Graphic: Mohsin Shaikh/Mint
Graphic: Mohsin Shaikh/Mint

Different vote for state and centre?

The inference that people deliberately vote differently in the national and assembly elections is a bit trite

After some spirited sloganeering by political parties in the recent Bihar assembly polls, we were treated to some bubbly jargoneering by political commentators on the outcome of that election. The race to explain the Bihar electoral outcome with attractive one-liners yielded a bountiful harvest of narratives such as “agro-povertarianism", “index of opposition unity" and so on. Chief among them is this notion that voters in Bihar preferred Narendra Modi to be their prime minister and Nitish Kumar to be the chief minister, suggesting that the Bihar debacle was not a verdict against the PM.

Implicit in this notion is a presumption that voters choose differently for the state vis-à-vis the centre. That voters in India have different political preferences for their state leadership versus national leadership is an entrenched notion among both the political fraternity as well as their observers. Except, empirical evidence shows otherwise. Had elections to Bihar been held at the same time as the 2014 Lok Sabha election, there is a 77% chance that the Bharatiya Janata Party (BJP) under Modi could have won the assembly election in Bihar.

Using electoral data from 1999, we tested voter behaviour for differentiated preferences between the centre and state. All else being equal, when presented with two ballots at the same time—one for choosing their representative in the Lok Sabha and the other for their state legislature—did voters really demonstrate different choices for the two? There have been four Parliament elections since 1999. We chose all the states that had coinciding assembly elections with each of these Parliament elections and compared assembly segment-wise winners for Parliament and assembly. This meant that we tested 302 million voter choices across 2,600 constituencies in six states in 16 different elections between 1999 and 2014. In 77% of the constituencies, voters chose the same political party for both the state and centre. In other words, there is a 77% chance that the winning political party or alliance will win both the Lok Sabha and assembly elections in that state, when held simultaneously.

To boot, this trend of choosing the same party has been rising since 1999 and not declining, as is widely believed. Contrary to that popular notion that the average voter is acutely discerning of the difference between voting for their state and national representatives, there is very little actual evidence of it.

As our analysis shows, when presented with an option to choose different parties for the Lok Sabha and state, with all other things being equal, a vast majority of voters did not exercise that choice. It is then unclear what forms the basis for such punditry of astute voters voting differently for centre and state.

We also note that when assembly elections are held six months before or after the national election, voters chose the same party in only 61% of the cases, down from 77% when elections were simultaneous. When the state elections were held beyond six months of the national election, the chances of voters voting for the same party drop even further. The closer the elections are for the centre and state, the more likely voters are to vote for the same party in both. The farther they are, the less likely. There could be various reasons, primarily despondency with their last electoral choice, that drive voters to vote differently the next time. But the inference that voters deliberately vote differently for the centre and state is a bit trite.

To be sure, the fact that most voters do not choose different parties for the centre and state in simultaneous elections is not conclusive proof that they do not prefer different leadership. But there is no evidence that they do. Thus, while “Modi for PM, Nitish for CM" may be an eye-catching one-liner to explain the Bihar electoral outcome, like most electoral narratives dished out by our pundits, it is devoid of any evidence.

In the Bihar election, of the 243 assembly constituencies, 129 had a BJP and a Janata Dal (United) and a Rashtriya Janata Dal/Indian National Congress candidate in both the 2014 Lok Sabha election (assembly segment) and the 2015 state election, thus presenting a similar bouquet of choices to voters. In these constituencies, while 41% chose the BJP in 2014, only 36% voted for it in 2015, a loss of 5 percentage points. If Bihar assembly elections were held at the same time as the Lok Sabha election in 2014, it is highly likely (77% chance, as per this analysis) that the BJP would not have lost the 5 percentage points of votes and perhaps even won the state assembly.

So, the narrative that voters wanted Modi as their PM and Kumar as their CM to explain the BJP’s loss in Bihar is banal. That a section of BJP supporters turned despondent within 18 months and shifted loyalties is a more plausible explanation for the BJP’s loss than some innate voter preference for a Bihari as their state leader and a bahari as their national leader.

Praveen Chakravarty is a visiting fellow in political economy at IDFC Institute, and founding trustee, IndiaSpend.

This column was written with help from Rithika Kumar and Swapnil Bhandari, associates at IDFC Institute. Comments are welcome at