The overall vote-share lead of NDA remained at around 5% in both the heartland states of Uttar Pradesh and Bihar in phase 2. The key difference was that in UP, NDA virtually means BJP on its own whereas in Bihar it is a strong three-party alliance of BJP-JDU-LJP. It is therefore pertinent to point out here that north India literally looks like a mirror image replica of 2014 after the two phases of electioneering, whatever be the state of opposition alliances.

In the state of Uttar Pradesh, once again our models are showing 6 clear leads for BJP and two narrower leads which have been classified as tossups. In essence, till now UP predictive models show 12 MP seats going to the BJP and 4 being in toss-up category. Even in the 4 tossup seats, 2 of them are showing a stronger tilt towards the saffron party rather than the opposition Mahagathbandhan. Once again there are four primary reasons for this kind of a political momentum in Uttar Pradesh.

  • The upper-castes + Non-Yadav OBCs + Non-Jatav SCs together form a solid overall vote-base of nearly 50% in these two supposedly minority dominated phases of Uttar Pradesh elections of 2019. All these three segments have stood solidly behind the BJP and Modi’s leadership in particular, despite of some pre-poll anxieties about the state BJP government being projected as a pro-forward-caste entity by many sections of media and intelligentsia.
  • There is a significant drop of anywhere between 4-6% in the Muslim voter turnout even as the overall voter turnout has remained more-or-less the same as that of 2014. We have already demonstrated this with clear raw data after phase 1 and this trend has continued in the second phase.
  • Vote transfer between the main constituents of the opposition alliance has not been smooth at all. One the one hand, Yadavs have literally seen a three-way split and only a segment of this community is reluctantly voting for the MGB, while on the other hand, Jats who were supposed to shift away from BJP due to the presence of RLD in the MGB have simply not made that shift – the situation is such in the Jat-land that even Chaudhary Ajit Singh is trailing. In reality, the MGB remains a virtually two-segment alliance as of today – a Jatav-Muslim combo at best.
  • The fourth and perhaps the least understood phenomenon is the way the poorer segments of UP society have voted based on the benefits they have received from the Modi government going beyond caste considerations – LPG, DBT, Toilets being the top 3 hits among the voters. Once again, we have proven this conclusively by tracking hundreds of families in these two phases where Modi remains the numero-uno choice for more than 2/3rd of these voters.


Among the 8 seats that went to polls, only 3 seats actually saw any contest while the other 5 – Agra, Aligarh, Bulandshahr, Fatehpur Sikri, Hathras and Mathura – hardly witnessed a fight as BJP was clearly ahead. In fact, in seats like Hathras and Mathura, especially the latter where yesteryear film star Hema Malini is defending her seat, BJP may actually maintain its 2014-17 lead of over 20% margin.

In Nagina which was one of the important seats that MGB wanted to make inroads this time, the ground realities were quite different from our pre-poll expectations. A small but significant segment of Jatav votes were seen to be going to Omvati Jatav of the Congress party (him being a former BSP Legislator) even as the Samajwadi cadre was simply absent on the day of polling, many refusing to work for the BSP candidate who had stood third in 2014.

In the high profile Amroha seat, the MGB was much better coordinated and the Congress too was virtually absent, thereby not splitting the opposition vote. This seat by far remains the sole hope for the Mahagatbandhan in this phase as Danish Ali (the former Devegowda acolyte of north India) seems to have better managed his resources as well as his political networking. Our turnout models indicate a close fight in this seat.

[Moradabad was wrongly mentioned in this article as having gone to polls, It had been picked up from our pre-poll tracker rather than post-poll. The error is regretted].


In the other heartland state of Bihar, this was a phase where the heavily Muslim dominated Seemanchal went for polls, leaving the NDA allies to do most of the heavy lifting. In Kihsnaganj, for instance, a seat won by Congress even in the face of 2014 Modi wave, owing to the fact that 70% voters here are Muslims, this time struggled to retain its base. AIMIM led by Owaisi gathered a significant section of Soorajpuri Muslims who had voted for Congress last time while JDU managed to get the other Muslim votes to its side. It is one of the tossup seats in our models, but NDA may well manage to win this seat.

Katihar, the other opposition stronghold where Muslims and Yadavs together account for around 52% votes saw a triangular contest again this time as NCP has fielded a strong Muslim candidate and a segment of Yadavs are a bit averse to vote for Tariq Anwar of the Congress party due to decades long animosity. Our models show that OBCs have voted virtually in the range of 85% to 15% in favor of the JDU, but the seat has still been classified as a toss-up.

Among the 3 remaining seats, Purnia and Bhagalpur have shown decent leads to NDA, while Banka which saw a triangular contest with Putul Kumari, the wife of late Digvijay Singh being in fray as a BJP rebel, has given a slender edge to JDU. BJP has remained virtually on the sidelines visually throughout the first two phases of Bihar elections which has served the strategic purpose of Chief Minister Nitish Kumar of not polarizing the elections along religious lines and managing to attract a segment of Muslim votes. The real election for the saffron party will only begin from now on.

