Why predicting a poll wave is tough

Should we define a “wave” in terms of vote or seat share of the winner? Or should we think of it in terms of the gap between the winner and the runner-up? It is likely that if the SP wins with even a slender majority, it will be described by some as a wave election, but at what vote- or seat share-level would this election look like a wave for the BJP? For example, though the BSP won only a simple majority in 2007, the backdrop was significant: No party had formed the state government on its own in two decades. Similarly, the BJP’s victory in 2017 shattered many records in the state. And yet, the gap between these two wave elections was 10 percentage points in votes and 100 seats. Therefore, in India’s first-past-the-post system, a swing of 2-3% votes can result in massive shifts in seats, making it almost impossible to pick these alignments without a large post-poll survey.

Election prediction is difficult in India, but picking up a wave is even tougher — beyond the grasp of even analysts who correctly predict the winner. Why?

First, our judgment about the perceived winner is often merely an informed guess, based on what fellow travellers are picking up from the ground. Everyone has their favourite mode of gathering information. Some dissect pre-poll surveys and try to find patterns in historical data. Many others travel to far-flung areas and speak to voters and politicians to reach a conclusion. Then there are those who decipher the body language of leaders, the tone of the campaign, and crowd presence in political rallies. While each of these techniques tells us something noteworthy, the triangulation of information is the most productive path. Unfortunately, such triangulation is rarely a part of our analysis and discussions.

Second, we often get swayed by colourful idioms such as “sasta rashan to gareebon ka haq hai [free ration is the right of the poor]” that certain voters use to indicate their choice. These are not representative voices, rather the voice of voters with strong sympathies for one political party or the other. They rationalise their political preferences using localised idioms. Thus, we not only get muddled information on what issues are dominating voters’ choices, but also let our previous understanding about voters’ caste or community identity indicate if there is a wave or undercurrent in favour of a party. For example, even if someone speaks to 100 voters at random, they are likely to meet, on average, 40 BJP supporters, and a few extra supporters in either direction would produce more noise than give a clear signal about the probable outcome. Very often, we are not cautious while extrapolating macro outcomes from such interactions.

Third, the mysterious, silent voter is always lurking in the background and pushes conclusions that are in line with our bias about the preferred outcome. The underlying premise behind the silent voter phenomenon is that a significant chunk of the electorate, especially the bottom half of the social pyramid (women, lower castes, and the poor) is reluctant to publicly admit their displeasure or admiration for a political party due to fear of physical threat or loss in beneficiary status. The phenomenon first surfaced in the late 1980s when parties started overtly mobilising lower castes, but political empowerment, social media and better law and order have ensured that the impact of the silent voter is, ironically, much more muted today. The empirical evidence on this phenomenon is relatively thin as well, though everyone, from politicians to election observers, keeps adding to this mystery.

A substantial section of the Indian electorate makes up its mind close to the polling day. Such voters are unlikely to discuss their choice, even in personal interactions, leave alone to election observers. But this reluctance to articulate often gets attributed to a self-conscious decision not revealing true preferences. It is likely that the distribution of support for various parties is largely in similar proportions as it is among more vocal supporters, and the net result of different types of “silent voters” in swaying results in one direction is negligible.

Why do terms such as “wave” or over-interpreting colourful idioms expressed by voters harm the election studies enterprise? It tends to stipulate every election as critical, rather than normal. In reality, very few vital elections realign the social basis of power for a substantial period. The 1989 general election, for example, presaged the decline of the Congress, the rise of BJP and the dominance of regional outfits — trends that were confirmed in subsequent polls that decade. Similarly, despite the common understanding of 2007, 2012, and 2017 being “wave” elections for UP politics, only the 2022 election results, wave or not, can confirm if the social realignment that happened between 2014 and 2019 has long-term potential.

Rahul Verma is fellow, Centre for Policy Research, New Delhi The views expressed are personal