The 2015 Bihar Assembly Elections Calls for Cautious Optimism from the Left

November 20, 2015

By Deepankar Basu

Left, democratic and secular forces in India are justifiably pleased at the results of the general elections to the Vidhan Sabha (Legislative Assembly) of Bihar in 2015. The National Democratic Alliance (NDA), led by the quasi-fascist, right-wing Hindu nationalist BJP, got a solid drubbing at the polls. Despite using its enormous money and muscle power, its fabled electoral machine, its list of star crowd pullers, including the Prime Minister himself, the NDA ended up with only 58, out of a total of 243, assembly seats. The primary opponent of the NDA was the Mahagathbandhan (MGB), an alliance of three secular parties, the Janata Dal United (JDU), the Rashtriya Janata Dal (RJD) and the Indian National Congress (INC). The MGB swept the polls by winning 178 – close to three-fourth of the total – seats

There can be two very different reasons for the phenomenal win of the MGB, and they lead to different political conclusions. One scenario could be that the electoral outcome might be the reflection of a massive shift in voter preference away from the right-wing BJP and towards the secular alliance of the MGB. If this is the case, then Left, democratic forces need to be especially happy because the threat of fascism, with its continuous attack on minorities and labour, might well have blown over. But there is another possibility. The electoral outcome might be more a result of the first-past-the-post system than the reflection of any significant shift in voter preference away from right-wing politics. If this is the case, then Left forces need to be more cautious in their celebrations. In fact, they need to redouble their efforts at fighting the growing tendencies towards fascism in the country.

One way to distinguish between these two very different logics underlying the electoral outcome in 2015 – MGB’s win and the NDA’s rout – would rely on a careful comparison of the current with past results to discern any possible changes in the electoral support of the BJP. Such a comparative analysis will need to deal with at least the following issues: (a) which elections should be compared?, and (b) which electoral outcome should be compared and how should one deal with changes in electoral alliances?

Which elections should be compared? There are at least two past elections in Bihar that can be used for comparison with the Vidhan Sabha elections in 2015. The first would be the previous Vidhan Sabha elections of 2010, and the second would be the previous Lok Sabha elections of 2014. Both comparisons have advantages and disadvantages. Comparing Vidhan Sabha elections seem most intuitive because they operate by certain common political, electoral and arithmetic logic, and such logics are usually different from those operating during Lok Sabha elections. But, in the case of Bihar, considerations of electoral alliances also recommend a comparison of 2014 and 2015. This is because electoral alliances have changed drastically between the 2010 and 2015 Vidhan Sabha elections. But the changes are less dramatic between the 2014 Lok Sabha elections and the 2015 Vidhan Sabha elections. Hence, in this article, I use both comparisons.

Comparing Electoral Outcomes?

Which outcome should be compared across multiple elections, including a comparison between Lok Sabha and Vidhan Sabha elections? Many analysts focus on the number of seats won by parties as the primary measure of electoral success. In one sense, this makes sense: the number of seats won by parties is really the factor that affects the composition of the government. But the number of seats won might not be the best measure to use if one is trying to gauge the level of electoral support of a party. This is because in India’s first-past-the-post (FPTP) electoral system, even small changes in vote share can lead to large changes in seats won and lost. Thus, a better and more stable outcome to analyze electoral support is a party’s vote share.

Table 1: Bihar Assembly Elections, 2010
Seats ContestedSeats WonVote Share in StateAdjusted Vote Share
Bharatiya Janata Party1029116.4939.29
Indian National Congress24348.378.37
Janata Dal (United)14111522.5838.91
Lok Janshakti Party7536.7421.84
Rashtriya Janata Dal1682218.8427.25
NDA (BJP+JDU)24320639.0739.07
RJD Alliance (RJD+LJP)2432525.5825.58
Left Parties (CPI+CPM+CPIML)19014.195.36
Note: adjusted vote share is the ratio of “vote share in State” and “fraction of seats contested”.  Sources. Data for the 2010 Assembly Election is from Statistical Report on General Election, 2010 to the Legislative Assembly of Bihar, Election Commission of India

But comparing vote share across elections is not adequate by itself. This is because changes in vote share can be driven by changes in the number of seats contested even if the level of support for each contested seat remains unchanged. A better measure seems to be what I will call “adjusted vote share”, which is the statewide vote share adjusted for the number of seats a party contested in an election. For all parties that contested at least 1 seat in an election, the adjusted vote share is computed as the ratio of two quantities: (a) the statewide vote share won by a party, and (b) the fraction of the total seats it contested.[1] Thus, if a party contests all seats, its statewide and adjusted vote share would be the same. Moreover, between two parties which had the same statewide vote share, the one which contested fewer seats would get a higher adjusted vote share.

