India has had seven consecutive elections (1989 to 2009) in which no single party won a majority of seats in the Lok Sabha, resulting in hung parliaments. In 1991, the Congress formed a single-party minority government (which achieved a majority half-way through its term) but in all other cases minority coalitions dependent on outside support were formed, these being large, multi-party coalitions with participation of several regional parties since 1996. In the run-up to such situations, one of the keys to victory for both the leading national parties, the Congress and the BJP, is the number of state-level pre-electoral coalitions formed, for pooling votes based on seat-sharing agreements.
Why pre-electoral coalitions? What are the incentives for national parties to form such coalitions, and under what circumstances? Given the plurality-rule (first-past-the-post) system, aggregation of votes at the constituency level is vital for winning seats. By implication, given the breakdown of the national party system into distinct state party systems, formation of alliances with parties commanding a significant state-level vote share, helps aggregate constituency-level votes shares in states where one’s own party is not strong enough to go it alone. Pre-electoral coalitions have the potential of increasing the number of seats won although at the expense of conceding a certain number of seats to allies, and also including such allies in a post-election government. The BJP, since 1989, has grown partly on the basis of its own ideological appeal and mobilization and partly by leveraging coalitions, while the Congress turned to coalitions with success in 2004.
In 2004, coalitions in Tamil Nadu, Andhra Pradesh, Bihar, Jharkhand, and Jammu and Kashmir, and the absence of a coalition for the BJP in Haryana, Assam and Jharkhand played a key role in the very narrow victory of the Congress-led UPA. In 2009, the Congress was critically dependent, despite a swing in its favour and a swing against the BJP, on pre-electoral coalitions in Maharashtra, West Bengal and Tamil Nadu, besides its longstanding one in Kerala.
The general finding on pre-electoral coalitions is that the seat-sharing ratio between partners tends to get stuck in a narrow band, and is not adjusted according to the demands of a partner perceiving its popularity to be on the upswing demanding more seats. For example, both the Congress-NCP and BJP-Shiv Sena coalitions in Maharashtra, the BJP-JD(U) coalition in Bihar and the BJP-Akali Dal coalition in Punjab, as well as the Left Front coalitions in West Bengal and Kerala, and the Congress-led UDF coalition in Kerala, have tended to remain stable in their seat-sharing ratios over the past two or more elections. It is only when old coalitions are discarded and new coalitions are formed that new ratios can be established. In this we would expect parties that are perceived to be on the upswing to be able to attract more allies and on more favourable terms.
If one compares the pre-electoral coalition tables for 2009 with 2014 for both alliances, NDA and UPA (on the eve of the elections without all candidates finalised yet), this is precisely what we can see. A BJP that according to forecasts over the past several months, is projected to be in the lead in votes and seats, has struck a range of new alliances, most of them on more favourable terms than in the past while retaining its key old alliances (Shiv Sena, Akali Dal) on the same terms.
NDA Pre-electoral coalition 2014 (till April 6)
UPA Pre-electoral coalition 2014 (till April 6)
Thus, in 2014 to date, the BJP has 10 pre-electoral alliances compared to 6 alliances in 2009 in which both it and its partner(s) are contesting seats, of which 7 are new alliances – Bihar, Seemandhra, Telangana, Tamil Nadu, Haryana, Kerala and UP – and on more favourable terms. By contrast, the Congress in 2014 has 8 alliances, one more than 2009, on essentially the same terms as before (except in UP and Bihar, with an improvement over the terms of 2004). In all other states the terms remain essentially the same. What this reflects is the willingness of a range of parties to ally with the BJP in response to the perceived swing in its favour according to forecasts, as against an unwillingness to ally with the Congress, patently so in Tamil Nadu, former Andhra Pradesh and West Bengal, as it is seen to be a party in relative decline and unlikely to be able to form a government and offer a share of power at the Centre. To the extent that pre-electoral coalitions make a difference, the odds seem to be against the Congress.
Coalitions at the state level also depend on various local factors. In bipolar state party systems (almost all major states except UP and currently Bihar), either of the two leading parties are attractive coalition partners for significant third or fourth parties, provided there are no basic ideological contradictions or aversion of their voters. One can, therefore, expect coalitions between the BJP and third parties not dependent on religious minority votes when the former is a leading party in a state, or between a regional party, even if dependent on minority community votes, if the latter is the leading party and can check the BJP as a junior partner at the state level (e.g. JD(U), earlier TDP, Trinamul, BJD). Likewise, one can expect coalitions between the Congress and a regional party if the Congress is a third/fourth party and the regional party’s main opponent is another party or the BJP (e.g., DMK/AIADMK in earlier years, or RJD in Bihar, Trinamul in West Bengal currently).
One would also expect pre-electoral coalitions, negotiated under pressure of vote aggregation under conditions of uncertainty about which party has how much popular support, to be ideologically indiscriminate, relatively speaking. How will a victory based on a diverse pre-electoral coalition affect post-electoral government formation? Depending on the exact numbers and the relative strength of the leading party, a government so formed will be ideologically, geographically and socially more diverse than the leading party’s ideological preferences, particularly if the leading party turns out to be the BJP, and still more so if post-electoral allies are needed.
Dr E Sridharan is the Academic Director of the University of Pennsylvania Institute for the Advanced Study of India (UPIASI) in New Delhi. E Sridharan would like to thank Adnan Farooqui for help with the data.