Nickeled & Dimed

Penny for your thoughts?

We are accepting articles on our new email:

, ,

Political Dynamics in Bihar: Electoral Competition and Representation

Reproduced from Economic and Political Weekly Issue with the authors’ permission. The authors Yugank Goyal and Arun Kumar Kaushik teach at OP Jindal Global University, Sonipat. They can be reached at and respectively.


Abstract: Upcoming Assembly Elections in Bihar are expected to be unprecedented on several counts. Following a miserable defeat of their parties in Lok Sabha 2014 elections, two so-called arch enemies Lalu Yadav and Nitish Kumar have joined forces, after a 25 years hiatus. Despite that, there is considerable uncertainty in the air. In this article, departing from the usual qualitative claims on who will win, we study numbers of previous elections to draw an informed judgment on the type of representation and political competition Bihar can expect to face in its elections. In some ways, we predict the nature of elections, rather than the result of it.

1. Introduction

The 2015 Assembly Elections in Bihar has become a playground for the coalition of Janata Dal (U), Rashtriya Janata Dal (RJD), their allies with Indian National Congress and National Congress Party on one hand, and Bharatiya Janata Party (BJP) and its allies on the other. The Nitish Kumar-led former alliance against BJP-led latter is one of its own kinds of election battles in the history of the state for two interesting reasons. Firstly, the eighteen years old partnership between JD (U) and BJP, which saw them through three Assembly elections (of which last two were favourable) is broken this time. Secondly, Nitish (JD (U)) and Lalu (RJD) have joined forces after being arch enemies for twenty five years. Election pundits say that given Nitish’s empty promises in last five years, coupled with internal rivalry between the foot soldiers of JD (U) and RJD, the elections may actually see a replication of 2014 Lok Sabha elections. On the other hand, people have also observed the power of unified Other Backward Castes (OBCs) through union of Yadavs (Lalu) and Koris and Kumris (Nitish) that will greatly diminish BJP’s vote share.

In this article, we tiptoe around the temptation to offer predictions as to who is expected to win. However, we build a possible scenario for (a) how (electorally) competitive the elections are going to be, and (b) how representative the government will be, no matter who comes to power (none of these will depend on which party actually comes to power). This way, our analysis departs from standard narratives which hinges on relying on qualitative rhetoric.

2. Election Dynamics in Bihar Assembly

Except for a brief period of Janata Party’s and few interspersed President’s Rule, Bihar was mostly ruled by unchallenged Congress’ Upper Caste Chief Ministers (CM) for forty years after independence. The Congress’ upper caste land owners secured votes from landless labourers through a patron-client relationship (Kumar et al, 2008). This continued until the 1960s when agrarian relations began changing, and upper OBCs (Yadavs, Koeris and Kurmis) began articulating their consolidation (Junuzzi, 1974). In 1967, the collective opposition against Congress led Bihar to elect the first non-Congress CM. It didn’t last long due to internal politics leading to elections in 1969 which again led to a hung Assembly, and paved the way for power switching between Congress and its opposition, interspersed with President’s rule until 1972, when Congress gained victory taking advantage of a fragmented opposition. Post emergency, Janata Party swept the elections with a seat share of approximately 66%.

That didn’t last long and Congress ruled the state from 1980 until 1989 although with fragmented leadership. However, the release of Mandal Commission Report in 1980 (implemented in 1989) coupled with already brewing caste consciousness began affecting Bihar’s politics deeply, with lower castes mobilization marking a clear shift in political dynamics (Jefferlot, 2003; Yadav, 1999). Bhagalpur riots in 1989 and Babri demolition in 1992 had already cleared Bihar off Congress. The 1990s witnessed a political transformation, with backward castes, Muslims and Dalits united under JD, while upper castes divided between BJP and Congress (Kumar et al, 2008). Under Lalu Prasad Yadav, JD swept the elections in 1990. However, due to internal rivalry between lower castes factions, JD got divided with Nitish Kumar siphoning off Samata Party from it, and Lalu Yadav retaining the rather sizable remains. Lalu managed to win the 1995 elections, rising as the new leader of the state. However, the fodder scam led to his resignation from JD and he formed his own RJD, installing his wife Rabri Devi as CM in 1997. RJD won the elections in 2000 Assembly Elections as well, although with poor performance compared to 1995.

