At-large versus ward elections: Implications for public infrastructure

Abstract

Since Gordon Tullock's (1959) seminal piece on logrolling, economists and political scientists have attempted to find empirical support for the argument that vote trading leads to larger municipal expenditure levels. In order for logrolling to occur there must be clear potential gains from trade, negotiations must be feasible and less costly than the value of the trade, and enforcement must be possible (Buchanan and Tullock, 1962; Mueller, 1979). The city council environment appears to satisfy these requirements. Apart from Zax (1985), however, who examined municipal employment and compensation, no empirical evidence of the effect of logrolling on municipal expenditures levels has been uncovered. Different institutional structures may provide different incentives to trade votes. In particular, ward election systems may provide a greater incentive to logroll than at-large systems (Bradford, 1911; Goetz, 1977). Ward councilors rely on neighborhood loyalties for their re-election. Thus, they are interested in providing public services whose benefits are geographically concentrated but whose costs are spread city-wide. At large councilors, whose constituency is city-wide, are more concerned with city-wide benefits. This difference in loyalties suggests that ward councilors are more apt to engage in vote trading behavior than are at-large councilors. One reason for the lack of empirical evidence of this behavior is that researchers have not examined the appropriate measure of municipal output. The local public good whose provision is most likely to be affected by logrol-