The 2020 Bihar state elections turned out to be more closely fought than expected. The two main contending alliances, the National Democratic Alliance and the Mahagatbandhan, ended up with identical vote share (37.2%) but with a different number of seats – 125 against 110 in the 243-member assembly.

The Rashtriya Janata Dal emerged as the party but the largest number of seats but lost the election. Both the Rashtriya Janata Dal (which was part of the Mahagatbandhan) and the Bharatiya Janata Party (a member of the National Democratic Alliance) were dragged down by their coalition partner.

This election also saw a resurgence of the Left, which wins 16 seats, and the emergence of a new player in Bihar, the All India Majlis-e-Ittehad-ul-Muslimeen, which won five seats in the eastern tip of the state.

The Lok Janshakti Party played spoiler by running against the Janata Dal (United) but failed to win more than a single seat.

Close elections easily create conflicting narratives and explanations, as most actors concerned tend to read the result in any way that suits them. In this article, we look at the data that this election generated to provide a fact-based explanation of the outcome and provide an explanation on the data’s possible significance.

The data is based on Election Commission of India data, scraped on the day of the results and added to the historical data compiled by the Trivedi Centre for Political Data (TCPD-IED: TCPD Indian Electoral Dataset). The team of researchers at Trivedi Centre for Political Data has added a number of variables that enable a detailed analysis of party performance, which we offer here.

A low-turnout state

For a state reputed for its interest in politics, one cannot say that the 2020 Bihar elections generated wild enthusiasm among voters. Of the 73.8 million registered voters, 42.1 million cast their ballot (57%), a similar ratio as in 2015.

Of course, participation was higher than in the two 2005 elections, where not even half of voters participated. But it is far below the participation of its neighbors West Bengal (84.3% in the last election) or Jharkhand (65.4% in 2019). Even Uttar Pradesh, which has also a tradition of low participation in state elections, does better (60%).

In fact, Bihar is one of the rare states where fewer people vote in the state elections than in the general elections.

One consequence of these low turnouts is that it throws off projections of results based on population demographics, as we cannot predict who is not going to participate in the election. Another consequence is that as the political space is fragmented between two alliances and a series of small parties, parties can get away with mobilising narrow segments of the electorate.

As we will show below, no party since 2005 got more than 25% of total vote share.

The one noticeable thing about this election, however, is the gap between male and female voters’ participation, to the advantage of women. Of all registered women voters, 59.76% cast their vote, against 54.7% of all male voters.

The gap between them (5%) is large and significant, but only slightly lower than in 2015 (7%).

However, as Prannoy Roy noted, there is a differential of registration of 6% between men and women in Bihar (53% voters are male against 47%), compounded by the unfavorable sex ratio (918 in the 2011 Census).

In total, at least six lakhs women are missing in the electorate.

The turnout map reveals that participation was lower in the Bhojpur area, around Patna (in central Bihar) and in the Magadh region, where the Mahagatbandhan scored its best performance.

This could indicate that the performance of the Mahagatbandhan was probably more driven by the disenchantment against the National Democratic Alliance (or perhaps more specifically with one of its members, the Janata Dal [United]) than by a popular surge in favor of the Grand Alliance.

A phase-wise chart shows how participation rose in the third phase of the election, which covered the northern and eastern subregions of Tirhut, Mithalanchal, Koshi and Seemanchal. The National Democratic Alliance performed particularly well in these regions, as well as AIMIM in the east. This should temper the analysis according to which the Rashtriya Janata Dal gained momentum during the campaign. As data will show later, it performed better during the first phase of the election.

The most popular election

If these elections did not attract masses of voters, they did bring out large numbers of candidates, contesting in an ever-growing number of parties. There were 3,976 candidates who took their chance, against 3,450 in 2015. Most of them run as independents or small parties’ candidates. As a result, 86.7% of them lose their deposit (meaning they fail to score more than one sixth of the total votes polled in their seat).

A record 213 parties contested the election, a sharp jump from 2015, where only 167 parties were in the fray. This does not translate into greater diversity in the assembly as none of these new parties succeeded to win a single seat.

The AIMIM did open its account, winning five seats, but it had contested elections in Bihar before, in 2015. In total, 12 parties and one independent found their way to the assembly.

The geography of party performance

The electoral map shows how party performance was concentrated in specific subregions. While the BJP won most of the seats in the Tirhut, Saran and Northern Mithalanchal, the Janata Dal (United) concentrated its wins in the Koshi, Ang Pradesh and Southern Mithalanchal regions.

The Mahagatbandhan sweeps almost cleanly the three southern regions of Bhojpur, Central Bihar and Magadh. The last cluster lies in the East, in Seemanchal, where AIMIM won five seats around Kishanganj.