Lalu, through bringing hitherto marginalized groups into the political mainstream and ensuring communal harmony, was able to secure election seats. However, law and order, health and education, and civic amenities were deplorably low during his fifteen years of rule. There was a disappointment amongst non-Yadav castes on the ‘yadavization’ of political elites (Hauser, 1997). In the meanwhile, Nitish had forged an alliance with BJP (late 1990s), and was therefore relying on upper caste votes in addition to his own Koeris and Kumris. In 2005, his party JD (U) with BJP emerged victorious. 2010 Assembly Elections saw a repeat, with a stellar performance, where Nitish led coalition won 206 of the 243 constituencies.

This narrative breaks down in the 2015 Assembly Elections given how coalitions have broken and new ones formed. Note that during the last twenty five years – the period which saw caste politics reaffirming in the Assembly election verdict – has been interspersed several times with President’s rule owing to unclear mandate. With that in mind, we engage with Bihar’s political dynamics, suggesting the type of outcomes in upcoming elections.

3. Political Measures

We look at the seat share and vote share of the parties in the Assembly Elections in Bihar. Two of the most popular indices, which we will use in this analysis are that of disproportional representativeness and electoral competition (Kaushik and Pal, 2012). In this way, we observe the undercurrents of election shifts in Bihar in rather nuanced estimation of results.

Disproportional index (Gallagher, 1991) is a measure of how representative a government is, i.e. how much does it represent the preference of the voters. This is measured using the difference in seat share and vote share, because a perfectly representative government will be the one where seat share and vote share are the same. In the index, sj denotes the number of seats of political party j in the elected body, vj denotes the number of votes received by the j-th political party and P denotes the set of political parties contested.

Formula 1

Clearly, DISPR takes the value zero, if the political system is perfectly representative. On the other extreme, if there is dictatorship, DISPR takes value 100. Therefore, DISPR∈[0,100]. It is easy to observe that higher value of DISP indicates lower representation of people’s preferences.

Electoral competition measures the effective number of parties contesting in the elections. It is not merely a measure of number of parties contesting because not all parties are equally strong. Hence, increases in the number of political parties contested does not necessarily mean an increase in electoral competition. However, the seat share or vote share is a better proxy for the size of the strength of that political party. Thus, the effective number of political parties (ENP), which is defined as follows, can be considered as an appropriate measure of electoral competition (Laakso and Taagepera 1979).

Formula 2

ENP can be interpreted as the number of hypothetical equal sized parties that would have the same total effect on electoral competition as have the actual parties of unequal size. It is evident that the lowest possible value of ENP is one, when only one party wins all the seats. On the other extreme, if all the parties win an equal number of seats, ENP is equal to the number of parties. Clearly a higher value of ENP indicates more intense competition among political parties in election. ENP is therefore a reliable measure of electoral competition (see for example, Chhibber and Nooruddin 2004; Bortolotti and Pinotti 2008).

When the ENP is estimated using vote share (instead of seat share), we denote it by ENPv. Note that ENP and ENPv need not be the same, unless vote share and seat share of all parties are the same. However, these two aspects should move in the same direction.

Formula 3


Based on these parameters, we construct Table 1. In the next section, we weave the political history of Bihar with these indices.

Table 1

4. Analysing Bihar’s Assembly Elections

We engage with suggested parameters to understand Bihar’s electorate. Our focus will be on Assembly elections held in 1990 and later. This is because Assembly Elections held in 1990 broke the narrative from the previous elections, and also because politics and leadership in successive elections became a path dependent imprint of the same.

4.1 From Independence until 1990

We hinge our analysis on Table 1. We notice the electoral competition increasing in 1967 and 1990. Indeed, 1960s were marked by changed in agrarian relations and upper backward castes (Yadavs, Koeries, Kumris) emerged as new land owning groups, and ready to articulate their political vision (Kumar et al, 2008). Congress suffered a major setback in 1967, whose seats were absorbed by other parties. The story of 1990s is evident in the emergence of consolidated backward caste parties, which were also strong in popularity, thus leading to high political competition. Interestingly, both in 1967 and 1990, disproportional index had low values compared to any other Assembly Election of Bihar. This shows a high level of representation. In other words, governments in 1967 and 1990 reflected people’s preferences more truly. One can surmise that those elections in Bihar where backward castes did not (or could not) articulate their voice, were rather unrepresentative of Bihar’s populace. Even in 1977, when post-emergency anti-Congress sentiment lead JD to sweep Bihar elections in an understandably one of the lowest values of political competition, the representation value is one of the highest.

