[This is a 2 part series on Uttar Pradesh elections 2017 based on Team 5Forty3 Survey conducted between June 28th and July 14th.]

In late May, BSP supremo, Mayawati had neatly divided the responsibilities of different castes among her lieutenants. Swami Prasad Maurya and R.S. Kushwaha where given the responsibility of bringing in Kushwaha, Maurya, Shakhya and Saini votes. Brahmin votes were the responsibility primarily of national general secretary, Satish Chandra Mishra, and secondarily of former minister, Ramvir Upadhyay along with 3 other leaders, O.P Tripathi, Ravinranath Tripathi and Subodh Parashar. Similarly, former minister, Thakur Jaiveer Singh and Jeetendra Singh Babloo were entrusted to bring in Kshatriya votes to the BSP fold. Finally, although Dalits were considered as the core vote-base of the party, R.K. Chaudhary was given the specific responsibility of attracting Pasi Dalit voters (the second biggest Dalit community after Jatavs).

After neatly dividing the UP demography among her “trusted” party men, Maya even addressed a group of Karyakartas in Lucknow and confidently told them that the party was going to come back to power in less than a year. Many Dalit ideologues and former Kanshi Ram loyalists wondered what was so special about this division of demographic labor. “In fact, weren’t these same ministers and MLAs working for the party in 2012 when SP defeated Maya or even more horrifyingly in 2014 when Modi pushed BSP to zero?” asked one former BSP leader.

Within 2 months an even bigger jolt awaited Maya. First Swami Prasad Maurya, then R.K Chaudhary and finally Ravindranath Tripathi all resigned from the party one after the other creating constantly negative headlines for the party. The actual rot though was far deeper than what the headlines told us. “The very same people whom Maya had appointed to guard the fields have deserted their posts” averred the same former BSP leader. Throughout June, there was a sense of nervousness among BSP workers and local strategists, these desertions had created a lot of anxiety within the party. Everybody was looking up to Maya to stem the rot and give a new direction to the party by bringing in some radical changes.

As if on cue, Mayawati called an emergency meeting of all the major leaders last week. She once again re-divided the state’s demography among her lieutenants. In the new scheme of things, R.S. Kushwaha was given the responsibility of all the Kushwaha, Maurya, Saini and Shakhya votes. Similarly, Brahmin votes were now divided among the remaining Brahmin leaders minus Ravindranath Tripathi and Pasi voters were simply included in the larger Dalit votes!

This is how Maya operates. She simply abhors any new ideas and totally believes in her old ways of winning elections. Lucknow folklore has it that after the humiliating defeat of 2014, apparently a group of young Dalit techies approached BSP leadership with an offer to create a separate centralized social media wing, but Maya dismissed them without even giving them a proper hearing. BSP is possibly the only large political party in India today which doesn’t have a youth wing. BSP’s presence in the digital world is also miniscule – no active Facebook interactions, virtually non-existent Twitter accounts and almost zero propagation of the party on the vast swaths of internet.

In the 2014 elections, BSP lost the largest chunk of Dalit votes in any Uttar Pradesh election of the new millennium which includes a total of 6 elections (both parliamentary and assembly elections). Our own polls of 2014 had shown that the biggest exodus of Dalit votes from BSP happened in the age group of 18-30. In fact, the 30+ Dalit vote had remained more-or-less intact for Maya, yet BSP won zero MPs in 2014 because the young deserted it.

Classically, with so much of anti-incumbency in the air, Mayawati should be the clearest alternative to Akhilesh Yadav in 2017 (as has been the case in the past). Yet, BSP is struggling to simply stay afloat. All the BSP rebels are planning a big political rally in the next few weeks and it is quite obvious that the rank and file are either deserting their post or are quite nervous.

Saffron caste micro engineering

In exact contrast is the political engagement of the other big contender to the anti-incumbency vote, the Bharatiya Janata Party. Prospective MLA candidates of the BJP have been asked to create FB/Twitter accounts and start engaging with voters online, digital campaigns are being designed specifically for different regions and most importantly, the party president is building a solid roadmap for creating sub-regional caste coalitions through massive doses of social engineering.

Most political observers believe that the recent cabinet expansion and reshuffle by the Prime Minister was a very carefully coordinated exercise aimed at the same social engineering. For instance, Anupriya Patel’s inclusion in the Modi cabinet may have caused heartburn within her family, but the fact is that she has by now emerged as the sole legatee of Sone Lal Patel’s unquestionable political capital among the Kurmi-Kunbi segment of the non-Yadav OBC pie. Anupriya Patel on her own is a force that can make a difference in some 37 assembly segments of Eastern Uttar Pradesh where the Kurmi-Kunbi vote is spread in a range of 12 to 30% and carries along segments of other OBCs too.

