Exit Polls: Manipulating the Odds or Predicting the Winner?

Banning exit polls is an impractical idea when elections are rigged with money power, writes Gautam Mukherjee.

Banning Not a Good Idea

• Among the four Bihar pre-poll surveys done so far, two have given heads-up to the Nitish-Lalu combine; the other have predicted an NDA victory
• There is little room for suspicion that these are manipulated reports, faked to favour one side over another
• Fallout of banning such poll surveys creates a buzz in the satta bazaar where speculation leads to betting
• As it is, money power is in full display during polls with contenders splurging during the campaign
• Researchers have indicated poll surveys likely to yield accurate results if they focus on what voter wants

Election pre-polling or exit polls, can be a controversial ingredient and a shadow player. But only if viewed as a staged and paid- for manipulation, a rigged projection. Those who want all election polls banned subscribe to this idea, supported invariably by the side that looks bad in early polling.

However, whatever be the pros and cons, the Election Commission has banned exit polls between October 12 2015, when the first phase of polling for the Bihar legislative assembly begins, and November 5 when the fifth and last phase of polling ends.

Notwithstanding the elimination of exit polls in this instance, except at the close, and before the results are expected on November 8; most people see polling as a professional exercise to predict who will win, and in a secondary and broader sense, who, and what, do the voters want. The jury is definitely out on the ability of polls to influence the outcome.

Opinion Polls or Faked Reports?

In case of the Bihar legislative assembly elections for 243 seats, it looks as yet, after four pre-poll surveys, that it will be a closely contested thing, with two of the polls giving it to the Nitish-Lalu combine and the other two to the NDA. The phrase ‘neck-to-neck’ is being frequently used, two days before the bugles sound. 

The latest poll, announced on October 9 does hand it to the NDA. It will get more popular votes, they all suggest, but the people of Bihar would like to see Nitish Kumar return as chief minister. While this may be an impossibility if the NDA does win, it suggests a close contest also.

One poll suggests a gap of 42% of the vote to NDA, against 38% in favour of the Nitish/Lalu/Congress combine. This gap is narrow, and within the margin of error of such pre-election polls.

So at least, in this instance, four times over, with the pendulum swinging one way sometimes and otherwise the other, with findings from different pollsters, there is little room for suspicion that these are manipulated reports, faked to favour one side over another.

Speculation Leads to Betting

The pre-poll surveys of varying size, have apparently not materially affected for or against - the ‘odds of victory’, ‘campaign morale’, ‘turnout’, the illegal ‘betting’ such as it is, and/or the ‘fundraising’ .

But banning the post poll exit surveys between the phases, leaves it to the whisperers, over nearly a month, and the satta bazaar, to indicate which way the wind is blowing. The satta bazaar, on the eve of the first phase of election, is saying that NDA will be able to form the government on its own.

Certainly, on both sides, the main contenders, have been spending a lot of money on the campaign, and will attempt to continue doing so between the phases. This even as there will be 80 observers just to watch money flows used to ‘buy votes’ state wide; and 243 other general observers,  throughout, and in between the polling on 12th, 16th, 18th, of October and the 1st and 5th of November.

All 62, 779 polling booths will be overseen by central armed police and monitored by drones, probably for the first time.

Deciphering Voter Psychology

Researchers in America, the birthplace of election polling, suggest that it is easier to predict who will win over who voters like. The is what Justin Wolfers and David Rothschild, economists at Microsoft Research had to say: “ Polls probing voters’ expectation yield more accurate predictions of election outcomes than the usual questions about who they will vote for”. Of course, this is America they are talking about.

But is voter psychology in urban and rural Bihar very different? In America ‘voter expectation’ surveys only got it wrong 8% of the time out of 214 surveys taken between 1932 and 2012. Neil King Jr wrote in the Wall Street Journal (WSJ), that the ‘expectation question has ten times the potency as who will you vote for’.

The rest will be made plain on November 8,

(Gautam Mukherjee is a plugged-in commentator and instant analyser)