After the first phase of polling on April 11, the initial estimates suggested that the turnout may have dropped a few percentage points in comparison to 2014. Many commentators went overboard in making claims about the effect of this drop. Fast forward to phase 3. When the voting took place on April 23, and there were reports of highest-ever turnout in states such as Gujarat, Kerala and Karnataka, the incumbent as well as the opposition claimed that the high turnout would favour them.
Some myths die hard; especially the ones in which logic could be used in either direction to support the claimant’s case. If turnout increases, the claim goes: ‘People are unhappy with the government and that’s why they are turning out in larger numbers to vote to express their anger’. If the turnout decreases, the claim is that ‘people are unhappy but as the opposition has nothing concrete to offer they have decided not to turn out’.
What are the effects of an increase or decrease in voter turnout on the re-election chances of an incumbent party in India? Several scholarly analyses of both parliamentary elections, as well as state assembly elections, have consistently shown that there is no relationship between turnout and incumbency. These findings are robust even after controlling several confounding factors and have been tested using sophisticated statistical tools. Yet turnout hypothesis is once again all over the news media, but as usual, it is missing the nuances.
Ideally, if there are no wild fluctuations, turnout rates do not invite any public scrutiny. However, turnouts matter for those in the business of forecasting elections, especially if there is a significant increase or decrease in the overall level (Type 1) or if the composition of who is turning out to vote undergoes a sea change (Type 2). In 2014, an increase in turnout at the constituency level was directly correlated with BJP winning that seat. Similarly, in the 1990s, though the turnout level had remained the same, increased electoral participation by marginalised communities led to the ‘second democratic upsurge’. Yogendra Yadav coined this phrase in a very influential paper which argued, comparing survey data from the 1970s and ’90s, that participation in election-related activities has substantially increased among the lower castes, religious minorities, and women. This had significantly affected the nature of political competition in the 1990s.
The data presented in the graph below shows that greater the increase in turnout in 2014, higher the probability of BJP winning the seat. The NDA won 67 out of the 70 seats (96%) where the voter turnout increased by more than 15 percentage points. In 145 seats, the increase in turnout was between 10 and 15 percentage points; the NDA won 125 of those seats with a success rate of 86%. The turnout increase in 267 seats was less than 10 percentage points, and the NDA won 123 of these seats (46%). Finally, the NDA’s strike rate was 34% in constituencies with no or negative change in turnout.
While we can safely rule out Type 1 scenario (significant increase or decrease in turnout) playing a role in 2019 verdict, Type 2 (the composition of who is turning out) is going to be crucial. Even in 2014, despite an increase in turnout across communities, the change was greater among the upper segments of the caste-class matrix. The poor and Muslims turned out less. And this was reflected in BJP’s ability to win seats with huge margins.
The pre-poll survey conducted by Lokniti-CSDS in March 2019 indicated an interesting pattern. While two in every three respondents interviewed for this survey said that they are highly likely to vote in the election, one in every three was less enthusiastic about turning out. The data in the second graph is unambiguous about the advantage the NDA has among the voters who are more likely to turn out. Yogendra Yadav describes this as the phenomenon of “active pro-incumbency” vs “passive anti-incumbency”. In his words, while Modi backers are enthusiastic about voting, those who are unhappy with the Modi regime are unenthusiastic about voting this time. The pre-poll data suggests that poor, Muslims, Scheduled Castes and Scheduled Tribes are less likely to turn out to vote.
Contrary to all the chatter in the first two phases, that the decline in turnout rates (which turned out to be not true) may harm the BJP, all the evidence seems to be indicating bad news for the opposition parties. Moreover, the cadre-based parties like the BJP (and some regional parties like the DMK in Tamil Nadu) are well equipped to mobilising, thus even in low-turnout scenarios they do well as they bring their voters to the polling booth. Parties that are poorly organised on the ground, such as the Congress, face challenges in getting their voters to the booth.
If the turnout plays any role in the verdict on May 23, it would be because of those who did not turn out to vote. The elections in 2004 indicated that the reach of the ‘second democratic upsurge’ was on the wane — participation among the marginalised had stagnated. And 2014 pointed to a process of reversal — decline in participation rates among the marginalised. If 2019 consolidates this trend, we can bid farewell to the promise of the deepening of democracy in India.
Verma is a Fellow at the Centre for Policy Research (CPR)