Who has the edge in the elections—NDA or UPA?

The NDA seems to be better placed for the Lok Sabha polls, suggests an analysis of the 2014 election numbers and the assembly poll results since then

Ten weeks ago, the momentum in India’s polity seemed to be in favour of the Congress party. Fresh from its victories in three key Hindi heartland states in December 2018, India’s principal opposition party seemed ready to put up a strong fight against the ruling Bharatiya Janata Party (BJP).

But since then, the momentum seems to have shifted against the grand old party of Indian politics as its younger rival has managed to stitch up effective regional alliances to boost the prospects of a new National Democratic Alliance (NDA) alliance government in May 2019.

A Mint analysis of the Lok Sabha (LS) numbers, taking into account the election results of state assemblies since May 2014 as well as factoring in erosion of votes for incumbents over time (anti-incumbency) suggests that the NDA could end up winning 252 seats in the next LS, 20 seats short of the majority mark (272).

If one were to assume that each party receives a similar share of vote in each constituency as in 2014, the NDA will win a comfortable majority in the next LS, with 319 seats.


However, it is unlikely that party vote-shares will remain static over five years. It is likely that incumbents across states will face some erosion in votes. One way to factor this in is to use the data that state assembly elections have thrown up.

We use the difference (or swing) between the 2014 vote-shares and the latest vote-shares in respective assembly elections to arrive at the likely state-wise swing for each party.

Assuming an anti-incumbency rate of 1% per annum (or 5% swing over the term of the government), we then deduct the anti-incumbency vote swing from the state-wise vote swing to arrive at the final vote-share estimates for 2019. Thus, for elections held less than six months ago, we assume that the assembly-level vote-share of incumbents will remain intact in the 2019 LS polls.

But for assembly elections held, say three years ago, we factor in a 3% erosion of votes compared to the vote-shares in those elections. Using this formula, we arrive at the estimated state-wise swings for each party. To estimate the constituency-level results, we allocate the losses (or gains) in votes for the incumbent party equally to the two closest contenders (in terms of the constituency-level vote-shares) to arrive at the final results.

Using this methodology, we arrive at the base case scenario of 188 LS seats for the BJP and 84 LS seats for the Congress, with their respective pre-poll allies winning roughly a third of the seats the two parties win. Thus, NDA in this scenario ends up with more than 250 seats while the United Progressive alliance (UPA) manages only 107 seats.


Even in Madhya Pradesh and Rajasthan, the Congress does not gain a large number of seats in this scenario given that the bump in its vote-share in the latest assembly elections is not enough to give it a winning margin in many constituencies. In Chhattisgarh, the Congress is indeed able to win many more seats in the base case scenario. It is worth noting that this analysis only serves to give us a rough idea of what we could expect in 2019 if there are no more changes in alliances. The precise number of seats each party wins and the distribution of those seats could end up being different for a variety of reasons, including—but not limited to—new pre-poll alliances in the coming weeks.

One factor that may change results dramatically is the re-emergence of a Modi wave. As a previous Plain Facts column pointed out, Modi has remained the most popular politician in the country over the past five years, and it is likely that the airstrikes on Pakistani soil will only help boost his popularity.

Conversely, the lack of a consensus Prime Ministerial candidate among the opposition could hurt the Congress party and its allies. If we add a 5% vote swing nationally to the NDA tallies in our base case scenario, the electoral map changes dramatically, looking much more like 2014.


The NDA would then win a comfortable majority, with the BJP itself managing to win 240 seats of its own.

What if the Rahul Gandhi-Priyanka Gandhi duo is able to counter the BJP’s narrative and create a wave in favour of the Congress? Assuming a 5% vote swing for the Congress and its (current) allies nationally, NDA tally falls significantly short of a majority, with 187 LS seats (See chart 4). This would leave the UPA, with an improved tally of 184 seats, to be within striking distance of the NDA tally, with the Congress (139 LS seats) only marginally behind the BJP (142). In this scenario, the likes of Samajwadi Party and Bahujan Samaj Party—currently classified as ‘others’—could hold the key to government formation.


At the moment, the last scenario appears most unlikely. But as the past 10 weeks have shown, a lot can change over 10 weeks in Indian politics.

Vishnu Padmanabhan contributed to this piece.