to favor the issue if and only if it is one which produces a (positive) net social benefit. Even with this optimal position, however, under general conditions the incumbent will nevertheless be defeated, by a challenger who opposed the issue and who will therefore not enact it, even though it would be socially optimal to do so. These results thus support the doubts expressed by Thurow and others, concerning the inability of a competitive democratic systems to deal effectively with major issues when distributional considerations become politically important. They also imply, however, that Thurow's proposed reforms, to strengthen party responsibility, would not help, since the problem lies in the nature of the competitive process itself