The adjusted vote share provides an approximate measure of the degree of electoral support of a party. If a party fights elections on its own, then this is possibly quite close to its true support base. But when a party is part of a coalition, we can only get an approximate measure. This is because of the phenomenon of “vote transfers”. In the seats that the party contests, the party might get votes from its coalition partner’s supporters – a transfer to the party. On the other hand, in the seats that it offers up to its coalition partners, the party’s supporters do not get to vote for it, but might instead vote for the coalition partner – a transfer away from the party. If these two transfers are equal, the adjusted vote share would give an accurate estimate of a party’s electoral support; if they are not, we will get only an approximate picture.

Table 2: Bihar Lok Sabha Elections, 2014
Seats ContestedSeats WonVote Share in StateAdjusted Vote Share
Bharatiya Janata Party302229.4039.20
Indian National Congress1218.4028.00
Janata Dal (United)38215.8016.63
Lok Janshakti Party766.4036.57
Rashtriya Janata Dal27420.1029.78
Rashtriya Lok Samta Party333.0040.00
NDA (BJP+LJP+RLSP)403138.838.80
UPA (INC+RJD)39528.529.23
Left Parties (CPI+CPM+CPIML)2902.743.78
Note: adjusted vote share is the ratio of “vote share in State” and “fraction of seats contested”.  Sources. Data for the 2014 Lok Sabha Elections is from the website of the Election Commission of India:


In Table 1, 2 and 3, I have collected together some information about major parties and coalitions that participated in, and made a difference to, the 2010, 2014 and 2015 elections, respectively. The first and second columns in these tables show the number of seats contested and won by each of the major parties and coalitions. The third column shows the percentage of total votes that went in favour of each party in the state as a whole, what I will refer to as the statewide vote share. In the last column, I report the “adjusted” vote share for each party or coalition.[2]

Table 3: Bihar Assembly Elections, 2015
Seats ContestedSeats WonVote Share in StateAdjusted Vote Share
Bharatiya Janata Party1595324.4037.29
Indian National Congress41276.7039.71
Janata Dal (United)1017116.8040.42
Lok Janshakti Party4024.8029.16
Rashtriya Janata Dal1018018.4044.27
Rashtriya Lok Samta Party2322.6027.47
Hindustan Awam Morcha2112.3026.61
NDA (BJP+LJP+RLSP+HAM)2435834.1034.10
MGB (INC+JDU+RJD)24317841.9041.90
Left Parties (CPI+CPM+CPIML)23433.583.72
Note: adjusted vote share is the ratio of “vote share in State” and “fraction of seats contested”. Sources. Data for the 2015 Assembly Election are from the website of the Election Commission of India:

How did the BJP Perform?

The BJP’s statewide vote share increased from 16.49% in 2010 to 29.4% in 2014, and then declined to 24.4% in 2015. Part of the increase, and then decrease, is because the BJP contested different fractions of the total number of seats in the three elections. In 2010, the BJP contested 42% of the seats; in 2014, it contested 75% of the seats; and in 2015, it contested 65% of the seats. That is why it is more informative about electoral support to compare the adjusted vote share in the last column of the three tables. BJP’s adjusted vote share declined mildly from 39.29% in 2010 to 39.20% in 2014, and declined further to 37.29% in 2015. Between 2010 and 2015, there is a decline of 2% in the adjusted vote share; between 2014 and 2015, there is a corresponding decline of 1.91%. While these declines are not insubstantial, it would be an exaggeration to portray this as a massive move away from the BJP or a strong anti-BJP verdict. The data in Tables 1, 2 and 3 suggests that the electoral support base of BJP is more or less intact through the three elections, with a small decline in 2015. There is no evidence of any significant move away from the right-wing, Hindu nationalist party.

This conclusion would be further strengthened if we compare the performance of BJP’s alliance partners – LJP and RSLP – between 2014 and 2015. The adjusted vote share for LJP and RSLP were 36.57% and 40.00% in 2014. The corresponding numbers in 2015 were 29.16% and 27.47%. Thus, while BJP’s adjusted vote share declined by 1.9% between 2014 and 2015, the adjusted vote shares of its main alliance partners, LJP and RSLP, declined by 7.41% and 12.53% respectively. This seems to suggest that the decline in NDA’s electoral fortunes between 2014 and 2015, when its vote share declined from 38.8% to 34.1%, is more likely the result of declining electoral support of the LJP and RSLP, rather than of the BJP.

Could the substantial declines in the adjusted vote shares of LJP and RSLP have been caused by lack of transfer of BJP votes to these parties (in the seats where they contested)? This is plausible but seems unlikely for several reasons. First, the fact that BJP’s adjusted vote share declined by far less suggests that BJP’s core voters supported the NDA across the state. If BJP core voters deserted the BJP/NDA in 2015, that should have emerged in the data for the adjusted vote share for BJP too. Second, the seat sharing arrangement must have been at least partly driven by the perceived relative strengths of these parties. Thus, seats where the LJP and/or the RSLP contested must have been places where they were perceived to have relatively more electoral support. The fact that they lost vote share suggests that their erstwhile voters must have switched to other parties.