Party-wise winners in the 2020 Bihar state election

A longitudinal view of vote shares shows the state of fragmentation of Bihar politics. The three main parties – the Rashtriya Janata Dal, BJP and Janata Dal (United) – are oscillating within a 16%-23% vote share bracket. The Rashtriya Janata Dal might have emerged as the first party in terms of seats (75) and vote share (23.1%) but the fact is that the performance trends of these three parties are quite stable.

The coalition configuration contributes to that stability but so does the fragmentation of voters, who split their vote between parties in a way that prevents any of them from being able to form a government on their own.

No party ever has secured a majority of votes in the Bihar assembly, not even the Janata Party in 1977. The second instance of a party securing more than 40% of the votes was in 1962, when the Congress was dominating the state.

Of course, all-seat vote share measures are misleading since parties contest in alliances. Looking at parties’ vote share in the seats they actually contest gives a better measure of their performance. The chart below show that while the Rashtriya Janata Dal emerged as the first party, the BJP scored its best vote share in 20 years, garnering 42.6% of the votes across the 105 seats it contested. The Rashtriya Janata Dal follows closely behind at 39%.

The Janata Dal (United) and the Congress both trail behind their alliance partners, at 32.8% and 32.9% vote share. The Lok Janshakti Party loss of vote share stems from the fact that as it contested alone, it could not count on the vote transfer from other parties’ supporters.

The Congress seems to be on a path of recovery but that is an illusion, as the Congress’ vote share is partly a function of the number of seats it gets to contest – 70 in this election, against 41 in the previous election. However, the Congress can find solace in the fact that it got more vote share this time than in 2010, when it contested all the seats (but then, there was no vote transfer from partners to benefit from).

Speaking of vote transfers, one cannot deduce from these numbers that BJP voters and Rashtriya Janata Dal voters did not transfer as many votes towards their party’s partner than the other way around.

The differential could simply come from the fact that the Congress and the Janata Dal (United) were less popular and attractive than their ally. The two sub-regional vote share charts below show how the BJP and the RJD outperformed their partner across every region.

The Mahagatbandhan chart shows that the Communist Party of India (Marxist–Leninist) Liberation, which won 12 of the 19 seats it contested, performed as well as the Rashtriya Janata Dal and even outperformed it across four subregions. This indicates the presence of an actual resurgence of the Left, and not just an effect of alliance politics.

This resurgence concerns mostly the Communist Party of India (Marxist–Leninist) Liberation and not the two more mainstream Communist Party of India and Communist Party of India (Marxist). Any talk of a resurgence of the Left in West Bengal is certainly premature at the moment.

The following maps give the detail of the main parties’ vote share geography. The regional concentration of party performance should invite to nuance the narrative about a wave. The BJP does have appeal beyond the seats it won but it was not sufficient to help its coalition partner, the Janata Dal (United) avoid the setback it suffered in this election.

Beyond the three eastern regions, the performance of the Janata Dal (United) is quite poor, which is indicative of a dwindling support for Chief Minister Nitish Kumar. The Janata Dal (United) finished in third position in 12 seats.

To a large extent, the geography of parties’ results matches the geography of the regions where they contested. The BJP contested most of the seats in the north-west, the Janata Dal (United) in the and so forth.

But in the southern regions, where the Mahagatbandhan swept, both the BJP and to a larger extent the Janata Dal (United) were badly defeated by various Mahagatbandhan candidates. The “winner by party” map showed earlier indicates well that all components of the Mahagatbandhan performed very well against the National Democratic Alliance there in that region, including the communists, who made important gains.

This is significant because Census data reveals that there is a greater Dalit population in the Bhojpur area. The Mahagatbandhan did particularly well in reserved seats this election.

By comparison, the Rashtriya Janata Dal seems to have much wider geographical appeal than the Janata Dal United or the BJP, for that matter. Obviously, it contested more seats (144) but its vote share performance is better distributed than the Janata Dal United’s across the territory, even if they could hardly win seats in the East and in the North.

With the exception of a small cluster of votes in the sestern tip of the Tirhut region, the Congress does not have any sub-regional anchorage. Its candidates ended in the third position in six seats and even fourth in Chainpur.

The Rashtriya Janata Dal was therefore the only major party that contested across the entire state. Since the Congress no longer has any subregional strongholds, its tickets were scattered across the map. This necessarily means that the Rashtriya Janata Dal campaign had to be stretched over a much larger territory.

That might explain that party leader Tejashwi Yadav’s 18 to 19 helicopter rallies a day during the campaign stemmed more from geographical necessity than only temperament.