4.2 1990 onward

Th 1980s marked a changing political rubric of the state, exhibiting the dormant sentiments in Assembly elections in 1990. Unsurprisingly, 1990 was characterized by low electoral competition (high ENP) and highly representative government (low DISP). In 1995, the statistic reversed. This cannot be explained by Nitish Kumar’s exodus from JD (taking the share of Koeris and Kurmis away from the consolidated backward caste votes) since his party got only 7 seats. Nor did the seat shares of BJP and Communist Party of India (CPI) suffer any major setback. The real change was Congress’ withering away, while JD’s strengthening power.[1] Hence, effectively, there were fewer parties in 1995. But the rate at which DISP has increased during this period, is unprecedented (see Figure 1).[2] This only shows that Lalu’s government even though had secured 167 seats (as compared to 122 in 1990) was not reflecting the will of the people (increase in its vote share was very small). Clearly, this indicated a dissatisfied electorate.

Graph 1

Surprisingly, in 2000, even though the electoral competition increased, explained by further consolidation of Nitish Kumar’s Koeri-Kumri base (which was weaker earlier) and Lalu’s (which was strong earlier) fodder scam hit to his newly formed RJD, the value of DISP has dropped dramatically, going lower than its 1990 levels. This massive drop in DISP indicates that seat share and vote share did not diverge as radically from each other as one would have expected. In some ways, wherever Lalu (and his allies) won, he (they) won by high mandate. Government was indeed representative. Notice that this is interestingly a rather different conclusion than most political commentaries, which argue that these results were unusual. Indeed, one can make sense of this mandate by understanding that 2000 was characterized by strong sentiments. At an aggregate level, victories were genuine and consistent – desired parties were really desired. Hence, a declining number of seats for RJD truly signaled diminishing of Lalu’s support base.

This became a reality in 2005. Most political pundits have argued that the 2005 change in political leadership from Lalu to Nitish was a major discontinuity in Bihar’s political narrative. While this may be right in terms of leadership, but the number-story of elections in Bihar do not suggest any unusual break. When we perform the calculations based on two major coalitions that had been formed before elections – Nitish-led JD (U) with BJP, and Lalu-led RJD with Congress – we do not observe any surprise. The political competition had considerably reduced, naturally. Disproportionality had increased expectedly, indicating big difference between seat and vote share. This means that more people must have been dissatisfied with the verdict. For most political commentators, this election was a clear divide between JD (U)/BJP, who were banking on backward castes (Koeris-Kumris) and upper caste votes; and RJD/Congress relying on backward castes (Yadavs, primarily) and Muslims.

We do not think this divide could have been as neat. In 2000, Lalu secured approximately 22% more seats, and 21% more votes than Nitish. In 2005, with 22% more votes, Nitish’s alliance secured more than double the number of seats as compared to Lalu’s. This shows the extent to which the Assembly’s representativeness got diluted in 2005. From our perspective, it was a turnaround of JD (U)/BJP alliance, with both securing significant number of seats owing to deplorable situation Bihar had been in, in last fifteen years. But it was not at all a grand sweep.

The 2010 Assembly Elections follow a similar story. The much touted victory of 206 out of 243 constituencies by Nitish-led alliance is a story of misplaced euphoria, concealing vote share. With a mere 50% more votes than Lalu-led alliance, Nitish (and BJP) secured more than seven times the number of seats. The victory margin has to be very thin. This is reflected in the unprecedented high level of DISP and corresponding low value of ENP. One can reconcile this observation by looking at the ENPv. Note that in 2010, there is hardly any reduction in electoral competition based on votes (ENPv). This means that for all practical purposes, Nitish’s win does not reflect his popularity. A DISP index of 35.4 shows that the 2010-15 Assembly was hardly representative of people’s will. Since the number of political parties contesting elections has historically been the highest in 2010 (90 as compared to approximately 50 in previous two), they have – without winning any seat – been only able to absorb a large amount of votes, making less of them available to other parties. No wonder the difference between seat share and vote share is highest in 2010.


