| This paper attempts to explain party unity variation at roll call level, as well as to analyze the asymmetries of voting patterns between different party groups. While conventional measures like party unity scores are effective at aggregate level, they fail to capture the variation in voting defection and the unique context of each roll call that contribute to the shaping of voting coalitions on the floor. Thus, the author chooses to adopt the Rice index that overcomes these limitations to estimate party unity.Drawing upon the findings of the previous studies on congressional organization and executive-legislative relations, this research employs a framework of the game between collective and individual preferences and falls back on two major theories: conditional procedural-unity-oriented voting theory and conditional presidential-position-oriented cross-voting theory. According to the former, party unity is strongest in procedural votes and least so in amendment votes, because different vote-types denote different inherent controversy, salience and party pressures. This association between vote-type and party unity is conditional in that it is only operative for partisan occasions (party unity votes) where party unity is essential to win the vote. Also, the theory is more likely to hold for the majority party than for the minority party, since the former has a bigger stake in and a greater power over the procedural proceedings, whereas the latter’s right to challenge is mainly invested in the amendment process. Based on these deductions, we point out three subordinate claims to complement the theory, defining its premises (partisan conflict, majority status) and shedding light on cross-party asymmetry in voting-unity in procedural and final passage votes is greater for the majority, whereas that in amendment votes is greater for the minority.The second theory addresses the influence of presidential public position taking on party unity-when the president disagrees with the party, the party’s unity decreases, because more legislators are willing to defect out of electoral incentives and other personal preferences. This theory is also conditional in that it works better for co-party presidents but less so for opposition-party presidents, since presidential reward for loyalty is exclusive to fellow partisans and the MCs have more incentives to follow their own party’s presidents. Additionally,the majority-party presidents are more influential than minority-party presidents, because the former work more closely with their co-partisans whereas the latter’s reliance on bipartisan support reduces their leverage to some extent.Based on the results of our multivariate regression models, when controlling for system-level influences like party polarization in the postreform era and congressional election years, our theories fare well in explaining variation in party unity from vote to vote. When the conditions are met, both vote-type and presidential position are significant determinants of party unity, and there is a clear cross-party asymmetry as to in what way and to what extent these factors influence party unity. As predicted, the above factors appear to exert an unbalanced influence over the majority party and the minority party.Through cross-party comparisons, this study contributes to the literature by offering valuable insight into how asymmetries in power shape asymmetries in unity. Also, it sheds light on the voting strategy of each party by delving into the interaction between collective preferences and individual preferences. Additionally, through the investigation of how presidential positions divide parties internally, this thesis analyzes the difference between co-party and opposition-party presidents, as well as that between majority and minority presidents. Such discussions contribute a congressional-centric perspective to the studies on executive and legislative relations, and provide a plausible explanation for the different legislative performance of different party presidents. |
PDF Full Text Request