Five Forty Three

Revolutionizing Indian Election Analysis

The first salvo of battleground 2014: Chhattisgarh Phase 2

Leave a comment

The last week of campaigning by star campaigners & top leaders and effective ground advertisement (not just by print/electronic media, but also by door-to-door pamphleteering etc.) is said to be the make-or-break feature of any Indian election, especially closely contested state assembly elections. Past fieldwork done by this author at polling booth levels has almost conclusively proven that close to 30% of the voters in state assembly elections and almost 20% of the voters in national elections are in the “undecided” or “fence-sitters” category till about 5 days before polling (will publish this work separately some other time). In fact, it wouldn’t be an exaggeration to say that on an average, approximately 5 to 10% of the fence-sitter voters decide on their voting choices either just a day before the polling or on the polling day itself. This is why the organizational strength of established political parties score over independent contestants and neo political entrants in converting popular support into votes and eventually seats.

Chhattisgarh election is a particularly close contest mainly fought along local issues no doubt, but the last mile campaigning could prove to be decisive as in most cases. As phase one had ended with a slight edge for Congress because of its 2008 washout in that region, phase two had to be the decider. Now let us try and analyze the 6 important factors of the last week of campaigning that could potentially leave an indelible mark on this election;

  1. Like all powerful mass leaders, Ajit Jogi is a consummate politician who doesn’t always play the game by the rule book. Thus his tacit understanding with Chattisgarh Swabhiman Manch (CSM), a third political force of the state, was a personal masterstroke in this election. An unlikely combination of Satnamis (Dalits) and Sahus would be a big winner in any election in central Chhattisgarh and Jogi had a winner on his hands. Although CSM had fielded some 54 candidates, they were serious contenders only in about 8 seats. The Jogi-CSM pact transcended even the Congress party’s high command, because it was rumoured that in the last week Jogi was tactically helping CSM candidates in a couple of seats – Urvashi Sahu (daughter of Tarachand Sahu) in Durg rural, former minister, Derhu Prasad, in Nawagarh and Ganguram Baghel (another former minister) from Arang are a few examples. As a return favor, he was getting the support of CSM in almost a dozen seats for his family and protégées.
  2. To counter this poaching of the numerically powerful Sahu votes, BJP had deployed Narendra Modi as not only their most popular national leader but also as a symbolic gesture to the Teli (Sahu) community to remain loyal to the party. Not surprisingly, almost 50% of the 8 odd, well attended Modi rallies were held in the Raipur-Durg belts which are the core areas of strength for CSM. One of the underlying themes of these Modi rallies was also about his “humble origins” which was a subtle way of attracting his fellow caste voters along with other backward castes
  3. Not for nothing is Ajit Jogi famous as the double-edged sword of Chhattisgarh, for he can be equally beneficial and detrimental for the Congress party’s chances in the state. In the last week of campaigning he once again proved his duality when he went haywire. What provoked Jogi is difficult to tell, but two factors could have worked; A] Narendra Modi’s constant mocking of Congress party not declaring the CM candidate and B] The total silence of Congress leaders when repeatedly asked as to who their leader is by the news media (Jogi expected his name to be spelt out by at least a few state leaders). So Jogi went ahead and acted in the only way he knows. He campaigned for votes in the name of “future CM’s family” for his wife in Kota and son in Marwahi, which he later reiterated in Gunderdehi. This created a great deal of consternation in the Congress camp leading to various factions working at cross-purposes in the last few days.
  4. Without sitting in moral judgment, let me discuss this interesting fourth factor. It is a well-known fact that almost 100% of the Muslim votes go against BJP in Muslim dominated polling booths and the party can do nothing about it… well, almost, at least till now. As an innovative strategy (maybe unethical) this time, hundreds of Muslim families were sent on a sponsored pilgrimage to Ajmer Sharif on the polling day by certain unknown (BJP leaning) local business entities in at least two divisions – Raipur and Bilaspur. Thus a vote that could have gone against the party was never cast
  5. In a surprising move, Rahul Gandhi and Sonia Gandhi (who had campaigned extensively in the much smaller phase one) were mostly missing in action in phase two. Instead Congress deployed other leaders, prominent among them was Bhupinder Singh Hooda, who camped in the state for two days and countered Modi. Although not as effective a public speaker as Modi, he did bravely counter the claims of development of Gujarat by pitting Haryana as a far more developed state. Hooda is emerging as the OBC counter of the Congress for the development plank of Modi and Chhattisgarh could be his first test
  6. Polling-booth level management by the trio of Saudhan Singh, Sachidanand Upasne and Rasik Parmar for BJP had been lackluster in phase one, but there seems to be a marked improvement in phase two. Although Vanwasi Kalyan Parishad was more active in the polling booths of Surguja in phase two rather than in Bastar of phase one, Dilip Singh Judeo’s calming effect is being missed sorely by the BJP which could have proven to be fatal for the party. But, for the first time, even the Judeo women also campaigned for the party who otherwise never move out of the palace. On the other hand Congress had almost totally outsourced its booth management strategy to a group of Charandas Mahant’s “non-political friends”, which is also an interesting strategy in a faction ridden state unit and has the potential to be a winning move by the Congress high command.