5. The Upcoming Assembly Elections

Using the above mentioned frame, we offer a fresh perspective to Bihar’s Assembly Elections 2015. Two events in upcoming elections stand apart. Firstly, with Lalu and Nitish joining hands, we see a repeat of 1990 coalition matrix. Secondly, the coalition between JD (U) and BJP is broken, unlike in the last three elections. We view these developments in light of estimates emanating out of seat and vote share variables.

ENP in 1990 was relatively high because of strong presence of Congress, BJP, CPI, Jharkhand Mukti Morcha (JMM) and indeed, JD. JMM has taken refuge in now separate Jharkhand and CPI is, by various analyses, out of the battlefield. This essentially leaves erstwhile JD (in a new coalesced avatar with Congress) and BJP with its allies (NDA). Hence, most definitively, the value of ENP should be fairly low. This would be the 2015 elections departure from the 1990 election.

More importantly, the difference between seat and vote share (ENP-ENPv) was very low in 1990 indicating appropriate representation in the Assembly. Do we see a similar pattern this time? In 2010, the winning margin was very thin, which is exhibited in historically the highest value of (ENP-ENPv). We see this trend to suffer some change in the changing dynamics of in caste calculus. This, in turn, means a non-representative Assembly. If the results of 2010 are of any help, we know that this is indeed possible.

The 2010 Assembly Elections in Bihar had the lowest level of electoral competition since independence, if one measures it using seat share. Interestingly, if one uses vote share, it was the biggest electoral competition. This discrepancy testifies to the fact that a large (and crucial) number of voters were in a swing mood. Low margins of victory means winners need to perform well, since a shift of a small group of voters is enough to topple them in the next elections. In some ways, the 2010 elections were highly deceptive – they had very clear majority, and yet an uncomfortably high value of |ENP-ENPv|. For us, the latter exhibits a true picture.

We expect that the value of |ENP-ENPv|  will reduce in 2015. One of the reasons for the high difference between seat share and vote share in 2010 was division of backward caste electorate (since 2000), coupled with lack of competition based on seat share. This may not happen in 2015 given the leaders of OBCs have joined forces, and are not competing against each other. Further, we have also noted that Bihar often replicated results of the preceding year’s Parliament elections (Lok Sabha) in Bihar, in terms of ENP and DISP , even though the election results were different. In other words, the nature of Bihar Lok Sabha and Vidhan Sabha elections have followed similar pattern even if the results are different. If this has some merit, we may expect BJP’s victory to yield some positive dividends for it in Assembly elections in November. Consequently, competition may increase, and representation may become better.

6. Conclusion

All of this makes us believe that Bihar Assembly Elections 2015 will witness better competition and a more representative government, regardless of the party coming to power, as compared to 2010. Yet, given the analogies flowing from 1990 and 2010, the level of representation may not reach 2000 levels. And if Nitish loses, it should not come as a surprise because his victory in the previous elections was built on a sizable number of voters inclined to swing.

Implications of this outcome could be very valuable. For any party which has come to power with high levels of disproportionality, the winning tenure becomes an acid test of its promises. Nitish may be smart in forging a pre-poll alliance with Lalu given how unimpressive his work has been. If the lower cadres of JD (U) and RJD do not bring their animosity in the ground, Nitish’s incompetence could be neutralized in elections. However, if BJP wins against the giant caste move of Nitish-Lalu alliance, it will not only – perhaps ludicrously – confirm the hypothesis that Bihar Assembly often replicates preceding Lok Sabha results, but also prove that an incompetent party occupying the Assembly with high disproportionality figure is sure to be out of business, whatever the electorate.


[1] Congress’ seat share reduced from 71 (out of 324) seats in 1990 to 29 in 1995, while JD’s increased from 122 to 167 during the same period. Also note that Jharhand Mukti Morcha also saw its’ seats halving between 1990 and 1995.

[2] DISP also increased at an alarming rate between 1967 and 1970. We are not sure why would that be the case, but we hypothesize that the political turmoil going on between 1967-1970, leading to installation of nine CMs and three President’s Rule resulted in voters emerging strongly in favour of one party – Congress – fueled by perhaps Indira Gandhi’s garibi hatao campaign.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: