Why Do Exit Polls Often Go Wrong? Think About Why Elections Have Secret Ballots

Secrecy is a key part of a process in which voters are often subjected to all sorts of pressures and even threats.

We love conspiracy theories in India. I am not surprised that after stock markets zoomed following the exit polls that showed Prime Minister Narendra Modi's BJP-led National Democratic Alliance winning upwards of 350 seats in the 542-member Lok Sabha and then crashed as the real figure came close to 290, tongues started wagging. Some want to probe if exit pollsters were in league with stock market manipulators.

However, as someone who has covered elections from the ground, I think there is another way of looking at it, and that is exactly what some of the erring pollsters are admitting sheepishly. The voters hid the truth or played around it!

To understand why, I am tempted to share a joke from my college days about a common puzzle in which you are asked, "An ant is saying there are two ants to my front and, another says there are two behind my back, and the third one says there is one ant behind and one in front." The clever, classic answer is "three" but in a wicked standup comic act, a boy once said: "The third was lying!"

Lies are often discounted in surveys, but there is a theory behind this. People forget that democratic elections in politics involve secret ballots for a reason. Secrecy is a key part of a process in which those who vote are often subjected to all sorts of pressures, temptations, and even threats.

The poorer or less privileged in education, community, or gender that voters are, the more likely they are to either fudge, lie, or mislead in some way. I call this the problem of power asymmetry. Unlike shareholder meetings of listed companies, in which voters express their views openly, the poor are often hustled into saying or voting in a certain way in political elections.

There is also something called the game of maximising benefits and minimising losses.

I often like to joke that Indians love to lower their caste when they seek a government job but elevate it when they want to get married. Anecdotal evidence suggests that caste and economic identities are often fudged to help get state benefits. This is not much different from businesses or individuals who show higher income levels, cash flows or asset details to secure bank loans.

In the instance of this year's general elections, especially in Uttar Pradesh, where pollsters and pundits have gone horribly wrong, I am not surprised. Chief Minister Yogi Adityanath became notorious for his "bulldozer" governance, in which he knocked buildings down in his eagerness to prove that he enforced law and order. The Ayodhya temple to Lord Rama was built after putting thousands of poor people in trouble by displacement and state pressures.

In such a context, do you expect those who are keen for a regime change to answer questions in simple formats? I should think not.

I have my own experience to share. In 1991, I got much of my election expectations wrong because, as a naive young reporter, I used to take campaign claims at face value. Political campaigners often exaggerate their numbers to encourage or motivate the cadres and boost their morale. That partially explains the "Ab Ki Baar 400 Paar" slogan of the BJP.

In BJP's Era, the Poor Reply With Strategic Silence

Political activists also know a strange fact about Indian elections. Many Indian voters do not like to "waste" their votes and hence join the perceived direction of a likely winner. This is one of the factors that lead to "waves" in Indian elections. Rival parties often claim "waves" precisely because of this. Undercurrents that may be truer or closer to reality may be missed in this melee. That is what seems to have happened this year.

I had my own success while covering the 1993 Himachal Pradesh assembly elections, in which the incumbent BJP government was expected to win. I was determined not to take campaign claims at face value. When I interviewed poorer villagers, I studied the tone of their voices, their eye movements, and their tendency to keep things close to the chest. I surmised that this looked like an anti-incumbency mood. I stuck my neck out and pointed to a BJP defeat, and I won praise from my editors for this.

But this can get tricky for exit pollsters whose troops ask systematic questions. Interpreting body language and the tone or choice of words is as tricky as taking derivative positions in futures and options in stock markets. Exit pollsters have virtually no choice. They have to add the devil of applying normal intelligence to detailed answers to the deep sea of projecting seat shares based on estimated vote shares. It is a slippery slope worsened by strategic red herrings thrown by poor voters!

I am more inclined to believe conspiracy theorists who say TV channels that hype these polls as exclusive events are likely gainers than stock market punters. Eyeballs are more certain than stock movements.

As someone who has seen shrill campaigns by the BJP for decades, I may add that this party is more vulnerable to exit polls going wrong. Its own supporters are the loudest in claims of success, the most belligerent in campaign messages, and the most aggressive in door-to-door canvassing. They exude an aura of power and dominance that can backfire in some quarters.

Here is where the poor reply with strategic silence, while the more affluent supporters dare to express what they feel. Asymmetry of power often goes with asymmetry of articulation. Words are mumbled, coded, or encased in meaningful eye movements that often cast a long shadow between expectations and reality.

I am not surprised that in 2004, then Prime Minister Atal Behari Vajpayee lost power as the BJP's "India Shining" campaign bombed. Shining for whom was the question that lingered in the air as the Congress-led Manmohan Singh government took power in a coalition called the United Progressive Alliance?

Pradeep Gupta of Axis My India, who has been cheered for getting right many exit polls in the past, was gracious enough to admit his mistake after this week's counting debacle. As a Reuters report noted, his survey "failed to capture a shift among voters in the less privileged sections of society in Uttar Pradesh, West Bengal, and Maharashtra."

This is in spite of the fact that 900 people were employed to conduct face-to-face interviews after buttons were pressed on electronic voting machines to track a huge, systematically classified sample of 582,000 voters. Gupta also said women asked males in their families to respond on their behalf in many cases. This is tricky because it is increasingly felt that women and men in the same family can often vote for different families.

What remains, in the end, is the power of silence inside the voting booths in the largest democracy in the world, notorious for its noisy, contentious campaigns full of misleading claims and loud propaganda.

(The writer is a senior journalist and commentator who has worked for Reuters, Economic Times, Business Standard, and Hindustan Times. He can be reached on Twitter @madversity. This is an opinion article and the views expressed are the author’s own. The Quint neither endorses nor is responsible for them.)