A dynamic test of the marginality hypothesis

Abstract

The marginality hypothesis is an attempt to relate the voting margins of members of Congress to their subsequent legislative behavior. A major corollary of the hypothesis is that members of Congress with small victory margins will be more responsive to constituents than those with large victory margins. This has been assumed to mean that electorally secure representatives can afford to be more loyal to their congressional parties, since they have less cause to worry about their chances for reelection. Previous empirical studies have produced mixed results. We ask the question in a different way: Do changes in marginality affect party voting within Congress? If so, major shifts in the electorate potentially can have a fundamental impact on the behavior of Congress itself. We find that this is not true. Electoral margin is simply not related to party loyalty.