The Tamasha of Exit Polls Continues: Axis & Chanakya slog it out

Exit polls on Gujarat and Himachal are out and the trend is clear, BJP is all set to win both these elections. Not a single poll predicts a Congress victory in either state. This is in line with the pre-polls, which, despite predicting a close contest, claimed BJP would win for a sixth straight time in Gujarat. You don’t need to be a political analyst to predict Himachal, which witnesses a change of guard at the helm, every 5 years. TV channels were full of noise, energy and drama, yesterday evening, all coming up with different theories. Anchors, commentators and politicians, tried their best to make sense of the numbers. While BJP spokespersons boasted of victory and didn’t forget to credit NaMo, Congress leaders rejected the exit poll results, highlighting how they failed in Bihar and Delhi.

Exit polls are generally more accurate than opinion polls as I put it in a co-authored article for Firstpost early this year. Though they are most of the times not spot on in seat tally, they do provide a general direction of the trend. However, it doesn’t mean that exit polls can’t go wrong.

Exit polls of 2017 showed a split verdict for all states and many of them got it wrong:

– MRC, CVoter, CSDS predicted a hung assembly in UP (BJP won)

– C-Voter predicted an AAP victory in Punjab (INC won)

– C-Voter predicted a tie in Uttarakhand (BJP won)

– Chanakya and MRC predicted AAP will be joint single largest party with INC in Punjab (INC won)

– Axis predicted a Congress victory in Manipur and a BJP victory in Goa (Hung assembly in both states)

– C-Voter, CSDS and MRC predicted BJP as single largest party in Goa (Congress was SLP)

One argument is that in close elections, exit polls could go wrong. This is true and was evident in Goa, Manipur and Punjab. However, agencies were also wrong in a sweep election like UP. That said, all of them going wrong at the same time, with Congress winning is possible, but it has a very low probability. A lot of agencies have propped up in last few years and one doesn’t really know about their ownership, research capabilities and background. Many new agencies do it for free for channels to get a foothold, which impacts the quality. Some do an online post poll survey.

The focus of this paper is on Gujarat elections. The average of poll of polls shows BJP getting 112 and Congress 68 seats. In terms of vote share BJP is expected to record 48% and Congress 41%. These results if true would mean a marginal improvement for Congress, however, on a very broad level, indicate status quo, as BJP won 115 and Congress 61 in 2012. The devil lies in the details though which again has been missed by many experts. The range for BJP seat tally is 99 (lower end of Axis-India Today) to 146 (higher end of Today’s Chanakya). Similarly, range for Congress seat tally is 36 (lower end of Today’s Chanakya) and 82 (higher end of Axis-India Today). Vote share ranges are 46%-52% for BJP and 35%-43% for Congress. These are very big ranges.

Three clear scenarios emerge. Though all of them show a BJP victory, they all depict a different story. What we are seeing is that the two of the most popular agencies with good track record have very divergent views on the results. Axis shot to fame getting Delhi, Bihar and UP correct. It does a seat by seat prediction. Chanakya got 2014 Lok Sabha, Delhi and UP correct. However, it goofed up in Bihar.

Scenario I: Axis lower end of range (BJP 99, INC 82)

If lower ends of what Axis predicts is correct, it shows Congress has managed to make a severe dent in BJP citadel of Gujarat. It places serious questions on the Gujarat model of development and will ring alarm bells for 2019. Modi can’t take 2019 victory for granted. It shows he is not invincible. BJP had won all the 26 Lok Sabha seats in 2014. Any losses here could mean it will struggle to retain its 282 seat tally of 2014.

Scenario II: Poll of Polls (BJP 112, INC 68)

This broadly means status quo is maintained in Gujarat. This result will provide something to cheer for both parties. BJP managed to hold on to its vote bank. 22 years of rule does bring in some element of anti-incumbency and it won despite many sections of society up in arms against it. People still don’t see Congress as a viable option. Congress should get solace in the fact that party managed to give BJP a run for their money. Most pundits had written off party’s prospects in Gujarat 3-4 months ago. It faced severe desertions having lost 28% of its MLAs who were poached by BJP.

Scenario III: Today’s Chanakya’s higher end of range (BJP 146, INC 37)

If higher end of what Chanakya predicts turns out to be true or even its median call of 135-47, then it is not good news for Congress. It shows BJP improving its position from 2012. Modi charisma still holding on. Congress unable to make dent despite creating a broader coalition. Its slow decision making and delay in forging alliance with the young troika – Alpesh, Jignesh and Hardik – and inability to project a leader versus Rupani as I put in an earlier post for Quint, despite getting rid of Shankersinh Vaghela, could be some of the factors.

If we consider forecasts for Scenario I and III as outliers, then we are headed towards status quo situation in Gujarat as per exit polls.

To conclude, the exit polls show three different possibilities and each has its own unique theme. We all need to wait till 18th to really know the actual results and then analyze the reasons. As my friend Subhash Chandra tweeted, it all depends on whether law of averages catches up with Axis, or the outlier forecaster Chankaya gets it right again. Huge stakes for both!

The article was first published here.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s