Exit Polls 2019 – 5 States, 89 Seats Difference
Exit Polls 2019 – 5 States, 89 Seats Difference
Kartikeya Goel, Founder, FirstTimeVoter.Org
India waited for nearly six weeks – from 11th April 2019, when the first vote for parliamentary elections 2019 was cast, to see the which way the nation has swung. The exit polls of May 19th, 2019 have predicted huge success of Narendra Modi, BJP and NDA. The TV channels and media were splash with numbers in excess of 300 seats out of 542 for Narendra Modi led campaign for the new government.
Given the results of 2014, everyone knew that Uttar Pradesh was the battleground. BJP led NDA had won 73 out of 80 seats here in a four cornered contest between NDA, Samajwadi Party (SP), Bahujan Samaj Party (BSP) and Indian National Congress (INC). Formation of Gathbandhan between SP and BSP led to the expectation of consolidation of votes on caste lines and most experts including BJP strategists expected BJP to lose anywhere between 10 to 50 seats, depending on what arithmetical assumption they took. Therefore, it was believed that BJP extended its “Go East” strategy to Bengal and Orrisa, where it expected better pay-offs due to changing in the mood of the voters. In nutshell, this election had three battleground states – Uttar Pradesh, Bengal and Orrisa.
As I sat through the TV debates, swapping channels, I found few interesting facts which I leave for your consideration.
- With 272 as the magical number for majority in parliament, at one extreme Axis My India (Axis) and Chanakya estimated more than 350 seats for NDA; on the other end was ABP Nelisen with 267 seats and NewsX-Neta with 242 seats.
- Chanakya comes with the reputation of having predicted 2014 election numbers most accurate. In 2014, they had predicted 340 seats for NDA against the actuals of 336.
- Chanakya also got the Uttar Pradesh Assembly elections 2017 most accurate by predicting 285 (+/- 17) out of 403 seats for BJP, when most pollsters had predicted a hung assembly. BJP ultimately got 312.
- Newsx Neta is an app with 1 million downloads. It claims that about 4.5 lakh people polled on the app, which forms basis of its exit poll.
It is well known that all exit polls got 2004 and 2009 elections grossly wrong.
- In 2004 undertook “India Shining” campaign and all exit polls predicted between 200 and 250 seats for BJP, whereas it got measly 138 seats.
- In 2009, the exit polls predicted 185 to 205 seats for INC led United Progessive Alliance, but they trumped the pollsters with 262 seats.
Some interesting divergences in 2019 exit poll are as below:
- For Uttar Pradesh, the main battleground state – Chanakya and Axis estimate 65 to 68 seats for NDA, while NewsX Neta and ABP Neilsen estimate 33 seats each. The variation is as much as 35 out of 80 seats.
- For Bengal, Chanakya and Axis estimate 18 to 22 seats for BJP, while Newsx Neta and ABP Neilsen estimate 11 to 16 seats. That makes a variation of 2 to 11 out of 42 seats.
- For Orissa, Chanakya and Axis estimate 12 to 19 seats for BJP, while Network18 Ipsos and ABP Neilsen estimate 6 to 9 seats. That makes a variation of 6 to 10 out of 21 seats.
- For Bihar, Axis has estimated 40 out of 40 seats to NDA, Network18 Ipsos has estimated 36 seats for NDA while Jan ki Baat has estimated 29 seats and NewsX Neta has estimated 22 seats for NDA. That makes a variation of 11 to 18 seats out of 40.
- For Tamil Nadu, Network18 Ipsos and India TV CNX have predicted 12 to 16 seats for BJP-AIADMK, whereas Chanakya and Axis have estimated 4 to 6 seats. Again, a variation of 8 to 10 seats on a total of 39 seats. The only difference is Chanakya and Axis have swung on the other side here.
Summing up, it is quite visible that exit polls have significant variations. For the five states discussed above which make 222 out of 542 seats, while on one extreme NDA is expected to win 165 seats, on the other extreme, it could possibly end up with only 76 seats. That makes 89 seat difference for NDA.
Also, while all pollsters have claimed that exit polls have become more sophisticated and accurate, the inter-se difference is large and cannot be
ignored for all their tall claims. As Sanjay Kumar of CSDS has said, “we aregood at getting vote shares accurately, but the models from vote share to seat share will need refinement.”
So, hold your braces until 23rd May when the final results of 2019 Lok Sabha elections are out. Perhaps there is a deeper lesson for all.