Five Forty Three

Revolutionizing Indian Election Analysis

Exit Poll analysis and the story of 30 swing seats of Delhi


Congress had 54753 more votes than BJP in the 30 swing seats of Delhi assembly. Of these 30 swing seats, Congress had won a whopping 21, BJP 8 and BSP 1 seat in the 2008 election. With the new additional voters in 2013, everything else being equal, BJP would have needed just 27 thousand more votes in these 30 assembly constituencies to win a maximum number of swing seats. It would have taken just an additional thousand votes in each of these swing seats to make an impact on this election. By and large, it is these 30 swing seats that will decide who wins the election, therefore, instead of focusing on unnecessary hyperbole, we will be concentrating our analysis on these 30 swing seats of Delhi.

The edifice of Congress in Delhi is built on these 30 swing seats, so it has to retain a majority of these at any cost. For instance, 49% (21) of the total Congress seats (43) are from these low-margin-big-fight swing seats, whereas BJP has only 34% (8) seats that fall into this vulnerable category. For the anti-incumbency facing Congress party, the big hope of 2013 is that AAP should manage to garner a chunk of the votes in these 30 swing seats and thereby helping maintain the lead of the Congress party. In many ways AAP’s performance in these swing seats will be the crucial deciding factor of who forms the government in Delhi.

Now consider this, about 61 lakh voters had exercised their franchise in 2008, whereas this time an additional 17 lakh voters voted in the Delhi assembly election taking the total votes polled to a whopping 78 lakh plus! Thus the entire math goes into a spin, for the Congress had a cushion of just about 55 thousand votes which is a paltry 3% of the additional votes polled on Wednesday. If that is mindboggling, then take this, 172000 voters of 37 assembly segments (more than half of the total 70) voted after the 5 PM deadline and nearly 10000 of those were yet to vote even at 6 PM. The irony is that many exit polls had started giving their final numbers by 5:30 PM, when roughly 4000 to 5000 votes were polled after the deadline in each of the 37 constituencies.

Our understanding of Delhi election tells us that a majority of the 17 lakh additional votes would be anti-Congress in nature mainly due to the anti-incumbent nature of the election this time; both against the state as well as the centre. For instance, there are 4 lakh 5 thousand first time voters, of whom 69% are believed to have exercised their franchise and we can easily surmise that a vast majority of those votes would be against the ruling party (somewhere in the range of 70-30 as per our exit poll survey). The problem though is a divided opposition and therefore the consequent division of the anti-incumbency vote. How exactly would the anti-incumbency vote be divided between BJP and AAP would be crucial to understand the Delhi election.

Although anything is possible in a wave election, it is less likely that much would be disturbed in the stronghold seats of both the established parties – BJP and Congress – because the number of votes needed to upset old social coalitions would be more than 8 to 20 thousand extra votes over and above base vote-share. So it is the swing seats that matter. In order to understand how the 30 swing seats are likely to vote this time, we have classified them into three different categories;

Category A: Zone where one of the party (BJP) is dominant – South Delhi and West Delhi – 8 swing seats

Category B: Zone where BJP and Congress are fighting each other directly – New Delhi and North West Delhi – 6 swing seats

Category C: Zone where all the three parties, BJP, Congress and AAP are in a tight race – East Delhi, North East Delhi and Chandni Chowk – 16 swing seats

This classification has been based on our opinion poll and exit poll survey and it is corroborated by two other opinion polls conducted by C-Voter and AC Nielsen (for our Delhi analysis we are discounting CSDS poll surveys as they are seen to have an inherent bias towards AAP).

Category ACategory BCategory CIn all three categories we see a uniform and almost proportional increase in voter turnout which is indicative of a single wave election of anti-incumbency with no localized factors affecting the turnout. Whatever minor discrepancies we see in Category C seats is due to multi cornered contests.

Among the 14 swing seats of Category A and Category B where either BJP is the dominant force or there is a contest mainly between BJP and Congress, BJP is leading in 9 seats, Congress in 3 seats and AAP in 2 seats as per our combined opinion poll and exit poll data. Whereas in 2008, Congress had won 11 and BJP 3 of these swing seats.