Even though the Lok Janshakti Party won only one seat, it did manage to make a dent into other parties’ vote base across most sub-regions. The Lok Janshakti Party had previously been a member of the National Democratic Alliance but fought this election on its own.

It also demonstrated that its vote base is not only large but also geographically scattered, as members of the Dusadh community (the party’s core electoral base) are present in significant numbers across most sub-regions. This makes the Lok Janshakti Party a perfect alliance partner but a poor contestant on its own.

While we can assume that the Janata Dal (United) suffered most from the Lok Janshakti Party’s lone ride in this election, we cannot assume that all the votes the Lok Janshakti Party got were taken away from the Janata Dal (United).

In any case, since the Lok Janshakti Party was a member of the National Democra in Bihar in 2015, those were not the JD(U)’s votes in the first place.

Finally, the map of the AIMIM’s performance helps realise that among parties, Asaduddin Owaisi’s formation is the one that won its seats most decisively. While theAIMIM failed to make a mark beyond the Seemanchal and Koshi regions, it won three seats with nearly 50% of the votes and more in areas where Muslim voters predominate.

In the seat of Amour (which means “love” in French, incidentally), Akhtarul Imam, a second-term MLA, won with 51.2% of the votes. This result is the outcome of years of groundwork and mobilisation by the party, which makes a remarkable entry into Bihar politics (especially when you consider that they have five times the representation of the Lok Janshakti Party).

NOTA, no please

In 2020, 7 lakh voters opted for NOTA, against, 9.5 lakhs in 2015. This year, the NOTA vote share was larger than the victory margin in 27 seats. But since we cannot tell how differently NOTA voters would have voted in absence of that option – or whether they would have voted at all – one cannot say with certainty that NOTA caused any particular seat to flip, even less in which direction.

The one thing that is certain is that the interest for NOTA, which was low at the beginning, is dwindling even more.

The seat share chart shows the decline of the Janata Dal (United) – which nearly won half of the seats in 2010 – to become the third party, by far, in this election. With 16 seats won, the Left gets its best seat share since 1995, when they won 11.7% of the seats.

As Vijdan Kawoosa indicated in the Hindustan Times, the seat share of the BJP kept increasing, phase by phase. The National Democratic Alliance won 29.1% of the seats in the first phase, 52,1% in the second and 66.7% of the seats in the third phase. The Mahagatbandhan followed a reverse path, winning 67.6%, 46.8% and 26.9% of the seats in each phase.

Small players still matter

An important fact of Bihar politics is that many voters still opt for parties other than the two main coalitions. Over the last four elections, about one voter out of five has choen a candidate who either belongs to a small local party or who runs as an independent.

If one includes the Lok Janshakti Party among small parties in this election, it means that 25% of all voters did not vote for either the National Democratic Alliance or the Mahagatbandhan.

Unlike in many other states, Bihar always had a few independent candidates winning seats. Before 2005, there were always seven to nine independent candidates who would win. This came from the fact that many locally entrenched leaders had the ability to win despite not being fielded by any party. Some of them were former party members, running on their own. Other were local bosses who can afford running without party affiliation.

Some of Bihar richest candidates are independent candidates, who have the financial surface to fund their own campaigns. That phenomenon is fading as barely one or two independents get elected nowadays. But they do collectively still receive 8% to 10% of all the votes, which is not negligible.

The net effect of this chunk of votes going to independents and small parties candidates is that it lowers the winning threshold required for major parties candidates.

Strike rates tell it all

Strike rates are a better measure of competitiveness in states with competing alliances. Simply put, they control for the number of seats parties contest, making the performance of parties comparable.

A quick look at overall strike rates of major parties gives away the key to understand this election. Within the National Democratic Alliance coalition, the BJP performed much better than the Janata Dal (United). It is usual that one party should perform better than the other. There is always some variation, but it is rare that the gap between members of the same coalition is as high as 30%.

While the BJP nearly won seven of every ten seats it contested, the Janata Dal (United) barely managed to win four. These numbers clearly expose the weakness of the Janata Dal (United) in this election.

The Mahagatbandhan strike rates reveal a similar gap between the Rashtriya Janata Dal and the Congress. While the Rashtriya Janata Dal managed to win slightly more than half of the seats it contested, Congress could not get three out of ten.

The Rashtriya Janata Dal must bitterly regret to have ceded to the Congress’ demand of contesting in 70 seats. This was clearly beyond the Congress’ reach and it should have been clear that arguing for a high number of seats by virtue of the national status of the Congress was a bad idea.

This gap in performance between partners, as political scientist Neelanjan Sircar succinctly put it, “is the story of this election”.