Opinion Poll indicators

Poll data

  • The two extreme ends of the spectrum are C-fore and CSDS, with C-fore predicting a dead heat and CSDS predicting a more than 2/3rd majority to the BJP
  • In terms of vote-share, no pre-poll surveys indicate Congress party crossing the 39% mark, whereas BJP vote-share varies between a low of 39% and a high of 46%, which gives us a trend-line that is advantageous to the BJP
  • Except for CSDS all other pre-poll surveys are projecting a neck-and-neck fight similar to 2008, with just about 1-2% margin separating both the parties – this can go eitherways in a close bipolar contest with a 3% error margin
  • Both CSDS and C-Voter project a better performance by BJP in central Chhattisgarh and by Congress in south Chhattisgarh (other surveys only give overall data) as compared to the previous election of 2008

Turnout Differential Factors

The overall turnout figure of 77% in phase two, which is up by about 5% as compared to 2008 in the same 72 seats is consistent with the turnout trends of recent years. What does it mean? Well, the old “high-turnout-equals-anti-incumbency” days are over for starters, but the new lazy analysis of high turnouts being pro-incumbency is also way off the mark. The devil lies in the details.

One thing is clear, in Chhattisgarh there is no anti-incumbency against the ruling party, but there is definitely micro-anti-incumbency against sitting legislators, which is why political parties try and drop unpopular MLAs and give tickets to new faces. But how does one pick up these signals of micro-anti-incumbency in such a crowded election? There is one system that we have devised and it is known as the “Turnout Differential Factor”.

As a first step, we pick the battleground constituencies based on two filters; A] low (less than 5000 vote) margins in 2008 elections and B] strong multi-polar contests in this election. So there are 13 battleground seats out of 49 in central Chhattisgarh. Now let us plot a turnout chart for these 13 seats.

Turnout Differential factor (CC)

Six of the 13 battleground seats were won by the BJP, while Congress had won 5, and 2 seats had gone to the BSP. As is visible from the above chart, the battleground seats of BJP have similar turnout patterns in 2013 as in 2008 indicating status-quo (except for Gunderdehi where a strong rebel CSM candidate is in fray). Whereas it is clearly visible that the turnouts in 2013 for Congress battleground seats are much higher and diverging from turnouts of 2008, indicating change.

Usually a party’s weakest link is the low margin battleground seats and in a small bipolar state this is the crucial difference between forming a government and being in opposition (as we have seen in Himachal Pradesh last year).

Now let us try and expand this to north Chhattisgarh to see if the same system holds true here too. In northern Chhattisgarh there are 23 seats, of which there are only 6 battleground seats since BJP and Congress have their own traditional strongholds in this region. Of these 6 battleground seats, BJP had won 2 and Congress 4 in 2008

Turnout Differential factor (NC)

Once again we see the same pattern; in fact, there is much lower turnout in Dharamjaigarh seat held by BJP and Ramanujganj is maintaining status quo, whereas the gap is wider in the Congress held battleground seats, once again indicating change.

