BackMaharashtra assembly election: Anatomy of an alliance
An analysis of political alliances formed in Maharashtra can help analyse future pre-poll alliance possibilities
Premium
Elections in Maharashtra are due in October and the two familiar alliances—Congress-Nationalist Congress Party (NCP) and the Bharatiya Janata Party (BJP)-Shiv Sena—are going to slug it out.
While these two alliances have been the major contestants in Maharashtra for a while now, what people tend to forget is that in the first election after the formation of the NCP (in 1999), the NCP and the Congress fought as rivals, only coming together to form a post-poll alliance. After that, however, they’ve been pre-poll allies and have fought all elections from 2004 together.
This artefact of the Congress-NCP alliance allows us to understand how alliances work and how the tie-up of the Congress and the NCP in 2004 affected their performance. This analysis can help us analyse future pre-poll alliance possibilities.
Table 1 shows the number of seats contested by the four parties in the 1999 and 2004 elections. One obvious result of the alliance is that the number of seats contested by the two new partners came down significantly.
Table 2 shows the vote share of the four major parties and independents in the two elections. While the combined vote share of the Congress and the NCP came down in 2004, it must be remembered that the two parties together were putting up fewer candidates and also the alliance was the incumbent—prior to the 1999 elections, the Shiv Sena-BJP alliance was in power.
When analysing alliance politics, absolute vote share doesn’t make particular sense—what makes more sense is the vote share in seats contested (something we’ve earlier referred to as “effective vote share" in Election Metrics).
Table 3 shows the effective vote share of the four parties in the two elections.
It is interesting to note that the effective vote shares of the existing alliance partners (BJP and Shiv Sena) actuall
The real story of Table 3 is, however, the sharp increase in effective vote shares of both the Congress and the NCP— this shows the real power of forming an alliance. Given this sharp increase in effective vote share of the two parties, the alliance romped back to power in 2004 despite a sharp drop in total vote share (from 50% to 40%). It is interesting to note that while the number of seats won by the Congress actually dropped, the NCP’s gains more than made up for it. (Table 4)
Another bit of analysis that this data affords us is to see how parties divvy up the seat pie. When you form an alliance, how do you decide which party gets which seat? Does it have anything to do with the party that is stronger in that particular seat?
In the 1999 elections, as you can see from Table 1, neither the Congress nor the NCP contested all seats. There were some where exactly one of the parties contested while others where both headed off against each other. Based on the 1999 elections, we can divide seats into five categories from the point of view of the Congress-NCP alliance—neither contesting, only Congress contesting, only NCP contesting, both contesting with Congress winning more votes and both contesting with NCP winning more votes. For each of these five categories, Table 5 analyses how the seats in 2004 were distributed.
Table 5 shows that the policy of allocating the seat to the party that got most votes the previous time has been fairly adhered to—of the 102 constituencies where the Congress got more votes in 1999, 92 were allocated to the party in 2004. Of the 95 where the NCP got more votes in 1999, 83 went to the NCP in 2004. However, the distribution was not so clear-cut where one of the parties didn’t contest in 1999, as the table shows. An interesting artefact that comes out from this table is that in one constituency (Digras) in 2004, both the Congress and the NCP put up candidates, showing that the alliance was not watertight.
Finally, how much did each party contribute to the alliance in 2004? We will calculate this regressing the vote share of the alliance in 2004 against the vote shares of the Congress and the NCP in the corresponding seats in 1999. We will restrict this bit of analysis only to the 193 seats where both the Congress and the NCP had put up candidates in 1999, and where one of the two parties put up candidates in 2004.
A regression gives the vote share of the alliance in 2004 in a particular seat by the following equation:
Vote share_2004= 59%* vote share_INC_1999 + 74% * voteshare_NCP_1999
In other words, in the 2004 elections, the Congress-NCP alliance got about 60% of the Congress’s vote share from 1999 and about 75% of the NCP’s vote share from 1999.
Among other things, it shows that supporters of the NCP were more supportive of the NCP-Congress alliance. The weights are interesting given that the Congress contested many more seats than the NCP in 2004 (158 versus 123).
It also makes sense to do separate regressions for the 2004 vote shares among seats that the NCP contested and among those that the Congress contested. This will perhaps give us a better idea of how the alliance shaped up.
For seats contested by both the Congress and the NCP in 1999, and contested by NCP in 2004, the regression is:
Vote share_NCP_2004 = 30% * vote share_INC_1999 + 90% * vote share_NCP_1999
For seats both parties contested in 1999 and contested by the Congress in 2004, the regression is:
Vote share_INC_2004 = 78% * vote share_INC_1999 + 44% * vote share_NCP_1999. (All three regressions are extremely significant with R square values of the order of 90%.)
What this shows again is that supporters of the NCP were much more supportive of the alliance compared to the supporters of the Congress. Ninety percent of NCP voters continued to vote for the NCP after the alliance was formed, and 45% transferred their votes to the alliance partner. For the Congress, only 78% of those that voted for the Congress in 1999 voted for the Congress as part of the alliance in 2004 and a paltry 30% transferred their votes to the alliance partner NCP.
The above analysis shows that it has primarily been erstwhile NCP voters that have been voting for the NCP-Congress alliance while more Congress voters have shied away from the alliance after it was formed. Yet, in both 2004 and 2009, the Congress has been the dominant partner in the alliance contesting 158 and 170 seats (out of 288), respectively. It will be interesting to see how the alliance will fare this time round, both in terms of seat sharing and in terms of performance.
Unlock a world of Benefits! From insightful newsletters to real-time stock tracking, breaking news and a personalized newsfeed – it's all here, just a click away! Login Now!