As per our exit poll data, of the 16 swing seats in Category C, BJP is ahead in 8, Congress in 3, AAP in 4 and ‘Others’ in 1 assembly segment. In the 2008 election, Congress had won 10 of these swing seats, whereas BJP had won only 5 seats (1 seat had gone to BSP).

If these opinion and exit poll trends hold true on the counting day, Then BJP would have gained 9 swing seats as compared to 2008, while Congress would have lost 15 swing seats as of last time. These findings are consistent with overall exit poll analysis as we shall now see.

Exit Poll Analysis

Small states are always a nightmare to project in any election, which is enhanced even further when we consider that Delhi is facing a very powerful third political pole which is accruing big vote-share. Then there is a wave like situation in Delhi due to very strong undercurrents of anti-incumbency. To complicate matters further there was a late, post 3 PM, voting surge in many parts which has upset many of the calculations – usually exit poll trends till 3 PM hold on for the end of the day, but a late reversal makes it even more difficult to project results. Finally, there is the unknown, unquantifiable commodity of AAP, which has no historic vote-share to have a proper comparative analysis. Under all these constrains, we at 5Forty3, with near zero resources, are trying to project an impossible election, so please do bear with us.

First let us see how the voting day unfolded just to give a background to the entire process of our post-poll analysis. After first two hours of polling, at about 10 AM, when the first set of numbers started coming in, it was shocking to see AAP do so well because the general assessment was that the fledgling new outfit lacks polling booth level infrastructure to bring in voters. The second shock was the extremely tight, 3-cornered fight in the New Delhi constituency, which we were specifically monitoring through exit poll numbers. Both of these shocks were corroborated by different party workers on the ground. Significantly, BJP was the single largest party and tantalizingly close to the halfway mark at this point.

Similar trends continued till about 3 PM, when we started to realize that Congress was losing Delhi in a big way and may end up being in the third position. It was after 3 PM that the trends started to reverse; a late surge of voting by Congress voters started to dent the AAP vote (borne out by exit poll time lines). Meanwhile, through all these upheavals BJP was always ahead of the pack and always hovering around the magic figure of 36. Thus we can surmise the voting day’s events as follows;

  1. Congress is unlikely to return back to power on its own
  2. BJP will be the single largest party
  3. The fight for the second position is strong between AAP and Congress, with Congress gaining late upper hand

Projected Vote-ShareIn all the six divisions BJP’s vote-share is between 32% and 36% (lowest being in East Delhi and highest being in North-West Delhi) and its total overall vote-share is 33%. Congress has a range between 21% and 30% in the six zones of Delhi – lowest being in Chandni Chowk and highest being in North-West Delhi and East Delhi. AAP has an even more volatile vote-share in the 6 divisions of Delhi ranging from 18% to 33% – lowest in West Delhi and highest in East Delhi. Due to these sub-regional discrepancies, Congress and AAP are both getting the same vote-share of 28%. BSP’s sub-regional vote-share is always in single digits (never crossing 8%) and is concentrated in 4 divisions – majorly South Delhi & West Delhi, and to some extent Chandni Chowk and East Delhi.

There is a lot of interchange of vote share among the parties, but the net swing is roughly 12% away from Congress, 8% away from the BSP, roughly 5% away from the ‘Others’ and 3% net swing away from the BJP. AAP is benefiting from this swing almost overwhelmingly.

It is quite possible that exit polls are overestimating AAP vote-share due to the first 6-7 hours of robust spread out voting, which has probably not given adequate weightage to the last 3-4 hours of concentrated voting. The last few hours of voting is generally believed to be party vote organized by the cadres. It is unlikely that AAP vote-share has been underestimated by exit polls. BJP vote-share has been consistent throughout the day and through all divisions so it is less likely to be wrong although minor corrections on the higher side due to “bandwagon effect” cannot be ruled out. As for Congress vote-shares, the anti-incumbency undercurrents may have led to lesser weightage given to traditional party voters.

