New Delhi: As polling for the 2019 general elections ends today, all eyes will be on the exit polls. The India Today network has claimed that its pollster, Axis MyIndia, has a 95 per cent success rate with exit polls.
Axis, interestingly, is the only pollster that puts out exit poll numbers in the media but does not forecast seats before the elections begin.
As news channels counted the “poll of polls” two days before the 2019 Lok Sabha elections began on 11 April, Axis was missing. It had stopped doing pre-poll seat forecasts two years ago, following assembly elections in five states — UP, Uttarakhand, Punjab, Manipur and Goa — in February and March 2017. Axis only does exit polls now.
Also read: All you want to know about Lok Sabha elections 2019 — seats, voters, parties and more
Political clients
Pradeep Gupta, who heads Axis, gives three reasons for this. One is that he also has political clients, and putting out opinion polls numbers just before elections upsets someone or the other. He doesn’t reveal who his political clients are but sources say the main client is the ruling Bhartiya Janata Party (BJP), apart from some state governments.
Opinion polls, Gupta says, alienate too many people — from party bosses in Delhi to party workers in the villages. “A survey will show someone as the winner and at least two other parties will be unhappy with the result,” he says.
This affects his relationships with parties since he also conducts surveys for them. Political parties and governments have been hiring Axis to conduct surveys to get actionable intelligence of election issues. It’s more useful to do surveys to see the impact of government schemes or figure out which candidates a party should choose.
It is also a conflict of interest, as some might wonder if a poll forecast put out in the media is influenced by the pollster having one of the parties as a political client.
Safety of surveyors
One reason why Axis has stopped doing pre-poll seat forecast is the hostility its surveyors face. Axis claims to have a team of 1,000 surveyors across India who visit the same neighbourhoods and villages at least once a month, sometimes two. When the surveyors visit after an opinion poll, to, say, do the exit poll, they often meet hostile locals who were unhappy to see the opinion poll result on TV.
In states such as West Bengal, Kerala, Jharkhand and Odisha, Gupta says Axis surveyors have been beaten up and dragged to police stations. On one occasion, in Rajasthan too. In West Bengal, one surveyor found himself in jail for 10 days.
In Andhra Pradesh, Gupta says, Jagan Mohan Reddy recently said the Telugu Desam Party was conducting door-to-door surveys to identify YSR Congress supporters and remove them from voter rolls. This has made it virtually impossible to carry out a survey in Andhra Pradesh. Gupta even spoke to an aide of Jagan Mohan Reddy but nothing came out of it.
Also read: Exit polls: How they are conducted, their relevance & accuracy
There are places where people — especially village sarpanches — want to know if the surveyors have taken permission to conduct a survey. “Our surveyors try to reason that no such permission is required,” Gupta says. “I’ve tried getting government officials to give me a document that clarifies that no such permission is required. The permission is called the Constitution!”
The conflict in some places lands up at the police stations, and the police, too, are often suspicious of the surveyors. There have been occasions when locals have snatched away the tablet computers of the surveyors, leading to an altercation that ends at the police station.
“Considering political surveys are only 10 per cent of our work, it’s not worth it,” says Gupta, whose company primarily does market research surveys.
Long election
Besides, he says, it is meaningless to do a survey on the eve of polling, because so much changes in a multi-phase election. “For instance, the Congress-AAP alliance were in talks in Delhi and Haryana until the very end. How could one have forecast Delhi seats when voters didn’t know if Congress-AAP alliance was happening? It is not useful to do a survey when candidates have not been announced in many places, campaigning is still on and people are in the midst of making up their minds,” he says. “Pre-poll surveys are too speculative and make everyone unhappy, except the news channels who gain TRPs.”
Axis did do a long series of opinion polls for India Today called the “Political Stock Exchange”, which gauged public mood but did not put out any seat forecast.
Track record
Axis has been putting out opinion and exit polls since the state assembly elections of December 2013 and Gupta says they’ve got only one election wrong so far — the Tamil Nadu assembly election in 2016. It predicted a loss for the then Jayalalithaa-government but her AIADMK retained power, becoming the first incumbent to do so since the MGR days.
The pollster shot to prominence with its exit poll for the 2015 Bihar assembly elections. CNN News 18, the news channel that commissioned the exit poll, had refused to run it because it showed an unexpected sweep for the Mahagatbandhan whereas other exit polls were playing it safe.
Axis’ exit polls tend to get not only the direction right but often the numbers too. In the Telangana assembly elections, its exit poll said the TRS would get between 79 and 91 seats — the result was 88. Few were willing to forecast this scale of the TRS victory.
In the 2014 Lok Sabha election, however, Axis like the others could not foresee the BJP winning a clear majority on its own.
Axis gets it right more often than others, Gupta says, because it does door-to-door surveys, not over the phone, and surveys every single constituency separately. That’s how they forecast not just seats but also which seats, although that is data they don’t usually put out.