What does the new Bihar assembly look like? What is the profile of the 243 MLAs who have been elected to serve Bihar’s citizens over the next five years? And what does a timeline of data on the sociological composition of the assembly tell us about Bihar’s political class and politics more broadly?
In this article, we look at the representational outcomes of the 2020 Bihar election in terms of demographics (caste, gender and religion) and other variables, drawn from the affidavits that each candidate signed when they filed for their nomination.
The demographic data has been collected before and during the campaign by a team of researchers at the Trivedi Centre for Political Data. The affidavit data has been processed from the data made available by the Association for Democratic Reforms.
Decoding the Bihar results in 32 charts: Turnouts, vote shares, victory margins and more
Except for perhaps the Congress, all the other parties can find some reason to be satisfied with the result.
Our data reveals that 40.7% of the members of the new assembly belong to the Other Backward Classes category, above the upper castes, who make nearly 30% of the Assembly. Scheduled Caste representation is fixed at 16.5% of the seats (seldom do any Scheduled Caste candidate win in general seats) and Muslim make for nearly 8% of the new assembly.
If we look at the data in absolute numbers instead of percentages, most upper-caste MLAs were elected on a Bharatiya Janata Party ticket (34), followed by the Rashtriya Janata Dal (13), the Janata Dal (United) (10) and Congress (7).
The Rashtriya Janata Dal contributes to the largest number of OBC MLAs (39), followed by the BJP (27), the Janata Dal (United) (22) and Congress, far behind (2).
These numbers reflect quite closely the distribution of major caste groups among the candidates of these parties. Representational outcomes are primarily the product of the selection operated by parties when they recruit their candidates, rather than by voters.
The BJP gave 52 tickets to upper caste candidates out of 110, against 39 tickets to Other Backward Classes candidates, 15 tickets to Scheduled Caste candidates, one Scheduled Tribes and no tickets given to Muslims (the caste identity of three BJP candidates could not be ascertained).
The Rashtriya Janata Dal gave 69 of its 144 tickets to OBC candidates, compared to 23 to upper castes. Scheduled Caste and Muslim candidates received nearly the same number of tickets (19 and 18). The caste of 14 Rashtriya Janata Dal candidates could not be ascertained. (We assume that most of them must belong to the OBCs but we cannot be certain.)
The Janata Dal (United) gave 59 tickets to OBC candidates, against 23 to upper caste candidates, 18 members of the Scheduled Castes, 11 Muslims, one member of the Scheduled Tribes. (The caste identity of three party MLAs could not be ascertained).
The distribution of tickets and seats by party therefore shows that both Congress and the BJP predominantly cater to upper castes representation, while the Rashtriya Janata Dal and the Janata Dal (United) mainly favour OBC representation. Upper castes make up for nearly half of all BJP MLAs, while the majority of Rashtriya Janata Dal and Janata Dal (United) MLAs belong to the OBCs.
In other words, these charts also show that some expected alignments between parties and caste groups persist but that there also a broad distribution of representation within parties that do distribute tickets across castes.
Obviously, one should look beyond these large caste clusters and consider jatis as a more relevant unit of observation. There again, we can find the expected alignments between specific castes within these groups and parties, and also some nuance.
For instance, the BJP distributed 24.5% of all its tickets to Rajput candidates, against 11.8% of tickets to Brahmins, and 7.3% of tickets to Bhumihars. Banias, who are considered as OBCs in Bihar, also find substantial representation compared to their demographic weight (9.1%). The rest of the tickets distributed to OBC candidates is split in half between Yadav (13.6%) and non-Yadav OBCs (22%).
The Janata Dal (United) continues to favour Kurmis among OBCs (14% of all its tickets) but gave quasi-equal representation to Yadav candidates (13%). The remaining 25% of tickets given to non-Yadav and non-Kurmi candidates were divided between at least 12 other groups, each receiving a handful of tickets.
The Rashtriya Janata Dal remains true to its base by distributing nearly a third of its tickets to Yadav candidates (31%). The remaining 70% was distributed among a large array of castes, chosen according to local demography and local patterns of dominance. 12.5% of its tickets were given to Muslim candidates, across most sub-regions.
