ADVERTISEMENT

Elections 2019 Exit Polls: The Devil Lies In The Detail

Some of the 2019 Lok Sabha exit poll numbers don’t really add up.

A microscope is seen in a lab. (Photographer: Geraldine Hope Ghelli/Bloomberg)
A microscope is seen in a lab. (Photographer: Geraldine Hope Ghelli/Bloomberg)

The exit poll results have been announced. Modi 2.0 looks possible now. The Sensex is up 1,400 points. Television channels witnessed dramatic moments on Sunday with supporters of both political sides stopping just short of landing blows, live in the studios. Most TV channels focussed on seat projections, but were rather silent about vote share, and divulged very little data. As Yashwant Deshmukh, the CEO of C-Voter India, has admitted, the job of exit polls is to project vote shares and indicate the direction of the trend.

Seat shares are derived using historical conversion ratios, adjusted for current political factors, and have often gone wrong. On seat shares, exit polls were wrong in 2009 as well as 2014 in judging the extent of the United Progressive Alliance and National Democratic Alliance victories in those years. The devil lies in the detail, as the famous adage goes.

All-India Projections

The average of vote shares projected for the National Democratic Alliance by four agencies is 43.98 percent (+5.48 percent from 2014) while that for the United Progressive Alliance is 25.8 percent (+2.5 percent).

However, the range projected is quite big: NDA: 41.1-48.5 percent, UPA: 25-31.7 percent, and Others: 26.5-28.1 percent.

Opinion
Elections 2019: Here’s What Brokerages Made Of The Exit Polls

All India Seat Conversion

  • All four exit polls predict a lower ratio for the NDA compared to 2014.

This is possible, as the vote share gains in West Bengal and Odisha may not translate into seats as easily as they might in stronghold Hindi heartland states. Also, the BJP’s conversion ratio is likely to decline in Uttar Pradesh, with similar vote tallies as 2014, but a lesser number of seats. Here again, there is a big range offered – from 5.7 times to 8.7 times. The Ipsos number is near the 2004 result for the NDA.

  • All four polls predict a higher conversion ratio for the UPA compared to 2014.

This may happen as even a lower vote share gain in Tamil Nadu (10-12 percent) is seen translating into a higher number of seats (30+). Again, the range offered is wide – from 2.9 times to 4.3 times.

  • The conversion ratio for others is a cocktail, ranging from 2.5 times to 5.2 times.

Two polls are predicting an improvement (Ipsos/C-Voter), and two see a decline (Axis/VMR). While the SP-BSP mahagathbandhan could get more seats in UP than last time with a similar vote share, an improvement in the ratio is likely to be offset by a decline in the conversion ratio of the Trinamool Congress and the Biju Janata Dal in West Bengal and Odisha respectively. So, it’s a toss-up.

Opinion
A Second Modi Term Needs to Be Bolder Than First

Vote-To-Seat Projections In UP, West Bengal, And Odisha

  • Axis and C-Voter suggest that the MGB has not been able to seamlessly transfer votes to each other. The leakage is 3 percent as per C-Voter, and 10 percent according to Axis. Anything less than or equal to 5 percent is healthy, as was seen for the JD(U)-RJD-Congress alliance in the 2015 Bihar assembly election.
  • Axis and C-Voter predict a vote share gain for the NDA in UP. C-Voter is predicting a 0.5 percent gain, while Axis expects a 4.4 percent jump. The four percent difference between the Axis and C-Voter surveys is resulting in Axis’ NDA seat-projection being higher by 27 seats.
  • The C-Voter survey plays it safe while Axis has put out an aggressive projection. A 48 percent vote share for the NDA would mean further consolidation of upper caste (15 percent) and non-Yadav OBC voters (10 percent) in its favour, with a minimum dent in the Yadav and non-Jatav support of 2014.
  • Both Axis and C-Voter suggest that the BJP is going to significantly improve its position in West Bengal, as has been expected. The Modi-Shah push seems to have worked, if exit polls are to be believed.
  • Axis predicts the TMC and BJP will share equal honours, both in terms of seats and vote share. C-Voter predicts a more middle-of-the-road case for the BJP, with a 15 percent vote share jump, and an increase of 9 seats. A word of caution, though, elections don’t follow any logic.
  • It all depends on the collapse of the Left in West Bengal. A 50 percent decline in Left votes, as suggested by C-Voter, would lead to a 10-seat gain for the BJP. A 75 percent Left collapse would net the BJP as many as 17 more seats, as per my calculations.
  • Here too, C-Voter has played it safe while Axis has taken an aggressive position.
  • Axis and C-Voter suggest that BJP is making significant strides in Odisha as well.
  • Axis is predicting that the BJP will sweep Odisha. C-Voter suggests honours will be shared. BJP is seen gaining significant vote share at the expense of Congress which has weakened in the state.
  • Like in the other two states, C-Voter’s survey plays it safe.

To sum up, some numbers don’t really add up. We’ll get more clarity on results day. As I write this, Axis has taken off its detailed report page.

Amitabh Tiwari is a political commentator, strategist and consultant advising political parties and leaders. He was a corporate and investment banker. He advised one of the political parties in Odisha.

The views expressed here are those of the author and do not necessarily represent the views of BloombergQuint or its editorial team.