Ms. Patel is not the only ace that Amit Bhai Shah has up his sleeve, for he has actually taken the social-engineering project to a whole new level of finesse in the east. It is in this context that BJP’s alliance with Om Prakash Rajbhar’s SBSP must be seen. In both 2009 and 2014 Lok Sabha Elections, SBSP had fielded about half a dozen candidates to secure some 3 lakh votes. That vote-share may look miniscule in the larger picture, but it is significant when we consider the realities of Indian elections. The fact is that Indian voters do not want to waste their votes, especially in national elections, and if 3 lakh odd Rajbhar voters simply decided to waste their franchise that tells us a story of its own. There are many such smaller parties that operate in the eastern districts of UP which cater to a dozen such small OBC groups who are fighting a battle against the Yadav hegemony. Fundamentally, these small parties enjoy a place of pride within that caste/community for fighting a lone battle but also face the geographic limitation of not having any recognition outside a few districts of their influence.

Another such example is Prem Chand Bind whose Janwadi Party exclusively caters to the Loniya-Chauhan community that has an effective strength of less than 2% in the UP caste matrix, but in a couple of dozen eastern UP assembly segments their vote share goes as high as 18%. Mr. Bind is already working in tandem with Amit Bhai Shah and a larger alliance is expected to be announced in the next 2 months. Together, these small OBC groups made of Kurmi-Kunbis, Loniya-Chauhans, Rajbhars, Kahars, Kashyaps and the biggest of them all, Mauryas, comprise of a whopping 30% vote base in east UP. When added to the BJP’s core of upper-castes and Lodh-Rajput voters, it becomes a lethal combo of nearly 1 out of every 2 votes in that geography.


  • A cursory look at the caste-vote matrix of Uttar Pradesh as it stands today tells us that BJP’s efforts have managed to retain at least some of the social coalitions it had achieved in 2014. AS compared to 2014, the biggest swings away from the BJP are happening in the Dalit, Jat and the Brahmin segments of the populace (in that order). The Dalit vote is partially returning back to Maya, whereas the Jat vote seems to be scattered as of now. What we must understand here is that all of these are relative terms. For instance, even today, BJP is the number 1 choice of Jat voters at 39%, but it is much less as compared to 2014 which stood at a whopping 70%+.
  • Despite massive doses of anti-incumbency, at least as of today, the Samajwadi Party seems to still enjoy enormous support among Yadavs which is along expected lines. What should really worry the SP is the state of the Muslim vote. An unusually big 21% of the Muslim vote in the state is still in the undecided category which means that every 5th Muslim voter is willing to experiment beyond the secure confines of Azam Khan-Mulayam duo. The undeniable fact of UP’s minority political landscape is that the mid-level and lower level Muslim leadership in the districts is largely part of the SP ecosystem, so if the Muslim vote has to go to other parties in substantial numbers, then the average Muslim voter must turn rebel against the traditional local leadership. It is in this contest that we must keep a close watch on the Congress strategists and a host of Muslim parties coming together under the leadership of AIMIM.
  • BSP has still not fully recovered its Dalit base as only 1 out of 2 non Jatavs are backing Maya today which is substantially lower from her 2007 peak. Even among Brahmins, BSP is getting around 18% vote-share as of today which is just about the bare minimum that the party has always got. Only when 1 out of 4 Brahmins start to go with Maya will the BSP attain the lethality of 2007. The big potential area here for the BSP is the Muslim vote, but the question is how will Maya woo them?
  • The Congress party is still as irrelevant as it had been in 2014 so it gets clubbed together in the “others” category. In our questionnaire, we had specially devised questions to ask the voter if they are likely to change their voting preferences in case Priyanka Gandhi or Sheila Dixit becomes the CM candidate. Less than 2.5% of our respondents showed any inclination to change their preferences. This is the humungous challenge that PK faces, the sheer disinterest among average Uttar Pradesh voters towards the Congress party.
  • Independent of caste silos, it is the youth vote that has emerged as the new segment which plays the biggest determinant in elections these days. BJP saw a gravity defying 292% jump in its youth vote-share in a matter of 2 years from 2012 to 2014 which enabled the Modi wave to eventually accrue 44% vote-share and 73 MPs in the state. Although relatively lower, even today nearly 4 out of every 10 youth of UP are positively disposed towards the BJP which is giving a big boost to the party’s chances. It is that 292% jump in just 2 years which calls for a deeper understanding of the youth vote which has two parameters; A] The fact that youngsters (18-30 age group) create the maximum noise both in the digital as well as the offline world creates a massive surge for political parties – this is the missing link for BSP and B] Today, the young increasingly find themselves in the decision making roles in a family structure which makes their vote far more valuable than it was a generation ago.

