Csaba Nikolenyi, "Recognition Rules, Party Labels and the Number of Parties in India: A Research Note," Party Politics, 14 (March, 2008), 211-222.

First paragraph:
According to Duverger's (1954) famous dictum regarding the relationship between the electoral rule and the number of parties, the first-past-the-post electoral rule should lead to a two-party system. Strictly speaking, the logic of this prediction applies only to party competition at the district level, but Duverger (1954: 288) himself assumed that local two-party equilibria would be aggregated upward to produce a national two-party system. However, recent research has shown that a Duvergerian two-party equilibrium emerges only under special circumstances at the district level (Cox, 1997; Gaines, 1999) and that beyond the districts the number of parties is governed by the process of party aggregation, itself a product of factors other than the electoral system alone. Such factors may include the degree of economic centralization (Chhibber, 1999; Chhibber and Kollman, 1998, 2004); party finance rules; the pursuit of upper-tier seats in the national legislature; the pursuit of a powerful presidency; or, in parliamentary systems, strong linkage between plurality status and formateur role in the government formation process (Cox, 1997, 1999).

Figures and Tables:

Figure 1. Impact of the actual on the effective number of parties in India, 1952-2004

Figure 2. Number of parties in Indian elections, by type

Figure 3. Effective number of parties in India, 1952 to 2004

Last paragraph:
Conventional wisdom attributes a central role to the electoral formula in shaping the number of parties in a polity. Yet, in cases, such as India's, where the number of parties, both actual and effective, changes significantly despite stability in the electoral rule, one must consider the possible impact of other relevant political institutions. I have pointed to the ways in which one such set of institutions, namely party recognition rules, has shaped the development of the number of parties in India's national party system over time. This, however, is not to suggest that the electoral formula has been irrelevant. Indeed, I have argued that the two sets of institutions have interacted with one another: party recognition rules could either mitigate the reductive effect of the electoral formula by encouraging an increase in the actual number of parties, or enhance the reductive effect of the plurality system by providing incentives for the formation of large national parties, as was the case early on after Independence.