This is an old electoral recipe in Hindi belt politics, which involves seeking to tie one’s core support base with whoever else is in a position of influence, if not dominance, locally. This can yield dividends when the electorate is as fragmented as it is in Bihar, but can also backfire when the votes that make a difference in closely contested races, come from groups that are not aligned with any specific party, or who do not enjoy the kind of local dominance that attracts the interest of parties. Courting local elites may come at the cost of broadening support among non-elite groups.
The Congress distributed 40% of its 70 tickets to upper-caste candidates and 17% to Muslim candidates, again across sub-regions. Beyond a handful of tickets given to Yadav candidates, other OBCs hardly found representation in the Congress’s roster of candidates.
The outcome is that, regardless of the volatility of these elections, the longitudinal trends of caste-based representations have not dramatically changed.
Long-term trends hold
A timeline of the data shows how these results fall within larger stable representational trends, as the share of seats occupied by major caste groups has not varied significantly since the mid-1990s, which followed a decade of deep social and political churning in Bihar.
The data reminds us that the real period of churning in Bihar caste politics took place in the late 1980s with the rise of regional parties and the subsequent ascension of Lalu Prasad Yadav. It was during this crucial period that the balance of power flipped from forward to backward castes, as Jeffrey Witsoe recounts in his book, Democracy Against Development.
After the rise of Lalu Prasad Yadav, the share of upper-caste MLAs starts increasing again, as traditional elites find representation across parties, and not just within the BJP or Congress. Other parties, notably the communists, also distribute a significant number of their tickets to upper-caste candidates.
In 2000, the Lalu Yadav’s Rashtriya Janata Dal sent more upper-caste MLAs to the Vidhan Sabha (20) than the Janata Dal (United) (15) and nearly as many as the BJP (22). The Congress still sends a few upper caste MLAs to the assembly and through the period, about a dozen upper caste, independent candidates succeed in getting elected.
So even if it is true that some broad caste-party alignments persist, there is more fluidity to party-caste relations than most people often think.
The resilience of Yadav representation
In 2020, the BJP and the Janata Dal (United), distributed between 13% to 14% of their tickets to Yadav candidates, most of whom were elected in the process. The simple fact that Yadav candidates find substantial representation in two major parties in addition to the Rashtriya Janata Dal ensures that they retain a dominant presence in the assembly.
In 2015, one MLA out of four was a Yadav. This year, the ratio of one out of five.
Since 1990, the ratio of Yadavs among OBC MLAs has not gone below 40%, with the exception of 2010. Kurmis and Banias get substantial representation but every other OBC group, particularly members of the Economically Backward Classes, hardly get represented, beyond a handful number of seats distributed across a large array of small groups.
Rajputs on the rise
A similar dynamic, although not as dramatic, plays out among the upper castes. Since 2000 and the rise of the BJP, Thakurs have received the lion’s share of representation, among the upper caste (above 40%). Bhumihars, once a dominant force in Bihar politics, have been on the decline over the same period.
Brahmins have maintained themselves around 20% of all upper caste MLAs but have not been able to gain from the progression of the BJP in the state.
In Bihar, much like in the rest of the Hindi Belt, the BJP becomes more and more a Rajput party – not so in terms of social bases but in terms of composition of its elected representatives. The fact that non-Rajput MLAs belong to a large array of castes that are individually poorly represented contribute to the strengthening of the Rajputs generally, including on local party structures.
A stable caste geography
Finally, a quick look at the geography of representation illustrates further that political representation of castes and party performance are linked – but only to an extent.
The two maps below display the caste identity of the MLA and show that important clusters remain nearly identical, between 2015 and 2020.
The Tirhut and Saran regions, in the north west, remain largely dominated by upper castes. Other clusters exist, overlapping central Bihar and the Munger region, or the western tip of the Magadh region.
There remains a cluster of seats held by Muslim MLAs in the Seemanchal region, where Muslim make the majority of the population.
Much of the Mithalanchal, Koshi and Munger region remain dominated by OBC representation.