Invisible Collectivism hypothesis and the RSSI scale

The biggest factor that works in Indian elections is “invisible collectivism” which gets zero importance in most psephological models. This is a unique Indian feature wherein people find safety in numbers which is also reflected in our consumption patterns across all categories. From cars to paints to toothpastes, India is a rare market where the top 1 or 2 players get the lion’s share of over 50% market-share. Average Indians usually want to buy products that are brought by others too because we find safety in numbers. Similarly, average Indian voters want to vote for a political party that others are voting for – this is why we invariably witness the “winner takes it all” phenomenon in Indian elections. If a lot of Brahmins and Thakurs and OBCs are voting for a political party in a particular geography, then a lot of Dalits and even Muslims are also likely to shift from their traditional choice and go with the collectivism of the others. This is how “waves” unfold in Indian elections.

For instance, in 2012, when Yadavs and Muslims and Thakurs started shifting towards SP, a section of Brahmins and Dalits too joined the bandwagon to take Akhilesh Yadav towards a clear majority. It often works like an ‘invisible hand’ that guides vast sections of the populace into voting towards a leading party. We saw a much bigger example of the same in 2014 when no political pundit expected a clear majority for Modi in the national election but the invisible collectivism of average voters ensured that BJP won 282 MP seats on its own.

A subsection of this “invisible collectivism” theory of our psephological model is “relative enthusiasm”. Over the years, we at 5Forty3 have always argued about the statistical significance of relative turnouts of different caste/ethnic groups which is the greatest determinant of electoral outcomes in India. For instance, in the Bihar elections of last year, the upper caste and Dalit turnout was relatively lower to the consolidated turnout of Yadav-Muslim-Kurmi vote. In fact, this is the reason why our indigenously developed tool RSSI (Randomized Social Swing Impact) gave 100% accurate results in Bihar while all other pollsters failed.

Over the years we have developed a unique RSSI scale which measures the relative enthusiasm of different castes/subgroups of voters on a scale from 0 to 10 by assigning different weightages to different responses in the survey of specific swing polling booths. For instance, we assign higher weightage to greater consolidation of votes behind a particular party in a particular swing polling booth. Thus the resultant curve will then show us how likely a caste group is willing to turnout on election day which in turn determines how effective the social coalition of a dependent political party is likely to be.


Thakurs, Yadavs and Jatavs are the three communities that have shown maximum consolidation swing on RSSI scale whereas Muslims seem to be the most disinterested and the most confused of the voting bloc. Surprisingly, even the non-Jatav Dalits and Brahmins are just at the median score of 6 which indicates a certain degree of lack of enthusiasm among these groups (it must be remembered here that the semi-urban and urban Brahmin voters in UP have historically been indifferent to elections and their turnouts have always been low, but 2014 had changed that to a large extent). The crucial segment really is the non-Yadav OBC groups which will decide who wins Uttar Pradesh.

It is here that Amit Bhai Shah’s strategy of tying up with small OBC parties of the likes of Rajbhars and Loniyas starts to make a crucial difference. These small parties like SBSP may not have a large vote-base, but they end up enthusing smaller OBC groups into turning out on voting day which helps BJP consolidate its social coalition against Yadav-Muslim hegemony.

The 4 pillars of BJP strategy in UP

The saffron strategy in UP is (unknowingly) based on channelizing the invisible collectivism by creating a hegemony on the winnability criteria. This strategy should be ideally built upon 4 pillars. We have already discussed the RSS as the first pillar and caste-coalition as the second pillar of this strategic base. Now let us consider the other two pillars.

Among the 5 key issues that the respondents picked, about 32% of them believe Law & Order situation in the state along with communal politics was their primary concern. These were voluntary choices that came up from the respondents. When specifically asked if the state government deliberately practiced communal politics of appeasement, more than 1 out of every 2 Hindu voters (55%) believed so, whereas only 32% non-Hindu voters gave a positive answer. This divergence gives us a very interesting insight into the psyche of average Uttar Pradesh voter. While most Hindus either clearly believed or were non-committal about the state government’s deliberate Muslim vote-bank appeasement, non-Hindus strongly believed the opposite.

In fact, in this case, the non-response of more than a fourth of Hindu segment of voters underscores a state of confusion due to either lack of good alternatives or a certain degree of ambiguity on the part of the BJP. Interestingly, more than 36% of the Hindu voters believed that many Hindu families are driven out of Muslim majority ghettos like that of Khairana (every 3rd Hindu voter). We have also demonstrated yesterday how BJP’s vote-share even goes beyond its peak in those assembly segments where Muslim population is higher.

