(HT file)
(HT file)

Opinion | The great electoral mandate and the wisdom of crowds

Large groups of people are often found to be collectively smarter than individuals with specialized domain knowledge

Even as the curtains were drawn on the biggest “festival of democracy" and the National Democratic Alliance (NDA) prepares to form another government, exit polls and psephologists seem to have got 2019 right. While there had been much speculation over the veracity of exit polls, especially in the wake of some of these going completely wayward, the question is: Is there a way for superior and more credible predictions of electoral outcomes than what is currently practised by drawing upon the opinion of experts? Can political parties adopt some of these methods to improve outcomes on D-Day?

In his 2005 book, The Wisdom Of Crowds, James Surowiecki posited that under certain conditions, large groups of people are collectively smarter than even individual experts when it comes to problem-solving, decision-making, innovating and predicting.

The notion of collective judgement itself can be traced back to the observation by statistician Francis Galton in 1907, based on an experiment involving a “guess the weight of the ox" competition at a country fair. Galton discovered that an average of the 800 entries (1,197 lb) produced a remarkably accurate estimate of the actual weight (1,198 lb), beating not only most of the individual guesses, but also those of alleged cattle experts.

The conditions for a crowd to be “wise" include ensuring diversity and independence of opinions, sufficient individual knowledge on the part of each decision-maker to form his/her own opinion, and the possibility of aggregating the individual opinions of the crowd into one collective decision.

We decided to test this concept to predict and test the Indian electoral outcomes of 2019. A simple questionnaire, ensuring that the four conditions were met, was administered to a group of 215 business school participants from diverse backgrounds and demographic groups under conditions that ensured the independence of elicited opinions. This was done two days before the exit poll results, at a time when most of the country’s states had finished voting.

The respondents’ ages ranged from 21 to 59, and they belonged to different income groups, as also occupations, with several belonging to business families and bearing corporate backgrounds. By design, the survey covered people from all parts of the country.

The survey was based on four questions: (a) the opinion of the respondent on the party/alliance that was likely to win the general elections of 2019; (b) their opinion on who they thought ought to win; (c) the probabilities of parties/alliances winning and forming the next government; and (d) the most important factor that would influence voter decisions in the 2019 elections.

The survey results were interesting in that there was a clear difference between what the crowd thought would happen versus what they thought should happen. A large proportion (62%) predicted that the Bharatiya Janata Party (BJP) and its allies (NDA) would win these elections, while 32% felt that the BJP alone would win the elections. The proportion of people predicting a third front (non-BJP-non-Congress) was abysmally low, at 2%, while the Congress and its allies (United Progressive Alliance or UPA) itself could muster only 3% of the share.

On the question of which party they thought deserved to win and form the next government, 50% thought it should be the BJP alone, while 34% felt it should be the BJP with allies. The proportion of people who thought a Congress alliance or a third front should win and form the government was 5% and 7%, respectively, while 3% felt that no one deserved to win.

The BJP and its allies, respondents felt, had a 62% chance of forming the government, while the odds of the third front and Congress of forming the government were placed at 14% and 15%, respectively, while 9% expected a hung Parliament.

On the most important issues that would influence voter decisions, 29% (the highest proportion) marked out economic growth as the most important factor. The “There is No Alternative" (TINA) factor followed with 17%, while unemployment and job losses came third at 15%. Interestingly, it was the economic factors (growth and unemployment), which dominated this poll, at 44%. Other factors surveyed included national security, especially India-Pakistan tensions (17%), social security (4%), caste and communal divides (7%), political ideology (6%), presence of local candidates (3%), and candidates of similar caste/religion (3%).

The results of this survey seem to match the actual outcomes. However, the process of eliciting such opinions (as also the outcomes) does generate interesting insights. There may be a discrepancy between what people think ought to happen versus what they think would happen. The former is a metric that some polls capture and may be different from the latter, which forms the basis of collective judgement. However, it is collective judgement that is more likely to converge to the actual outcomes than the former metric.

Again, the collective judgement of our sample reveals that it was not Pulwama, social security or TINA, but economic issues that would still have reigned supreme in the mind of the typical voter as he voted for the BJP. The moot question then, with collective judgement converging to the actual result, is this: What lessons should opposition parties draw from this exercise?

Tulsi Jayakumar is professor of economics at S.P Jain Institute of Management and Research

The author acknowledges the discussions with Dr. Ranjan Banerjee and Dr.Anshul Verma in writing this piece

Close