The 2020 map shows that representation of Muslims has actually receded in the Tirhut and Mithalanchal regions, on account of most parties reducing the number of tickets given to Muslim candidates.
This is a phenomenon that we have observed in other states, like Karnataka. Regardless of the political volatility that exist in a state, regardless of the competitiveness of elections and uncertainty of outcomes, social patterns of representation remain quite stable.
This is a product of demography (castes are not evenly distributed across the territory) but also of the fact that power is anchored locally, through the control that locally dominant group exert over politics, institutions, and the economy. This is an indication of the resilience of local patterns of social dominance, which still shape Bihar society.
An examination of other socio-demographics variables enables us to deepen that idea further, by considering disparities of representation alongside gender, religious, financial and integrity lines.
Women candidates up, women MLAs down
One of the great disappointments of this election is the decrease in women representation, despite the fact that more women contested, including on major parties’ tickets, and despite the fact that women outvoted men by 5 percentage points.
In the 2020 elections, 371 women contested, compared to 273 in 2015. The two alliances – the National Democratic Alliance and the Mahagatbandhan – fielded 62 women candidates. The National Democratic Alliance had 37 women candidates while the Mahagatbandhan had 25.
Two parties, the Janata Dal (United) and the Lok Janshakti Party, put in more effort than usual by fielding 22 women candidates each (out of 115 and 135 candidates respectively).
But despite this modest progression, the representation of women in 2020 decreased, from 28 women MLAs to 26. The map below offers a depiction of how small and dispersed women representation is in Bihar.
The effort of the Janata Dal (United) and Lok Janshakti Party is visible when compared to 2015 nomination data, when only 8% of all candidates were women. In 2020, the Janata Dal (United) gave 19% of its tickets to women candidates and the Lok Janshakti Party gave 16% of its tickets to women.
The chart also shows how women are even more absent among the three communist parties, who this year nominated only one woman candidate out of 29. It is a tradition of the Left to exclude women from political representation in state and national elections.
In Bihar, various Left parties have fielded 2,419 candidates since 1962. Only 80 of them were women and only five of them have won a seat, in the 1980s and 1990s (Sunaina Devi, Shakuntala Sinha, Manju Prakash (twice) and Madhvi Sarkar).
Why this contrast between nomination and representation? One needs to look at variations in nomination among parties and within alliances. Within the National Democratic Alliance, the Janata Dal (United) fielded more women than the BJP (22 against 13). Both the Rashtriya Janata Dal and Congress distributed 11% of their tickets (16 and 8) to women candidates.
As the Janata Dal (United) underperformed in the election, so did its women candidates. Only six of them won their race.
A further examination of the data shows that in most aspects, women candidates do not differ from their male counterparts. They share the same distribution of caste, age and educational background. They have about the same amount of political experience as men (54% of all women candidates were first-time contestants, against 46% of all men).
More women tend to belong to political families than men but that comes as a patriarchal imposition from parties who tend to select their women candidates among the relatives of politicians.
AIMIM keeps Muslims’ representation stable
If it hadn’t been for the All India Majlis-e-Ittehad-ul-Muslimeen, it is likely that no one would have spoken about Muslims in this election. They were absent from the discourse of other parties and most major parties abstained from giving many tickets to members of India’s largest minority.
Nineteen Muslims have been elected in the 2020 Vidhan Sabha, compared to 24 five years ago. Eight were elected on a Rashtriya Janata Dal ticket, five from the AIMIM, four from Congress and one each from the Bahujan Samaj Party and the Communist Party of India (Marxist Leninist ) Liberation. Muslims made up less than 10% of all candidates, and 8% of the MLAs. Muslims in Bihar represent 16.5% of the total population.
Among major parties, the share of Muslim candidates gradually increased from the early 1990s. The electoral strategy of the Rashtriya Janata Dal (Muslims + Yadavs) led that party to increase the number of tickets given to minority candidates. The Left also contributed a bit to their representation, as well as a number of emerging Dalit parties, who opened their ranks to Muslim candidates.
After 2005, there has been a sharp decline of Muslim candidates fielded by major parties, from 17% of all candidates to 11%. This includes the Rashtriya Janata Dal and the Congress. In 2010, the Congress gave 19% of its tickets to Muslim candidates, against 6% in 2015 and 2020.