Thus we can safely surmise that if turnout is a factor, then it is possibly favoring the ruling BJP more than the opposition Congress in phase two. Micro-anti-incumbency is not hurting the ruling BJP as much as it is impacting the opposition Congress. There are three important reasons why turnouts are helping the BJP rather than the Congress;

  1. Unprecedented women voter turnout suggests that Dr. Raman Singh government’s welfare schemes centered around women members of the family has had its positive impact on the fairer sex – there has been 6% more female turnout this time
  2. Despite promises galore by both the parties, voters seem to have greater faith on the ruling BJP because it has seemingly delivered on most of its past promises
  3. The general lower voter turnout in BJP held battleground seats is indicative of a relaxed voter attitude who do not want to punish the ruling party

Exit Poll Survey:

Methodology 1

This survey was limited to phase two only as it was extremely difficult to conduct an exit poll in the Naxal affected phase one. So effectively it was an exercise limited to 72 seats of 19 districts of central and northern Chhattisgarh only.

The aim was to cover at least one out of every three constituencies, picked randomly but by keeping the state’s composition intact – 15 General category seats, 6 ST reserved ones and 3 SC reserved seats made up the 24. About 3 to 5 random polling booths were chosen in each constituency by ensuring adequate geographic spread and urban-rural representation – 24% urban polling booths and 76% rural polling booths overall. 1647 random respondents were achieved by strictly adhering to the following criteria;

Methodology 2

Percentages have been rounded off for ease of usage. Urban and woman respondents were kept lower than the actual data by design, because urban turnouts are usually lesser than rural turnouts and men voters still tend to decide the voting preference of a family unit in greater numbers than women. About 3% of the respondents in semi-rural polling booths were difficult to qualify either as “rural” or as “urban”.

Lesser number of SC and ST respondents were achieved than the actual data – a lesson on why SC-ST voter turnout is usually underrated by psephologists – but this anomaly was addressed in the final percentage derivatives to the extent possible, adjusting for noise.

Exit poll predictions

[Only for the 72 seats of phase 2 and percentages have been rounded off]

BJP is gaining marginally (about 1%) and Congress is losing about 2% votes in phase two, while they had almost equal vote-shares in 2008 and were separated only by 0.40% in these 72 seats. Not surprisingly, the “others” are gaining most and the swing is away from the BSP (one caveat though, lesser SC votes were reported in the exit poll and despite “noise” adjustments, there could be errors).

This exit poll only goes on to confirm two aspects that many opinion polls have predicted;

  1. The contest is tight and only 1 or 2% separates both the parties
  2. BJP has a slight advantage over Congress

We have not attempted seat conversion based on exit poll findings because it was done only partially (18 seats of phase one missing). But this exit poll data can be used for trend confirmation. Having calculated a margin of error of about 4%, three following scenarios can be assumed;

  1. If there is an error of 4% and the margin moves towards the Congress, then Congress would have a lead of 2% over the BJP and the scenario predicted by the exit poll would be reversed
  2. If there is an error of 4% and the margin moves towards the BJP, then the gap would widen to a big 6%, which would ensure a sweep by Dr. Raman Singh of the proportions predicted by the CSDS pre-poll survey
  3. If there is an error of 4% and the margin moves towards “others”, it would be a status quo election vis-à-vis 2008.

[In case Congress sweeps the election, this exit poll survey would have erred beyond the assigned margin of error and the findings would be null and void]

A note on internal post-poll surveys:

After the elections, political parties usually have their own internal assessments which is a deeper level district wise analysis and throws much better light on what is likely to happen. BJP has claimed that it is winning 52 seats and Congress has claimed that it is winning 48 seats based on their own internal assessments.



Note: The Exit poll survey was sponsored and conducted by two independent organizations in collaboration. Questionnaire was kept simple by just putting all the names of the ballot paper of that corresponding polling booth along with party symbols and not asking for any other inputs from the respondents. Five Forty Three’s role was limited to analyzing the numbers only. Since this is a one of a kind exit poll survey done without preceding historic tracker survey data (unlike larger organizations like CSDS and C-Voter etc.), the usual error margin of 1-3% is calculated to be higher, at about 4%.


Author: Dr Praveen Patil

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

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