Legislator turnout and the calculus of voting: The determinants of abstention in the US Congress

The belief that turnout and abstention depend on the interaction of the ben efits and costs of voting is one of the longest held in public choice. The interplay between benefits and costs has typically been studied within the context of voter turnout in elections, but these results are confounded by the fact that participation in such elections is a low-cost, low-benefit ac tivity. Analysis of voting in legislatures, where the potential returns and costs to participants are greater, is more promising. This paper examines participation on roll calls in the U.S. House of Repres entatives during the recent 104th Congress (1995-1996). We analyze all cont ested roll calls during this two-year period, using a negative binomial cou nt model that accounts for legislator heterogeneity, to determine what fact ors associated with such votes induce turnout. Our results suggest that, wh ile turnout is predictable, its primary explanation does not lie with our s tandard calculation of expected benefits. We find that the ideological pola rization of the roll call alternatives is not a very important factor for t urnout and, even more strikingly and contrary to what past analysis has imp lied, the likelihood of any given legislator being pivotal is completely ir relevant. Rather, what is most important is that a non-trivial number of me mbers of Congress decide not to vote when contextual factors force them to choose between electioneering and legislating. Consequently, while our resu lts are not sufficient to induce a full-fledged "paradox of congressional v oting'' analogous to that of mass voting, they do indicate that participati on in Congress depends most heavily on factors beyond a desire to affect th e outcome.