The Eastern Saffron Sunrise Continues Unabated

First, let me state the obvious. BJP has shown significant leads in all the 3 MP seats of West Bengal that went to polls on Thursday last. This is an unprecedented development in Bengal’s politics and will have many long-term repercussions irrespective of how the other parts of Bengal vote in subsequent phases. 

To begin with, let us take the case of the tea belt seats of Darjeeling and Jalpaiguri. In both these seats, BJP seems to have built a solid base from which the electioneering has taken off. “The state government is completely apathetic to our concerns, we struggle to get even a minimum wage of 160-170 rupees whereas in neighboring Assam ruled by the BJP, everybody gets a minimum wage of 350 rupees” many tea workers told our teams in Jalpaiguri. These two constituencies together account for over 3 lakh voters belonging to tea workers community and their common refrain against TMC had literally become a tidal wave on polling day.

The second most important issue here is the NRC – national registry for citizenship – which has really stirred the pot as Hindu voters are largely united in favor of the BJP’s move. Contrary to popular perception, TMC barely managed to win the Jalpaiguri seat for the first time in 2014, and since then, a large section of ground cadre and workers have made a beeline from both TMC and the Left. Darjeeling in any case has been voting for the saffron party for more than a decade now and will continue to do so this time too.

In Raiganj, a four-cornered contest has helped the BJP snatch it away from the Left which had won it from the Congress in 2014. This is in fact a common feature of BJP’s growth in many parts of India as it keeps occupying the vacuum left behind by the Congress party in its erstwhile strongholds – we saw  in the late Priyaranjan Dasmunshi’s pocket borough on Thursday and are observing similar trends in Malda district. Two factors that are working in BJP’s favor here are – 1] Congress party’s Islampur MLA, Kanhaiyalal Agarwal shifting to TMC and fighting this election has hurt the Congress party a lot and 2] BJP stood second in the Panchayat polls of the district last year which has given it strong rural presence. The perception in the constituency was that it was a direct fight between the Left and the BJP wherein the incumbent Left MP, Mohammad Salim has a big handicap of being totally absent for the last 5 years because of his preoccupations with Delhi and national politics.

In the neighboring Barack valley of Assam, once again NRC was the biggest factor in the election which has deeply polarized the electorate along Hindu-Muslim lines. One can understand the impact of the same by looking at how Congress party leaders and workers are unable to explain their stand to the public. For instance in Silchar, sitting Congress MP, Sushmita Dev, sensing the anger of voters had to repeatedly promise to support Modi’s CAB (Citizenship Amendment Bill) going against the official stand of the Congress party.

What has also helped BJP is the fact that it had replaced 5 of its 7 sitting MPs thereby neutralizing localized anti-incumbency to a large extent. On the other hand, Congress could not even replace its oldest MP, Biren Singh Engti from the Autonomous District of Assam despite of the fact that he had very little chance of winning. As a result this is now a seat that has given BJP big leads on election day.

Of the 5 seats that went to polls on Thursday, BJP is definitely ahead in 4 and there is a lone tossup seat. In fact, 2019 will be remembered as a turning point election when virtually the whole of North-East had been saffronized. Of course, one man who would get a lion’s share of the credit for BJP’s emergence in the seven sisters as the most dominant political force will be Hemanta Biswa Sarma who left the Congress party and has caused incalculable loss to the grand old party – had Rahul Gandhi not been interested in his dog, pidi, and the dog biscuits on that fateful afternoon, Congress party would have had a different trajectory in the North East.

In Odisha too, the onward march of the BJP continued in phase 2. Of the 5 seats that went to polls on Thursday, our models have projected saffron leads in three seats, a gain of 2 seats as compared to last time. Of the remaining two seats, 1 is going to the BJD and 1 is in tossup territory. Once we have full data from all of Odisha, we will be able to do a much deeper analysis of this momentous 2019 election in the state.

Maharashtra trends in the second phase

The NDA’s good showing in Vidharbha, despite a section of media reporting about anti-incumbency continued in the second phase too, albeit with a glitch or two. As per MAPi – Micro Analytics Projections (INTELLIGENCE) – trackers, of the three seats that went to polls from Vidharbha in this phase, NDA is ahead in 2 and the NCP in one – NDA had won all 3, last time around. The lone seat where our models are showing Shiv Sena trailing is in Buldhana, where possibly the localized anti-incumbency is at play. In the highly publicized Amravati contest wherein a local actress is challenging the sitting Shiv Sena MP, Anandrao Adsul, the NDA was way ahead on polling day. Even in Akola, one of the two seats where Prakash Ambedkar is contesting as the VBA nominee, BJP had a reasonable lead.

In the Marathwada region, voting pattern this time virtually reflects the 2014 trend. For instance, in Latur, now infamous for its water woes, despite all the campaign by the Deshmukhs’ (late former CM Vilasrao Deshmukh’s family) efforts, BJP seems to be retaining the seat effortlessly. In Beed, where Pritam Munde contested again this time, BJP has taken decisive leads on our trackers. The fact that there was virtually an exodus of NCP leaders from the district has helped Ms. Munde’s campaign a lot, even as BJP had 5 out of 6 local MLA’s in its kitty to fall back upon.

