THE MARGINALITY HYPOTHESIS Electoral Insecurity, Self-Interest, and Voting Behavior

The issue of legislative pay raises offers several advantages as a perspective from which to study the connection between voting behavior and electoral insecurity (the marginality hypothesis). The incentives for voting either yes or no are obvious and allow us to employ the powerful assumptions of self-interest in our analysis. Data for the Ninety-Fifth Congress supports the hypothesis that electorally secure legislators are more likely to support a pay raise than are less electorally secure legislators. Political scientists long have suspected that legislative behavior. is predictably affected by the relative electoral insecurity of indi-vidual legislators. Most often, the concept of electoral insecurity is defined in terms of representatives ’ margins of victory in general elections-hence the suspected relationship commonly has been labeled the &dquo;marginality hypothesis.&dquo; Legislators who win office by narrow margins are thought to pay closer attention to constituency preferences and more often to tailor their legis-lative behavior accordingly. Three basic assumptions underlie this hypothesis: (1) legislators desire to be reelected, (2) con-gruence between constituency preferences and legislative be-havior increases the probability of reelection, and (3) &dquo;marginal&dquo; legislators have a greater incentive to try to increase their re-election chances than do &dquo;safe&dquo; legislators. In short, the narrowly elected representative ignores constituency preferences at his or her peril, but the safe representative has more leeway