I seek to determine whether or not political parties have significant independent effects upon the roll-call behavior of their members. Taking advantage of a natural experiment, I analyze the roll-call behavior of those members of the House and Senate from 1947 to 1997 who changed party affiliation while in office. Using data from the 80th to 105th Congresses, I find that Democrats who become Republicans, for instance, start to vote like Republicans at the time they “cross aisles.” This finding is consistent with the claims made in a growing literature that emphasizes the partisan aspects of congressional organization, and it supports the contention that party plays a direct role in determining members’ roll-call behavior.
The paper is by Timothy Nokken (gated). I’ve posted his graphs of what happened to Ben Nighthorse Campbell and Richard Shelby after their switch to the Republican Party. Pretty dramatic changes in party support.
Here is the abstract of a second paper:
In this paper, we analyze the roll call voting behavior of those House and Senate members who changed their party affiliation during the course of their political career. We analyze members who switched during the stable periods of the three major twoparty systems in American history: the Federalist-Jeffersonian Republican system (3rd to 12th Congresses), the Democratic-Whig System (20th to 30th Congresses), and the Democratic-Republican System (46th to 106th Congresses). Our primary finding is that the biggest changes in the roll call voting behavior of party defectors is observed during periods of high ideological polarization, and that party defections of the past 30 years are distinct from switches in other eras due both to high polarization and the disappearance of a second dimension of ideological conflict.
It is a basic premise of this study that politicians who switch parties are likely to do so when they are ideologically out of line with their current political party and more in agreement with the other party. Nor do we quarrel with the notion that sometimes the switch may also be related to electoral incentives.
Of course, lots of people fit this bill but don’t switch. King and Benjamin argue that this is because switching typically entails a combination of pressures:
Estimates from our statistical models indicate that switching is often explainable by systematic variables representing pressure on the legislators from the external, electoral, and congressional environments. Large changes in these three categories of variables each led to only fractional increases in the propensity of members to switch. It is only when there were large amounts of political stress coming simultaneously from several of these categories—such as is the case during critical realignments—that there is enough reason to cause members to change their political party affiliations.
I think this fits Specter pretty well. We’re not in a realignment, but clearly there has been a secular rightward shift in the ideological location of the Republican Party. Moreover, there is a large legislative agenda that would continually force Specter to confront his ideological discomfort. Add to that the electoral challenges of running in a closed primary, and, more generally, the challenges of being an elected Republican in a state trending Democratic.