In Parbani and Osmanabad, the two Shiv Sena MPs faced anti-incumbency, but may scrape through due to the popularity of Prime Minister Narendra Modi. Indeed, Modi still enjoys huge popularity across Marathwada and there is a clear consensus among most voters that the PM needs to be given another chance.

In the erstwhile Congress bastion of Solapur, BJP’s Mahaswami Jaisiddheswara Shivacharya has managed to get nearly 90% of the 2.8 lakh Lingayats who voted on Thursday, even as Prakash Ambedkar of VBA has got a big chunk of SC and Muslim voters (his candidature is also backed by AIMIM). Maratha voters are crucial here and our models suggest a 40:40 split between Congress Party’s Sushilkumar Shinde and BJP. Due to various x factors, MAPi has put this seat in the tossup category with a slight tilt towards BJP. Net-net, 2019 could end up exactly same as 2014 in this phase for Maharashtra with just a couple of seats exchanging hands – Buldhana and Hingoli.

Tamil Nadu: The candle in the wind?

It requires a psephologist extraordinary courage not to declare a landslide in Tamil Nadu. 94% of the times, Tamil Nadu votes for a landslide in favor of one party/alliance or the other, especially in the Lok Sabha elections. The latest example was 2014 when AIADMK won 37 out of 39 MP seats in the state and became the third largest party in the Lok Sabha just 7 short of the Congress total!

After carefully analyzing the different voting patterns of all the 5 regions of Tamil Nadu, we are tending to believe that Tamil Nadu has possibly defied its longstanding tradition of voting for a landslide, although we do not rule out that possibility. We believe there are 6 major reasons why TN may defy its historical tendency this time around.

  1. For the first time in many decades, the two powerful icons of Dravidian movement – J. Jayalalita the legatee of MGR school of thought of charisma, oratory & populism and K. Karunanidhi representing the ideological purity of Dravidian movement – are both absent from the electoral scene leaving vast gaps in the political landscape of the state.
  2. TN politics has possibly never been so fragmented with so many claimants to leadership position (imagine Rajnikant also entering the fray a couple of years down the line) and yet there being a big vacuum of leadership leading to a possible fractured mandate.
  3. Newer entrants with nebulous political base creating micro-trends and micro-countercurrents (that are even difficult to capture by vast polling exercises) leading to unprecedented fragmentation of voting – Kamal Hassan’s party and TTV’s AMMK being the prime example with hard to determine vote-share of anywhere between 3% to 20% in different sub-regions.
  4. For the last 3-4 election cycles (both national as well as local), the Amma machinery of distribution of vast sums of money during elections enjoyed a clear edge over its main rivals. This time, our on the ground observation is that both the groupings have been equally efficient.
  5. MAPi – Micro Analytics Projections (INTELLIGENCE) – has shown 12 tossup constituencies which means nearly 1 out of every 3 MP seat is virtually “unpredictable” at the micro-trend level (beyond overall vote-share) which is by far the highest uncertainty that we have observed this election season in any geography.
  6. Historically, on an average, AIADMK enjoyed 56% to 39% (DMK+) edge among women voters across the state (remaining going to others). There hasn’t been a dramatic fall in that support base which is again not giving us enough confidence to project an opposition sweep.


Keeping all these factors in mind our models are projecting a slight 1% edge to the DMK alliance. In TN, even a 2-3% edge for one grouping or alliance could mean a landslide in terms of seats, but a 1% lead may not be enough as yet. Sub-regionally, the DMK alliance is performing best in the Cauvery delta where it has a clear lead over AIADMK alliance. Even in the greater Chennai region, the former has an edge. Kongu land is seen to be giving a thumbs-up to the ruling AIADMK alliance, while the Vanniyar belt and the south are more equal fights.

Kamal Hassan could be a major x factor in this election although his party may not win any seats per se, because of very poor booth-level infrastructure. Similarly, TTV Dinakaran’s AMMK also has gained micro-traction, but he seemed to have a better booth-level strategy with presence in many parts at the tier 3 level.

As for BJP, our models are giving the party an edge in Kanyakumari and Coimbatore, although the latter is a bit more dicey than the former. At the basal level, NDA will vastly improve its vote-share, but how many seats will it add is a moot question. The chances of a DMK sweep are brighter than that of an NDA sweep which by default will give the Congress party much needed breathing space in an election where it is facing defeat almost everywhere.


Projections for Phase 2

As you already know, we are using two different models for projecting vote-shares and seat-shares in the 2019 election journey. Our vote-share model is essentially a macro-analytics projections model that takes the aggregate support of different parties in different sub-regions to arrive at vote shares.

For projecting seat-shares, instead of relying on the old methodology of simply converting vote-shares to seats which is fraught with lot of errors, we are instead deploying MAPi – Micro Analytics Projections (INTELLIGENCE) – our polling booth level tracker algorithm that uses voluminous amounts of data and thousands of simulations to arrive at the most probabilistic projections of number of seats.

[We have already analyzed Karnataka here. No polling/tracking was done in the state of Chhattisgarh because the local teams have been dismantled there after the recent assembly election debacle. Projections for Chhattisgarh are added into the model by extrapolating recent assembly election data].