All of these are clear pointers to the fact that a polarized Hindutva agenda brings out greater unity among the pro-BJP voters which essentially defangs the minority vote that tries to tactically defeat the BJP in every seat. This is a very important tactical strategy that should be deployed with subtlety, for a large segment of Hindus are also less enthusiastic about overt Hindutva and more interested in the day-to-day economics. Thus, like 2014, the Hindutva agenda should be more an undercurrent built by the RSS ecosystem while BJP’s campaign should be built purely on a development agenda.

In fact, an economic agenda is the 4th pillar on which the BJP campaign rests in Uttar Pradesh. We live in an era when voters are no longer enamored by ambiguity, they expect a clear agenda driven campaign from a political party. BJP needs to identify at least 3-4 major issues that really matter to the voters and weave a specific campaign around those themes of what are the solutions that the party is offering to the voters.

“We have got a bank account, but really don’t know what to do with it”, this has been a constant query from a large number of respondents across the state when asked about the PMJDY scheme. In Uttar Pradesh, a whopping 2.5 Cr PMJDY accounts have maintained a positive balance till now. As per estimates, roughly 4-5 Crore rural households now have access to a positive balance bank account, but a vast majority of them are still unaware of how it is going to benefit them. The Prime Minister must now clearly articulate how he plans to transfer all the subsidies directly into the people’s accounts (already more than 1 Cr accounts in the state have been Adhar seeded). In fact, going forward, the prospective BJP CM candidate should set a clear timeframe of less than 6 months for subsidy transfers into PMJDY accounts. This is a game-changer idea that needs far better articulation on the part of the BJP because it has a very specific solution to a very specific problem.

Fact is that there has been no communicator better than Narendra Modi who has articulated complex economic issues to the voters in a simplistic way that they can relate to. The average voter in UP is quite intelligent and politically savvy enough to understand the complexities of governance. For instance, the PM has to explain various agrarian and rural policies of the government to the average voter in terms of how they are likely to yield long term benefits; as also the digitization of all the APMCs and how that measure could end up revolutionizing Indian agriculture. Another case in the point is the Pradhan Mantri Ujjwala Yojana which has already given over 4 lakh LPG connections in Uttar Pradesh and aims to provide 5 Crore connections across India. This is a great initiative that needs to be constantly hammered into the voters’ minds creatively by the BJP.

The Headline numbers of UP


As things stand today, even without a CM candidate and despite a 12% away swing from the 2014 peak, the saffron party has a solid 7% lead over its nearest rival, the BSP. Yet we must not forget that 13% or roughly 1 out of every 8 voters in the state are still undecided about their vote. Even more importantly, a large chunk of those undecided voters belong to the Muslim community. For BJP to maintain its pole position, it has to strengthen the 4 pillars on which its victory in UP will rest upon.

Sub-regionally, BJP enjoys a clear lead only in eastern Uttar Pradesh consisting of a whopping 29 districts and the small region of Bundelkhand, whereas in all other areas it is facing stiff competition. Especially interesting is the massive away swing of over 18% from western UP (including Rohilkhand) as compared to 2014. This is the real challenge for BJP which has not nurtured any serious leadership in that prosperous Jat dominated region of UP. For instance, somebody like Dr. Sanjeev Baliyan (the recently reshuffled water resources and Ganga rejuvenation minister) could have emerged as the young suave face of Jat aspirations because he had all the credentials for it, but unfortunately his political graph never took off after 2014. This is a big failure of BJP’s leadership development programs. Even now BJP can build a solid Jat leadership in west UP by creating special advisory and public relations teams to nurture the Satyapal Singh-Sanjeev Baliyan duo; especially of importance is to improve their social media interactions and digital presence to address the restlessness among Jat youth. If the saffron listlessness in western UP continues then it would once again create a political vacuum for Chaudhary Ajit Singh to make a comeback.

