Debating India’s election cheat sheets

May 14, 2014

(Any opinions expressed here are those of the author and not of Thomson Reuters)

As the sun set on the final phase of polling in India on May 12, newsrooms were waiting impatiently for 6.30 p.m. — the deadline set by the Election Commission for airing survey results on post-poll predictions.

Elaborate studio sets packed with guests and news anchors flanked by psephologists armed with data sets were all waiting to declare that Narendra Modi is coming to Delhi.

The 63-year-old prime ministerial nominee of the opposition Bharatiya Janata Party (BJP), perhaps one of the most revered and polarizing figures in India’s political history, was being given a unanimous verdict. Once the clock struck half past six, Indians were glued to their television screens.

“Abki Baar Modi Sarkar” (This time it’s a Modi government) — the slogan championed by the Modi campaign was being splashed across media platforms, but what stands out is the variation in tallies. The six sets of exit polls predict that the BJP-led National Democratic Alliance (NDA) would manage the magic figure of 272 seats — the simple majority needed in the 543-member lower house of parliament to form a government — or at least get close to it.

The good news for the BJP is that it is emerging as the single largest party, with impressive performances in non-BJP ruled states such as Maharashtra, Bihar, Uttar Pradesh as well as in Karnataka and the recently bifurcated Andhra Pradesh.

The exit poll numbers confirmed the incumbent Congress party’s worst fears, foreseeing its coalition United Progressive Alliance’s (UPA) lowest election tally with an average of 100 seats overall and an embarrassing rout in various states. Regional parties were being seen as kingmakers in this election — with the Jayalalithaa-led AIADMK in Tamil Nadu and Mamata Banerjee’s Trinamool Congress in West Bengal emerging as the two largest regional parties with more than 20 seats each.

If there was one upset in the big picture, it’s the result predicted for Arvind Kejriwal’s Aam Aadmi Party, a new entrant in India’s political process after its success as a nationwide anti-corruption movement.

Election pundits say that after sweeping the Delhi assembly polls in December, the party spread itself too thin for the general election by contesting over 400 seats. Stuck with a tally of four-seven seats, the exit polls don’t see the AAP as a game-changer.

The trouble with these numbers are the variations in tallies, which could affect larger trends. For instance, the numbers projected for the NDA range from a conservative 257 to a decisive 340 seats, which can be problematic if one is second-guessing possible post-poll alliances. In fact, one exit poll predicts 291 seats for the BJP, a verdict that would mean a clear majority for the party that rules out any dependence on allies.

Despite the optimism of the exit polls, media reports have said the BJP is keeping all options on the table and looking to increase allies. On the other hand, only three of the six exit polls put their neck out in terms of seat numbers for the Congress party. The tally provided is for the UPA, ranging from an abysmal 70 seats to an optimistic 135. With such a difference in projected numbers, it could be argued that the poll predictions are far from decisive and come with a lot of caveats.

One of the most popular tweets in India on Monday said that “the first rule of Indian polls: don’t trust the exit polls”. Most national editorials began the discussion with cautionary tales of India’s experience in both 2004 and 2009 when poll pundits got their calculations wrong. In 2004, the numbers backed the NDA but the result was a shocker with the UPA coming to power. Again in 2009, NDA was hailed as the clear winner, but the people’s mandate got UPA a second term.

Many have blamed the sample size of these surveys for past blunders. This exit poll has also seen a wide range in terms of survey size — from 30,000 to 500,000 voters. Praveen Rai, a political analyst at the Centre for the Study of Developing Societies (CSDS) in New Delhi, has made a strong case for re-evaluating both the methodology and intrinsic challenges unique to the Indian case study, to find answers to why exit polls have gone wrong in the recent past.

Complicating this further is India’s vast socio-cultural diversity, block voting practices and so on. Also, in a polarized election many voters may respond under pressure, making calculations go haywire. Exit polls, after all, are surveys conducted among voters as they come out from the polling station. With so many overlapping factors and challenges, the margin of error is arguably high.

Before every election in India, we’ve seen a debate on whether opinion polls and exit polls should be banned to ensure an election free of influence and abuse by political parties. Each time, supporters have quoted the constitutional right of the freedom of speech to subvert any such attempts.

In fact, many ‘experts’ have told me cheekily, “why ask me about a ban, do you want an entire industry out of work?” Their media colleagues seem to have taken this to heart with TV networks competing with each other, promising over 100 hours of coverage till the results are out.

Notwithstanding the hype, the sensible thing to do is to take these numbers with a pinch of salt and be happy with the high voter turnout in this election. Perhaps the breathless election coverage needs to take a cue from the cautious market sentiment.

Despite the markets rallying on Monday riding on optimism generated by the exit polls, brokerages have warned investors that the numbers may not reflect reality. It’s clear India is poised on the cusp of change; a little caution and patience is all that is required.

Click for Reuters coverage of Election 2014

No comments so far

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see