Party organisation and polity-wide parties in India

Why do polity-wide parties in a multi-level context dominate in some sub national units but not others? Existing scholarship in comparative politics has either focused on regionally-based social cleavages or on political and economic decentralisation to explain the variation in the dominance of polity-wide parties at the sub-national level. Deviating from this predominant approach in comparative politics, I argue that political and economic decentralisation and the presence of distinct regional cleavages are necessary, but not sufficient conditions to explain the dominance of polity-wide parties. In other words, polity-wide parties are able to maintain their dominance at the sub-national level in a politically and economically decentralised polity with distinct sub-national regional cleavages if their sub-national branches have autonomy from their central branches. If intra-party autonomy is important in explaining the dominance of polity-wide parties at the sub-national level, then why do polity-wide parties have different levels of intra-party autonomy? Existing literature in comparative politics has focused on federalism, electoral systems, incumbency, the nature of regional party competition, the timing of regional elections, and internal party features in explaining the differing levels of intra-party autonomy. While these explanations are of particular importance in understanding cross national party organisational variations, I evaluate their merit at the sub-national level within a federal country with similar electoral systems. Empirically, this dissertation uses evidence from India, the world's largest democracy. It uses a multi method approach that combines statistical analysis of all major Indian states with a sub-national comparison of two Indian states to present the empirical findings. Building on previous work on the Congress Party and the Bharatiya Janata Party (BJP) at the national level, it looks at the dynamics of these two polity-wide parties in two Indian states, namely Gujarat and Maharashtra. Both the states are selected based on the Most Similar Systems Design, which further helps rule out alternative explanations in a systematic manner. Furthermore, the findings from Gujarat and Maharashtra are based on a long period of field research with semi-structured interviews with sub-national and local level politicians in Ahmedabad, Gandhinagar, Mumbai and New Delhi, India. Finally, in the latter part of the thesis, I use statistical analysis to test the qualitative findings across all the major Indian states using Chhibber, Jensenius and Suryanarayan (2014) dataset on party organisation. The findings from the statistical analysis are consistent with the main findings from my qualitative study providing further confidence in support of the primary theoretical arguments put forward in the thesis