Comment | How much does anti-incumbency sentiment matter in Indian elections?

Representative image

Manas Chakravarty

Economists Ajit Karnik, Mala Lalvani and Manali Phatak have recently published a paper in the Economic and Political Weekly, titled ‘Determinants of Electoral Outcomes’. The authors analyse constituency-level data from the last ten parliamentary elections, focusing on constituencies where the incumbent has won the previous election by narrow margins.

What are the findings? The researchers find that, for elections held after 1998, there is a strong disadvantage for incumbents. However, the trend is not uniform across states. In states where the share of the rural population is higher than average, there is strong anti-incumbency sentiment. The same goes for states that are poorer than average.

The authors have divided states into two groups---poor, rural, less-well-educated, where there is a strong incumbency disadvantage and rich-urban-educated, where there is no discernible incumbency effect. In the former grouping are Assam, Bihar, Chhattisgarh, J&K, Jharkhand, Madhya Pradesh, Rajasthan, Odisha, Uttar Pradesh and Uttarakhand. Are these the states where incumbent BJP candidates are likely to lose?

Not so fast, the researchers have divided the states into other groups too. For example, they find that in the Hindi-speaking regions of Bihar, Chhattisgarh, Haryana, Himachal Pradesh, Jharkhand, Madhya Pradesh, Rajasthan, Uttarakhand, Uttar Pradesh and National Capital Region of Delhi, chances of winning decline by more than 50 percent for an incumbent. That will be terrible news for the BJP. What’s worse, is that, in the non-Hindi speaking states, there is a weak advantage for the incumbent.

But for the current government, here’s the silver lining---the study finds that ‘being a BJP candidate confers a huge advantage in terms of chances of winning elections.’ Being a Congress incumbent is a disadvantage.

So if the conclusions this paper arrives at are correct, the BJP should do badly in the Hindi belt, which are its strongholds, but this is qualified by the fact that incumbent BJP candidates do well. That’s a bit of a let-down, as it is not much use in forecasting.

The other takeaways from the paper are that incumbent parties in the non-Hindi belt should put up a good show. The Congress will fare poorly.

A similar study in 2016 (‘Anti-Incumbency, Parties, and Legislatures: Theory and Evidence from India’ by Alexander Lee of the University of Rochester)  had also arrived at the conclusion that incumbents suffer from a disadvantage. The paper found, ‘It appears that this bias is attached to both individuals and parties. When we focus on party vote share, the party of the incumbent…. is still less likely to win than they would had they lost the previous election, though the effect is smaller than for individuals. Candidates from incumbent parties where the incumbent does not choose to run have a much smaller, though still perceptible, disadvantage.’ It might make sense for the ruling party to redistribute tickets.

Like the EPW paper, Alexander Lee too found that ‘incumbent disadvantage does not hold in Tamil Nadu, Karnataka, Kerala and West Bengal’, but also said, ‘The rest of India tends to discriminate against incumbents, but does so to varying degrees, with incumbents faring the worst in Maharashtra, Assam and Haryana.’ There’s no mention of anti-incumbency sentiment being particularly strong in the Hindi speaking belt.

Some of these conclusions are difficult to digest. Kerala, for instance, has a history of changing its government every five years -- is this not anti-incumbency? Could the finding that anti-incumbency doesn’t work against the BJP candidates be the result of studying a period when the BJP has improved its tally strongly? And if being a Congress incumbent is a disadvantage, how did it win in 2009? And why should incumbents be weakest in the poorer and less literate states -- surely better educated and richer folk are less likely to be bamboozled by politicians and throw them out at regular intervals? The problem is there are so many moving parts in an election, especially elections in a country as large and as diverse as India, that it’s difficult to pin down the impact of one factor alone.

On the whole, however, anti-incumbency sentiment is a good thing, keeping politicians and parties in power on their toes and preventing them from getting delusions of grandeur.​