For Mayawati, the really potential demographic pocket is that of Muslim votes. If BSP can demonstrate unequivocally that it is holding on to its core Dalit votes and getting strong support from Brahmins and a segment of OBCs, then Muslims may join the Maya bandwagon, but otherwise the Muslim vote is really weary of once again wasting itself on a party which got zero seats in 2014 and eventually proved to be counterproductive in stopping Modi. Maya’s problem is threefold;

  • BSP has not yet made any meaningful inroads into the non-Yadav OBC segment, especially the Jat segment in west UP which has severely limited her traction
  • Core Dalit vote, especially among the youth is not as solidly behind Maya as it was a few years ago and the Dalit czarina has not yet understood the importance of courting the young
  • While Congress has emerged as another challenger on Brahmin landscape, the emergence of AIMIM is also taking away sections of Muslim votes. Both these factors are hurting Maya the most, at least at present

There are many observers who believe that eventually Maya will be left with no option but to tie-up with the Congress party in order to plug the above gaps. In fact, it is widely rumored that Nitish Kumar and his coterie is actively courting Maya to join the PK strategic universe of Congress game-plan. It would be interesting to see how BSP reacts over the next few months if it is unable to gain traction to beat BJP.

In fact, we tried to indirectly assess voters’ minds by recording responses to a specific set of questions as part of the addendum at the end of the survey. This was more of an exercise to decipher why the voters are still supporting BJP and how strong is that bond. Two interesting questions we asked the respondents were regarding the Mohammad Akhlaq episode and the Kanhaiya Kumar fracas at the JNU campus.


Although only 39% of our respondents had heard of Kanhaiya Kumar’s JNU fracas, it is still quite a large number considering that it was merely a campus issue in the national capital (goes on to show how information flow is far more robust today). What might come as a big surprise is that 73% of those respondents actually believed that the Kanhaiya Kumar kind of anti-nationalism should not be tolerated at any cost which goes on to show that the saturating media coverage of JNU actually brought renewed support to Modi and BJP. In fact, the biggest chunk of those respondents were youth who are the primary narrative disruptors in today’s society which only helps magnify the media hounding of a Modi led BJP. Even on the cow slaughter issue, a whopping 67% of respondents (cutting across Hindu caste lines) believed that BJP was right about the cow slaughter issue. Whatever the intelligentsia may preach on these issue from the confines of their TV studios, average voters are concerned about growing anti-nationalism among the elite and the deliberate disrespect to Hindu religious symbols like the cow.

A brief note on the Uttar Pradesh poll survey

This UP poll survey was conducted between June 28th and July 14th spread across 145 polling station areas covering 51 specially chosen assembly segments of all the different electoral zones of the state. Our survey covered 196 geographic locations of 23 districts and had a target sample size of 4680, giving adequately weighted representation to various castes and socio-economic groups of both the rural and urban populace (see the below tables).

Table 01

Table 02

This 5Forty3 survey has a statistical error margin of 3.1% and a historic error margin of 2.7% (based on our past record). Our polls are especially accurate because of our path-breaking swing polling booth surveys based on indigenously developed tool RSSI (Randomized Social Swing Impact) which gave us near 100% accurate results during the Bihar elections where every other pollster failed. Based on our long experience we can say that there are three main error zones for political poll surveys in India;

  • Selection bias: Our methodology has possibly a near perfect selection probability, therefore this poll survey has produced one of the most robust results.
  • Random Sampling Error: Since we have used a controlled random sampling methodology this problem has been almost completely negated; for instance, most pollsters simply use a randomized sample across the geography to arrive at findings, but our system has extensively used electoral rolls in conjecture with our own ethnic (caste) database as the sampling frame.
  • Non-Response evaluation error: Most poll surveys completely ignore “no responses” and simply concentrate on the positive responses, which artificially inflates findings. Our system actually treats “no responses” as positive findings and assigns value to them, which adds a whole new dimension to understanding people’s choices.

Clear and precise questionnaires containing 65 different questions based on 3 criteria – preferences, opinions and factual information – were prepared in Hindi/English/Urdu which were used to elicit the opinion of the respondents by carefully trained fieldworkers. The entire survey was conducted by direct face-to-face interviews without employing any CATI surveys or phone surveys. The cost of our first survey escalated due to a larger sample size necessitated by the dynamics of a large state like Uttar Pradesh – while originally we had planned a survey capped with a sample-size of 3600, the same was redesigned to expand up to 4680. We achieved an extremely frugal hit-rate of 111 rupees per response which is a new record even for the historically cost effective team 5Forty3 (industry standard rates in India are roughly 200 to 270 rupees per response).

P.S: Seat Share Projections

In the complex electoral arithmetic of India there are no fool-proof statistical models to derive seat-shares based on vote-shares. This further becomes an exercise in futility when being done almost 6 months before the elections when so many voters are as yet either undecided about their choices or are not fully firm. Yet, for vast majority of our readers, an electoral analysis without seat-share projections becomes a half-baked product, so we are reluctantly including this section as an epilogue and do not take guarantees for these numbers. In fact, good psephology ends at the point of vote-share analysis and all activities after that are merely speculative exercises.