As a result, Muslim’s representation has stagnated between 8% and 9%, with a dip at 6.5% in the second election of 2005. The compensation came from Muslim elected on the tickets of other parties, like the Lok Janshakti Party or now the AIMIM.
An examination of Muslim candidates’ nominations by major parties over time shows how after 2000, most Muslim candidates have been fielded by smaller parties. In 2020, 61% of all Muslim candidates contested under tickets from parties like the Lok Janshakti Party, the Samajwadi Party, the Bahujan Samaj Party and of course the AIMIM. Barring the AIMIM, which won five seats, all these other parties are marginal players, at least in terms of seats.
Of course, the fact that there are more parties now contributes to the change in ratio between major and minor parties. But the overall ratio of Muslim candidates fielded by major parties compared to their other candidates also diminishes over time. After 2010, the absolute number of tickets given to Muslims by both the Janata Dal (United) and the Rashtriya Janata Dal has been reduced by half, from around 60 to 32.
The party-wise distribution of Muslims MLAs shows how the Rashtriya Janata Dal remains the main provider of representation to Muslims in Bihar and how the Congress lost that status early 1990s onwards.
For the first time, the Janata Dal (United) does not send a single Muslim in the state assembly. None of its 11 Muslim candidates succeeded. One can assume that an alliance with the BJP did not work in their favour.
In the following sections, we look at the educational profile of the new assembly, as well as the distribution of self-declared occupations among MLAs, their level of wealth and crime.
An aging, highly educated assembly
In terms of education, there is not much change in 2020. The bulk of MLAs hold graduate degrees and above, in a slighter higher proportion than in 2015. Twenty three MLAs hold a PhD, nearly 10% of the assembly (14 come from the National Democratic Alliance and nine from the Mahagatbandhan).
Overall, the educational profile of MLAs by alliance does not offer much contrast. There are slightly more graduates and above among the Mahagatbandhan MLAs than among the National Democratic Alliance MLAs, but the difference is thin.
In terms of age distribution, the 2020 assembly is not as youthful as the 2015 assembly. Obviously, those elected in 2015 have aged, but only 86 incumbent MLAs have been re-elected. Between the two election, it is the 35-45 age bracket that has diminished, while nearly 40% of the assembly is above 55 years.
That is still a high proportion of younger politicians, compared to most other states.
The differences between parties are not striking. The Left parties and the Rashtriya Janata Dal have a higher proportion of MLAs below the age of 55 (75% and 68%), against 54% and 58% for the BJP and the Janata Dal (United).
There are high ratio of fairly young MLAs across the board.
These ratio were even higher five years ago for all parties, which shows that no one in particular is responsible for making the assembly younger or older.
In terms of self-declared occupation, only one MLA out of five declares themselves to be in agriculture, a sharp contrast with the past. One third of all MLAs are in some business, while another fifth declares politics to be their principal occupation.
Liberal professions, which include lawyers, doctors, engineers and so forth, make up for only 3% of the assembly. Also under-represented, salaried professionals, educationists and social workers, a category usually large in other North Indian states.
Of course, these are self-declared occupations that can overlap. They are also quite hazy. What does social work mean? Many politicians claim “seva” to be their main occupation, while they maintain some kind of business elsewhere. Many businessmen once elected also choose to become “social workers”.
These categories are useful in as much as they tell us which occupational profile MLAs choose to publicise.
Compared to 2015, the share of self-declared businessmen increases, from 25% to 33%. The ratio of social worker drops significantly, as well as the number of pensioners.
When we break down the data by party, we see some contrast. Forty five percent of Rashtriya Janata Dal MLAs declare to be in some business, against 33% for the BJP and 25% for the Janata Dal United.
More Janata Dal (United) MLAs are likely to introduce themselves as professional politicians. Particularly those who have been reelected and may have opted indeed for politics as a life career. The share of farmers among Rashtriya Janata Dal members has also gone down significantly.