A closer examination of those strike rates shows that the BJP and Janata Dal (United) performed best against Congress candidates and that the Rashtriya Janata Dal had a much better strike rate against the Janata Dal (United).

This points to the notion that there were really two elections fought at the same time. One between the RJD and its ally, the Janata Dal (United), in which the former is a clear victor. And a second, between the Mahagatbandhan and the BJP, in which the latter showed great strength and resilience, despite the multiple crises that have affected India and Bihar in particular over the past few months.

Another sign of the BJP’s strength is that its strike rate improved over time in the campaign, between phases, while the Rashtriya Janata Dal campaign was gaining momentum. Interestingly, the discontent expressed against Chief Minister Nitish Kumar during the campaign did not seek to affect BJP candidates very much, who still have the figure of Prime Minister Narendra Modi to motivate them to vote for the saffron party.

And perhaps there was even a third election, fought between the Lok Janshakti Party and the Janata Dal (United). Many rumors are circulating about an alleged plot of the Lok Janshakti Party and the BJP to undercut Nitish Kumar in order to gain weight within the alliance. One cannot prove or disprove such notions, but it is clear that the Lok Janshakti Party approached this election with the sole mission of undercutting the Janata Dal (United).
Historians will have to figure what exactly happened and why. Here is not the place to speculate on this matter. Suffice to say that if the plan was indeed to undercut the Janata Dal (United), it worked.

The Janata Dal (United) trails in 60 seats and the Lok Janshakti Party scores an average of 12.3% of vote share across the 115 seats it contested against the Janata Dal (United). If one assumes that nearly all the Lok Janshakti Party vote share would have gone to the Janata Dal (United) had it stayed with the National Democratic Alliance (which one cannot assume), it could have helped the Janata Dal (United) win, hypothetically, 33 more seats.

There were also 37 seats that the Janata Dal (United) would have lost anyway, even if you add to it (again, hypothetically) the Lok Janshakti Party vote share.

Had the Janata Dal (United) performed in those circumstances, it would have argued that the popularity of the chief minister was intact, if not improving. In a way, the lone ride of the Lok Janshakti Party gives a better measure of the dwindling popularity of Nitish Kumar.

Thin margins

Those who followed the results live have been unnerved by the uncertainty of the outcome. One cause of uncertainty were the thin margins that led many seats to flip from one alliance to another until they settled for their winner.

An examination of party-wise victory margins shows that the smaller parties won their races more decisively than major parties. The three communist parties and the AIMIM won their seats with large margins, which indicate popular support for their platform.

In the case of the Left parties, this shows again that their candidates received support for who they were, rather than simply as transfer of votes from other parties’ supporters.

Among major parties, the Janata Dal (United) won with the thinnest margins, followed by the Congress. Forty seats were won with margins of less than 2%, which means that small numbers in a limited number of seats could have produced a different result.

But in reality, the distribution of those seats is pretty fair between the two alliances (20 for the National Democratic Alliance and for 18 the Mahagatbandhan). Given the difference of seats at the end of counting, it is unlikely that the election could have easily flipped for this reason.

Conclusion

No electoral outcome gets shaped by one factor alone and the Bihar election was of course no exception. What the data makes clear is that Chief Minister of Bihar starts his fourth term with a diminished stature.

The National Democratic Alliance won the election by the skin of its teeth and the diminishing popularity of Nitish Kumar almost cost them the election. Considering the multitude of factors going against him – mishandling of the floods, the pandemic, the migrant issue, joblessness and rising criminality and the cunningness of the Lok Janshakti Party – it is impressive that the National Democratic Alliance still managed to pull through.

However, the data shows that the outcome was not determined by the under-performance of the Janata Dal (United) alone. The BJP did score majorly, seemingly unaffected by the dire economic context.

The Rashtriya Janata Dal proved that it remains a competitive force in Bihar politics. It performed well but not as well as it could have, dragged down by the Congress Party. Its wide coverage of Bihar’s political map means that its votes get also more distributed across space, giving an edge to the National Democratic Alliance in a first past the post system.

The Lok Janshakti Party may have proven its point that it is a great party to have as an ally and an effective spoiler when running on its own. Except perhaps the Congress, everyone can find some reason to be satisfied with the result. In the polarised times we live in, it is not the worst of outcomes.

All authors work at the Trivedi Centre for Political Data. Mohit Kumar is a Data Scientist and GIS Engineer, Neelesh Agrawal is a Research Engineer. Gilles Verniers is the Co-Director, and Assistant Professor of Political Science at Ashoka University. The data is drawn from TCPD-IED: TCPD Indian Electoral Dataset, 1962-current. Priyamvada Trivedi, Ananay Agarwal and Basim-U-Nissa contributed to the data.