Projected Seat-ShareEven if there were to be minor discrepancies, overall trends will likely remain the same when actual votes are counted on the 8th, except for some upheavals in the AAP-Congress vote-swings. Seat conversion based on these vote-share findings is an extremely hazardous job, but one thing is clear, BJP is likely to cross the halfway mark. We have given lesser seats to AAP despite same vote-share because it is more spread out, whereas Congress has traditional strongholds where its vote-share is concentrated.

A Note on New Delhi assembly segment:

We were specially tracking New Delhi assembly constituency as it was a part of our swing seat category and had one of the keenest VIP contests of recent times. Two different exit poll data points show Shiela Dixit as losing the New Delhi assembly seat to Arvind Kejriwal of AAP, the only difference is that in one of the polls BJP is at number two and in another Congress is at number two. Interestingly, in both the surveys, the difference between number 2 and number 3 is just about 2 percentage points.

[With valuable inputs and various exit poll data points from our psephologist and in-house numbers man, Chinmay Krovvidi]


Author: Dr Praveen Patil

A story teller and aspiring writer with special interests in Indian electoral politics

15 thoughts on “Exit Poll analysis and the story of 30 swing seats of Delhi

  1. The Akalis are likely to get 1-3 seats in areas with strong Sikh population. Has this factor been specifically studied?

  2. Awesome write up, we can understand state like Delhi with multi corner fight with a new party on the horizons …I think BJp will get more 39 in Delhi.

  3. Excellent Analysis of the last minute surge ! Expect lot more in the days to come !

  4. This was undoubtedly one of the most fascinating elections in the history of India and Delhi, in the sense of predictability and surprise. Strangely it points to a less fascinating election at the time of LS-14. The reason is that whatever is the early peek to LS-14 provided by this Assembly election will remove some element of surprise and also because I believe, most of the voters have already made up their minds about whom to vote. [I certainly have in favor of Modi, coz I want my vote to result in a working govt, not wasted in a senseless verbiage of secularism/communalism]

    LS-14 aside … we are currently on Delhi-13. Though it is merely two days away, I just can’t hold my curiosity about the result. There are a few factors which I feel should be paid attention to –

    1. There is a good sikh population. [let me be quiet about this point as it is already addressed in the comments]

    2. A huge bihari migrant population which was so attentively tuned to the ‘Hunkaar rally’ of Patna. This population has by now turned extremely partial towards Modi. I suspect a good part of the late surge was by these migrant workers and their non-Muslim votes would largely accrue to Modi. Since this was not Lok Sabha elections, some of it (~30% of non-Muslims) wud go to AAP. Congress is unlikely to garner more than 5% of this (non-Muslim) vote. The Muslim votes among this segment is more or less guaranteed to go to AAP.

    3. There is a simmering discontent among non-migrant muslims against Congress (both national and regional). This is a group which feels that AAP is, for better or worse, given them a viable alternative. I believe about 50% of these votes cud go to AAP. Given the good rapport that some existing BJP leaders maintain (and also because Modi will not be the CM of Delhi), about ~10% of these votes may fall to the BJP kitty.

    4. Delhi has a huge migrant student population. And also a large number of IT/ITES workers. Non-Muslims among these educated class have an overwhelming support for Modi. Significantly, many of them lie at the time of opinion/exit poll … merely for the heck of it. A quick visit to the college campuses and their facebook pages wud reveal yearning for a ‘strong’ Modi. Kejriwal gets only moral support from this group. Muslims herd towards AAP more or less completely. This group may also contribute to the late surge particularly as the reports of early exit poll numbers poured in and worried the BJP voters tremendously.

    The above is not based on any analysis. It is purely my anecdotal/personal observation. And I must add the caveat that my pro-BJP tint may color my opinions.

    All said and done, I feel the exit polls underestimate BJP’s tally by atleast five. Come what may, Harshvardhan is replacing Ms Dikshit.

  5. I am a BJP supporter and voted for Dr. Sahab but I have met around 20-25 people, more than 20 went to AAP. In my office BJP supporters have voted for AAP. Not sure of AAP has cut Congress votes too or not. Do you have any idea how much votes of both parties each have been cut by AAP?

