|Abstract:||The first chapter is a paper in which I develop a sequential voting model to study roll-calls in the United States Senate, and empirically test its implications. In this procedure, senators cast their votes in a sequence, exogenously determined by the alphabetical order of their last names. I categorize senators into multiple types whose utilities depend not only on the roll-call outcome, but also on the vote they personally cast. When, among certain types, preferences over the senator's personal vote and the roll-call outcome are not aligned, abstention emerges as a nontrivial choice. For example, on a specific Republican-sponsored bill, a Republican senator representing a moderate constituency may want the legislation to pass, but not to be pinned down by a vote for it. As a result, the senator may want to vote for the legislation if that vote is necessary for the legislation to pass, but to abstain or vote against it if the vote is unnecessary for victory, or if it is clear that it will not be passed. I prove that, in the sequential voting context, the opportunity to make such strategic decisions is not uniformly available to senators with different alphabetical ranks, which determines their positions in the voting queue. More specifically, the model predicts lower levels of abstention for those in the middle, implying a U-shaped relationship between rank and likelihood of abstention. Using the data from Senate roll-calls between 1867 and 2015, I find a significant U-shaped quadratic relationship between rank and likelihood of abstention, which provides empirical evidence that senators do indeed take strategic advantage of their last names.
In the second chapter, I assess the impact of abstentions in the U.S. Senate on senators' re-election chances. I analyze 1,255 senatorial re-election campaigns from 1914 through 2014 and find that there is an overall negative relationship between abstention rates and likelihood of incumbents winning re-elections. The relationship is not driven by extremely high rates of abstentions, and is stronger in swing states. I then use the Partisan Voting Index (PVI) as an instrumental variable to show that more politically polarized constituencies tend to care less about their senators' "laziness." In other words, senators who represent states with greater polarization are less likely to be punished for skipping their voting responsibilities in the Senate, as compared to their colleagues who represent more moderate states.