Sanjay Kumar's column: Vote-share forecasting is more important than seats

The exit polls released for Assam, Tamil Nadu, Kerala, Bengal and Puducherry on the evening of April 29 did not bring any solace to many leaders and parties. The reason is, the exit poll projections for at least two states — Tamil Nadu and Bengal — are quite different from each other. But in Assam and Kerala, there seems to be a broad consensus among almost all agencies on which party or alliance is winning. Most exit polls give a decisive lead to the BJP-led coalition in Assam, while the Congress-led UDF indicates a victory in Kerala. But there is uncertainty about the outcome in Bengal and Tamil Nadu. Some agencies have even considered it appropriate to postpone the release of exit poll estimates for Bengal as the second phase of polling there ended on April 29 and they want to take some time to understand the data better. I would call it a prudent decision, because many times voting continues due to long queues even till the official time is over. However, it should also be remembered that stopping one more day does not guarantee that the estimate of seats will become more accurate. After all, in Tamil Nadu, polling ended on April 23 and the agencies had enough time to carefully analyse the data, yet the estimates of different agencies differ from each other there as well. Almost all the exit polls for Assam and Kerala are pointing in the same direction, but this is not because the agencies had extraordinarily more time for data collection and analysis. There could be many reasons behind this- such as relatively little political competition, the nature of the challenge posed by the opposition to the ruling party or coalition, and other factors. Nevertheless, the exit polls of Assam and Kerala need to be read carefully. In the 2024 Lok Sabha elections, all the exit polls were missed; He had projected more than 400 seats for the NDA. Similarly, in the 2024 Haryana elections too, there was a consensus among almost all agencies about the big victory of the Congress. The same situation was seen in the 2023 Chhattisgarh elections, where the exit polls predicted a certain victory for the Congress. The success record of the exit polls in the Bihar Assembly elections was mixed. Many had predicted the NDA's victory, but no one had predicted such a big victory. The exit poll projections in Tamil Nadu appear to be somewhat complicated. Of the four exit polls that caught my eye, three have predicted a victory for the DMK alliance, with the AIADMK alliance at the second position and the Vijaya-led TVK at the third position, while one exit poll in the context of Tamil Nadu stands out as an exception. The state which has held the maximum breath is Bengal. The elections have been fought in a very tight contest. Most of the exit polls made public so far are predicting the BJP's victory in Bengal, though there is one exit poll that stands out as an exception, aside from this trend. It is difficult to say at this time how accurate these estimates will prove to be compared to the actual results due on May 4. But I believe that if poll-analysts are proven wrong, it could be mainly due to an error in the vote-share estimates. Even a small change in the vote-percentage can make a big difference in the number of seats. We have seen elections where a party has won more seats despite having a lower vote-percentage than its rival. This complexity is further compounded in our electoral system – the first-past-the-post system. In this, a candidate wins by one vote or by one lakh votes, in both cases he gets only one seat. The recent exit polls, especially in the context of Tamil Nadu and Bengal, have added to the confusion and anxiety instead of predicting how the votes would have been cast. The number of exit polls is also relatively low this time. (These are the author's own views)