How did the Secular Parties Perform?

The Janata Dal (United) has recovered the electoral ground it had lost between 2010 and 2014. The JDU’s statewide vote share fell from 22.58% in 2010 to 15.8% in 2014, and then climbed back up to 16.8% in 2015. But the JDU’s gains can be better highlighted by comparing the adjusted vote share, which had more than halved from 38.91% in 2010 to 16.63% in 2014. In 2015, it moved back up to 40.42%.

For the RJD, the 2010 assembly elections were the worst moments. Since then, it has been gaining and consolidating support (even if that did not always get converted into seats). The RJD’s statewide vote shares were 18.84%, 20.10% and 18.4% in 2010, 2014 and 2015, respectively. This fluctuation was partly the result of different fractions of total seats contested. So, let us look at the evolution of the adjusted vote share, which controls for changes in seats contested. RJD’s adjusted vote share has continuously increased over the last three elections. In 2010, it was 27.25%, in 2014 it was 29.78%, and in 2015, it increased to 44.27%.

The INC has also seen a continuous improvement in its electoral fortunes once we take account of the number of seats contested. Its adjusted vote share increased from 8.37% in 2010 to 28.00% in 2014, and then jumped up to 39.71% in 2015.

The most significant fact relating to the performance of the secular parties in the MGB (JDU, RJD and the INC) is the substantial jump in their adjusted vote shares between 2014 and 2015. The JDU’s adjusted vote share jumped from 16.63% to 40.42%; the RJD’s adjusted vote share increased from 29.78% to 44.27%; and the INC’s adjusted vote share increased from 28% to 39.71%. This suggests that the transfer of votes between the three coalition partners was very effective, and highlights the effectiveness of their pre-poll alliance.

How did the Left parties perform? If we consider the three main Left parties together as a group – CPI, CPI-M, and CPI-ML-Liberation – they have not really gained much in terms of vote share over the three elections (even though they might have gained in terms of seats). In 2010, these three Left parties together had managed to get an adjusted vote share of 5.36%. In the 2014 Lok Sabha elections, this had declined to 3.78%. In the 2015 elections, the adjusted vote share has inched down further to 3.72%.


What could these numbers mean?

The trends represented by the numbers collected together in Table 1, 2 and 3 seem to indicate that the electoral rout of the BJP-led alliance in the 2015 Bihar Assembly elections is less the result of a massive shift of voters away from the right-wing alliance to the centrist, secular alliance of the RJD, JDU and INC, than an instance of the first-past-the-post electoral system showing its magic. The data for statewide and adjusted vote share shows that the BJP’s electoral support base is largely intact, with a slight decline in 2015. Despite this, the secular alliance managed to convert its vote share into a large number of seats because the anti-NDA votes were not divided, as had happened in the 2014 Lok Sabha elections. The coalition strategy and the implied vote transfer between RJD, JDU and the INC – highlighted by the significant increase in the adjusted vote share of each of these three parties between 2014 and 2015 –worked quite well. This has two implications.

First, the so-called electoral juggernaut of the BJP can be easily stopped if the main secular parties can come together in pre-poll alliances. Bihar’s 2015 assembly elections show that BJP’s divisive agenda cannot generate an electoral majority if the major secular forces are united. Despite the BJP’s attempts to generate communal polarization by invoking Dadri and Pakistan, that strategy could not be translated into a larger vote share. While this is very encouraging, it is also going to be an important factor – that secular parties need to consider – in the Assembly elections in Uttar Pradesh in 2017.

Second, the BJP’s ideological support base is much more robust. This is the upper-caste, upper & middle class, Hindu population. The real challenge for Left, democratic and secular forces is to engage that support base and erode it away through withering ideological critique. That is a far more challenging task, but no less pressing. The 2015 Bihar Assembly elections should not make Left forces complacent about that larger battle. It is far from won, and it needs our renewed attention, commitment and energy.

(I would like to thank the Sanhati editorial team for comments on an earlier version of this article.)

[1] An alternative measure that is often used is the vote share per seat contested, which is the ratio of the statewide vote share and the number of seats a party contests. Since the total number of seats differs between Vidhan Sabha and Lok Sabha elections, this measure would not be very useful for the purposes of the analysis in this article. That is why I use a different measure, the “adjusted” vote share, and define it as the ratio of the statewide vote share and the fraction of the seats that a party contests. Since fractions of seats contested are comparable across Vidhan Sabha and Lok Sabha elections, the comparison becomes more meaningful using the “adjusted” vote share.

[2] In Table 1, 2 and 3, I have included the three main left parties together as an alliance even if they did not always fight the elections as a coalition. The reason for clubbing them together is to study the evolution of the electoral support of these Left parties together since 2010.

No Comments »

Leave a comment