Party Institutionalisation and Switching from Parties: Evidence from India

Earlier studies on party switching have given primacy to the vote (Young, 2013), office (Gherghina, 2014) and policy motivations (Desposato, 2006) that drive politicians to switch party affiliation. However, there has been less attention given to how the internal attributes of political parties contribute to switching. My paper attempts to contribute to the conversation by identifying a new feature of political parties that influences switching i.e. the degree of institutionalization of the party. I expect that party members of institutionalized parties will show lower propensity for switching than party members of non-institutionalized parties. The basis for the argument stems from the characterization of party institutionalization by Panebianco (1988), which identifies that institutionalized parties offer longer and more predictable career paths for party members. I make distinctive arguments for how each of the two attributes (length of careers and predictability of careers) has an impact on party switching. For long careers, I argue that parties that encourage longer careers facilitate higher degrees of internal socialization and cultural assimilation within the party, which raises the costs for members to move to another party. For predictability of careers, I argue that by presenting predictable careers, parties reduce career uncertainty for members thereby breeding more loyalty. To test the arguments I constructed an original data-set on switching and party institutionalization from the unique political context of India, which showcases a medley of parties that vary in the degree of institutionalization. Based on the earlier arguments, I conceived of two novel measures for party institutionalization: average party career length and variance of party career paths. Each of these are measured for 35 major sub-national parties using publicly available information on the career paths taken by over 500 party legislators. The validity of these measures is checked with measures drawn from earlier attempts at operationalizing institutionalization of Indian parties (Chhibber, Jensenius and Suryanarayan, 2014; Kitschelt 2012). For the dependent variable on probability of switching by individual party members, I use original data on party switching out of more than 2,400 parliamentary candidates in India across four national elections over the 15 year time period of 1999 to 2014. The result of the multi-level model, after accounting for controls such as the ruling status of the party, nature of political competition and attributes of the individual MP lends partial support to the hypothesis. I find that party career length, as expected, has a significant negative effect on the propensity for parliamentary candidates to leave the party. However, party career variance has no effect on switching. This suggests that the socialization argument is potent at impeding switching but there is no support for the career uncertainty argument. Further to this I also run an instrumental variable estimation by incorporating party age as an instrument for party institutionalization (in this case for party career length). I find that the effect of party career length persists. The findings suggest that parties can reduce switching by instituting rules on career progression that promote longer party careers.