    Fingers Crossed!!!

  6. “After first two hours of polling, at about 10 PM” don’t you mean 10AM? Very fascinating science ….what I fail to understand is how..just how can congress party still get 28% of votes?? Do we have that many ‘secular’ voters in Delhi?

    • 10 AM… you are right, that mistake has been corrected.

      How can Congress party get 28%? Good question. The answer is a bit complicated; in reality, raw data is only indicating about 21-22% vote-share for the Congress party. Since we always tend to underestimate Congress vote-share because its core voters (Dalits Muslims etc.) are under-represented by field data collection officers, I, as an analyst, took a call to give additional weightage to Congress vote share. Plus the late vote surge, which is believed to be organized by party cadres, also had to be taken into account. Then, we also had to add weightage to the fact that in the last 3 odd hours voting was brisker than in the first 7 hours, but a field officer collecting data would have possibly functioned at the same pace. All of these factors had to be mathematically accounted for, so the Congress vote-share was inflated (who says an election analysts job is easy??)

      The bottom line was very clear to us, BJP’s vote-share of 33% was mostly accurate and could only go a bit higher (for we have under-weighted it). But the AAP-Congress vote-share could vary and that is what has been clearly mentioned in our exit-poll analysis.

      As to why Congress still manages to get so many votes even after 60 years of misrule, well… will have to write a lengthy post someday…

      • Sir, In which scenario AAP can reach half way mark? How much percent vote increase would make it reach this point?

      • “As to why Congress still manages to get so many votes even after 60 years of misrule, well… will have to write a lengthy post someday”

        I think below can be reasons :-
        1. Minorities
        2. Dalit/Backward Class
        3. Gandhi Surname surely helped till now. We have to see after SM penetration it may not help much.
        4. Lots of Schemes which are on Gandhi surname surely help in rural India.
        5. Congress used to have good strategist earlier but seems they are losing the plot after NaMo’s rises.
        6. rest would like to hear from you.


  7. One of the best election analysis I have read anywhere in any newspaper, mind you I have read as many as ten articles in Hindi and English. I am an ardent Aam Admi Party supporter and Arvind Kejriwal fan so let me put forward my views here. Yes you are right everybody under estimated Aam Aadmi Party but we did fantastic boot management. The late surge of votes was also because of our volunteers who came late in the evening bringing many many voters. We got mostly congi votes than bjp votes so it will be Sheila who will suffer.
    Me and my group of friends had all volunteered for Aam Aadmi Party for the last 1 month and we know how hard we have worked, we really hope Arvind Kejriwal will become CM of delhi, we need good people like him, bjp in delhi is as corrupt and nepotic as congress.

    P.S all of us group of volunteers support Modi for parliament elections and we will work for him without any money or expectations, most aap volunteers have same idea for modi, but for delhi AK is the best.

  8. AAP will win in delhi, you see how the late vote will help AAP. Late vote was AAP vote. East or west Kejriwal is the best.

    Your analysis is extraordinary, better than on TV, why are you not on TV dr patil? But only mistake u are making is that u are not understanding delhi’s passion for AAP. No exit poll can predict a wave in favor of AAP, but still good analysis by you.

    Namo will be PM, Kejriwal will be CM! India will be the winner!

  9. excellent analysis .. but,nobody has commented on Akali dal seats & how it fared. If you can,please share it.

  10. Well researched analysis, Sir. Mostly logical as well as rational too. However, my apprehensions on two fronts: 1) Cong can’t get this many % of Vote Share in present situation. 2) BJP’s Vote share must be underestimated. Looking to the united front engaged through out by BJP, the Chief Minister & PM Candidate’s credentials, many new Voters must have attracted and Opted to vote for BJP rather than AAP Candidates barring few top guns. Finally, in India booth management and networking helps a lot at husting – a factor which could have played a major role for BJP as Cong workers were demoralized in last phase itself. Late Surge can’t be engineered by Congress in the given scenario. In all, BJP should be safely getting more than 40.

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s