Rekha Diwakar, "Duverger's Law and the Size of the Indian Party System," Party Politics,13 (September, 2007), 539-561

First paragraph:
In the literature on the size of the party system, considerable attention has been given to the effects of electoral rules, particularly those governing politics in single-member plurality systems (SMPS). This interest follows Duverger's (1954: 217) argument that 'the simple-majority single-ballot system favours the two-party system'. Over time, this proposition has taken the shape of a law, famously known as Duverger's law. However, in empirical analysis of this law the examination of electoral data has often been overlooked, and for the most part the focus has been at the national level. India follows SMPS, but has many parties (not just two), especially at the national and state level, which is a situation that has led to many scholars treating India as an exception to Duverger's law. Others, however, have argued that the Indian party system at district level follows Duverger's law, and is therefore not a correct counterexample.1 While there has been some work addressing the question as to why India is an exception, the main focus has been on the national party system. In my study, in contrast, I use a comprehensive dataset to test the applicability of Duverger's law in the district-level party system in India. I find that a large number of Indian districts do not conform to the Duvergerian two-party norm, and that many elections involve competition between many parties and candidates.

Figures and Tables:
Figure 1. Histogram of effective number of parties (1952-2004)
Figure 2. Histogram of effective number of parties by election
Table 1. Distribution of effective number of parties by election
Table 2. Distribution of effective number of parties by region (1952-2004)
Table 3. Inter-election trends at the district level
Table 4. Effective number of parties - trends at the individual district level (1952-2005)
Figure 4. Nagayama diagrams for Indian districts by election
Figure 5. Segmented Nagayama triangle template
Table 5. Proportions of districts in segmented Nagayama diagram

First Paragraph of Conclusion:
I have used alternative methods to study the application of Duverger's law to the Indian party system at district level, and have come to three main conclusions. First, there is no unequivocal support for it in Indian districts. While many districts witness competition between two or fewer than two parties, there is a non-trivial number of districts where elections involve competition between many parties. This is similar to Gaines's (1999) finding on the Canadian party system, but in contrast to the existing beliefs and findings about the size of the Indian party system at district level. The last five elections during the period 1991-2004 have seen a move towards a more competitive Indian party system at district level. Furthermore, there is no clear movement towards the Duvergerian equilibrium, and negative inter-election trends have not lasted beyond two electoral cycles.