According to data from the Association for Democratic Reform, 33% of all candidates to the Bihar election in 2020 owned more than Rs 1 crore of assets, against 25% of all candidates in 2015. Among major parties, the ratio of crorepati candidates and MLAs in 2020 increases to 86% and 78%, respectively.
For most major parties, the median wealth (the average of total assets minus liabilities) of MLAS largely exceeds the average wealth of the candidates. The Hindustani Awam Morcha and two communist parties are the exception, where the successful candidates are not among the wealthiest of their candidates (but the numbers are in any case small. A aase in point, one of the two Communist Party of India MLA, Ram Ratan Singh from Teghra, is a multi-crorepati).
The Lok Janashakti Party has been removed from this chart as it is an outlier. Its lone elected MLA, Raj Kumar Singh from Matihani, is much richer than his party’s average (Rs 12.2 crore vs. Rs 1.9 Crore), although not the richest of the party’s candidates (there are at least eight candidates who are officially richer than him).
The same goes for the one Vikassheel Insaan Party MLA, who is ten times richer than the average it his party’s candidates (Rs 6.3 crore vs. Rs 64 lakh).
The contrast with 2015 is impressive, as the median level of wealth of both candidates and winners of major parties almost double. The candidate from Mokama, Anant Kumar Singh, has the distinction of being the richest MLA in the new assembly, with Rs 51 crore of net declared assets.
The affidavit data also reveals that the ratio of candidates with criminal charges has also increased significantly. In 2015, 54.5% of all major parties candidates (the two alliances and the AIMIM) and 58.2% of their MLAs had criminal cases against them. Those figures (including the same parties) in 2020 jumped to 61.7% and 66.8%, respectively.
A comparison of the distribution of number of cases per MLAs in 2015 and 2020 shows that as the overall number of MLAs with criminal cases increases, so does the number of cases per MLA.
Anant Kumar Singh, already seen as Bihar’s wealthiest MLA, is also the MLA with the highest number of criminal charges (38, including 11 charges related to attempt to murder and 4 charges related to kidnapping or abducting in order to murder).
A party-wise comparison shows that the level of alleged criminality among candidates and MLAS is well distributed among parties. The Janata Dal (United) has lower level of candidates with criminal charge 48.7%) than the three other main parties and smaller parties, with fewer candidates, tend to have the highest levels (except the Hindustan Awam Morcha and the LJP).
Interestingly, the ratio of MLAs with criminal charges tends to be higher than the ratio of candidates with with criminal charges among the Mahagatbandhan MLAs. For the BJP, Janata Dal (United) and Hindustan Awam Morcha, that ratio is slightly smaller.
But still, the data makes it obvious that parties seek to recruit candidates with a criminal pedigrees, for the simple reason that they tend to do well in the polls.
A comparison of 2020 and 2015 seems to suggest that this phenomenon is growing rather than receding, despite (and perhaps because) of the increased publicity that these candidates have received, under Election Commission of India order, in 2020.
To sum it up, the Bihar 2020 assembly looks very much like the 2015 assembly. Most MLAs are male, fairly well-educated, rich, in contradiction with the law and tend to belong to groups that exert local dominance. On all these variables, the under-privileged segments of society remain largely excluded from the assembly.
If anything, 2020 represent a slightly exacerbated version of the 2015 assembly, in which the representation of all privileged segments across major caste groups Improves. This suggests that the overall elite character of Bihar’s Vidhan Sabha is not going away any time soon.
Politics in Bihar was pitched by regional parties in the 1990s as a formidable competition between the forward castes, associated with national parties, and the backward castes, supporting the regional parties. This was articulated as the politics of privilege against the politics of dignity.
The politics of Bihar over the past 15 years show that this binary imagination is well behind us, as all parties seek power by catering by and large to the same entrenched elite interests, with some variation between them. Time will tell how long the elitism of Bihar politics will be able to sustain itself.
All authors work at the Trivedi Centre for Political Data. Basim-u-Nissa is a Researcher, in charge of caste data collection. 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-ILD: TCPD Indian Electoral Dataset, 1962-current, part of the CNRS SPINPER project. Priyamvada Trivedi, Ananay Agarwal and Samridhi Hooda contributed to the data.