Exit Polls Prove the Modi Increment

BY ASHOK MALIK

For everything that is true of India, an old line goes, the opposite is also true. Keeping this caveat in mind would be useful while analysing the results of exit polls that were telecast on several television channels on Monday evening.

In bipolar democracies, exit polls rarely go wrong and are usually seen as near infallible. Yet even in such polities there have been occasions, in tight finishes, when exit polls have proved to be inaccurate. In India, exit polls - like opinion polls - suffer from the additional handicap of converting vote share to seat share in a three or four-cornered contest. Depending on the wasted vote - the extra vote a candidate gets in a seat that only adds to his or her winning margin but is not needed to simply defeat all other candidates - an individual party could find its numbers inflated or reduced in a state. (India Votes 2014: Complete Coverage)

Take Uttar Pradesh, where the BJP is projected to win about 40 per cent of the popular vote. This could yield it 45 of the state's 80 seats; it could as easily yield it 60. An analogy would be appropriate here. In the IIT entrance examination, the gap in scores - not to speak of the scholastic aptitude - of the candidate who finishes second and the candidate who finished 52nd is probably very tiny. Yet, that tiny gap can yield a very different rank. Something like this could well happen in an election.

While most of the exit polls gave the BJP a similar tally for Uttar Pradesh, states like Tamil Nadu, Orissa and Bihar threw up large variations, even if the overall, all-India numbers were within a narrow range for all the exit polls save one. Why did this happen? Frankly, attempting to decipher state or more so constituency-level results from a national sample makes for great television but is actually risky psephology. It is one thing to attempt a state-specific poll and choose a random sample from that state. It is quite another thing to draw up an all-India sample and then try and read its opinion in terms of state boundaries.

In the May 12 exit polls, the Rashtriya Janata Dal's seat performance in Bihar was recorded extremely unevenly. Similarly, in Orissa (total seats: 21), one exit poll gave the BJP 23 per cent of the vote and one seat, and the Congress 27 per cent of the vote and five seats. Another exit poll gave the BJP 10 seats. It is likely that the BJP vote percentage was within the same band for both polls. Both polls could be completely honest but only one can possibly be right (or of course they can both be wrong).

Given this, the exit poll data is more important for vote share than for seat predictions. The big picture to emerge from the exit polls is that the BJP is now head-and-shoulders above the Congress when it comes to being India's leading national party. In 2009, the BJP had an all-India vote share of 18.8 per cent. In 2014, the vote share is expected to almost double.

This is a dramatic jump and much more than the seats the party may or may not win, it suggests the impact Narendra Modi has had. For anybody with a doubt as to the protagonist of the 16th Lok Sabha election, the CNN-IBN-CSDS-Lokniti poll was revealing. A massive 44 per cent of NDA voters said they "would not have voted for the alliance had Modi not been the PM candidate of the BJP".

Most of the exit polls argue this incremental or Modi vote will be enough to take the NDA to a majority. They may or may not be right and may or may not have got their vote-to-seat conversions right. In that sense, they may have misread the quantum and impact of the Modi increment. Even so, it is unlikely that any verdict on May 16 will be able to completely deny the existence of the Modi increment. The man is now in a league of his own.