Interview: Substantial partisanship as well as large share of volatile voters drive election swings

Indian voters don’t care about ideology and are not particularly attached to parties. They “vote their caste” rather than cast their vote. They also rarely coordinate well in picking candidates, leading to tremendous volatility in election outcomes. Those are just some of the long-standing stereotypes that recent research on India has sought to upend.

In the fifth interview of the CASI Election Conversations 2024, CASI Consulting Editor Rohan Venkat speaks to Francesca R Jensenius, Professor of Political Science at the University of Oslo, about fresh data that reveals a larger stock of “partisans” among Indian voters than was previously understood, why the massive sizes of Indian constituencies skews our understanding of voter dynamics and what we still have to learn about how political parties mobilise voters.

Could you tell us about the questions that have animated your research?

Broadly speaking, I am interested in how state institutions shape society, how we can best design institutions and how they affect power dynamics between people. In the Indian context, I find it particularly interesting how formal institutions interact with informal institutions locally.

I am also interested in contributing to building bridges between debates in comparative politics and Indian politics. You will see in a lot of my work that I demonstrate that Indian voters are more similar to voters anywhere else in the world than what some stereotypes suggest.

Are you talking about a stereotype in the academic literature or more in conventional wisdom?

More in conventional wisdom. For example, in the media discourse, or how I see Indian politicians talk about voters – as if they just follow the instruction of their community leaders, that they are irrational, or are easily fooled. My experience from field work, and even from analysing survey data, is that Indian voters have strong opinions on many topics and that they respond positively to politicians doing a good job.

We can’t think of all the Indian voters as behaving the same way either. It is often thrown in as a cliché how large India is and how many voters there are, and we should not forget that this is really true. The Indian electorate is a large part of the people in the world and there are many different motivations for voting. Put simply, although it is definitely true that some people “don’t cast their vote, they vote their caste”, this is far from true about all voters.

In the vein of challenging stereotypes, you’ve had recent work out on the surprising resilience of partisanship in India. Tell us about that work.

That is joint work with one of the PhD candidates here at the University of Oslo, Ankita Barthwal. We find it interesting that studies of Indian politics have hardly talked about partisanship, while in the rest of the world, there is enormous literature.

Partisanship is the term we use for an emotional attachment to a party. That is, it is not the same as supporting a party or voting for a party, but the idea of feeling close to a party. Partisanship is considered an important stabilising factor of elections in the US, in many European countries, and across the world.

Yet in India, you hardly see any discussion of it. This is surprising because partisanship is thought to build over time as you see the same party competing for elections. And in India, while there are many parties that come and go, there are also some parties that are old and have been around for a long time.

We wanted to think more about partisanship in India – how prevalent is it, what does it mean to people and has it changed recently? The first thing we found really interesting is that contrary to what some people are saying now, when we look at survey data, we actually don’t see evidence of an increase in partisanship over time. There is a substantial share of the population that feels attached to a party and that share has been really similar for as long as there have been large election surveys. [Data are sourced from the NES surveys conducted by the Lokniti Programme of the Centre for the Study of Developing Societies]

Of course, the quality of the partisanship – what it means to people – or the motivations that have brought people to a party, that might have shifted. In a paper we’re working on based on data from Himachal Pradesh, we’re trying to tease out whether people’s partisanship is motivated more by ideologically or by expectations of group-based favoritism.

We find strong evidence that most of our sample in Himachal Pradesh seems to be much more motivated by ideological concerns than by community-party link. We also find evidence that respect for individual leaders matters. So, here too, like I mentioned for voting patterns, there seem to be several types of motivations for voters feeling close to parties.

What implications does this have for our understanding overall of the Indian voter?

Well, it implies that there is room for a lot more discussion of ideology in Indian politics and that all Indian voters cannot be mobilised the same way. Some might respond to signals about group-based favoritism or to what their community leaders think.

But then there’s another group of people who don’t necessarily respond to that, who have individual loyalties to a political party because of their convictions for what is best for the country irrespective of their own economic benefits. And those are not necessarily elite people. You also have other voters who are probably just really indifferent and don’t pay much attention at all. There are many things going on at the same time.

The group of voters who consider themselves partisan are particularly interesting, because when we look at literature on partisanship internationally, partisanship shapes how we process information. It becomes a set of tinted glasses that we see the world through.

It shapes how we understand politics in at least two different ways: by being a heuristic or a shortcut, where if the party says something, we’ll just say, “Oh yeah, that’s also my opinion because I’m generally with the party.”

The positive perspective on that is that ordinary citizens can easily form opinions on a broader set of political questions and thereby probably become more politically engaged, but it has the negative side that they might not be thinking critically about whether they actually agree with the party.

That links to the second effect that we know partisanship has, which is to make you much more likely to evaluate your party positively and be willing to forgive negative things you hear about the party. When you have partisan voters, they'll accept more of the party's behavior, whatever that behavior is.

In the context of polarisation, these effects of partisanship can be amplified by the fact that voters are often not exposed to other perspectives. That can result in a hostile political environment where people are not open to other political perspectives, and this can mean that we miss out on a lot of healthy democratic deliberation that can make everyone wiser.

This finding is different to the ideology argument made by Rahul Verma and Pradeep Chhibber, because it’s more about attachment to a party – which can shift its ideology?

We build on their work, but say something slightly different. An ideological commitment is about believing in some overarching longer-term goals for how a country should be run. You can have overlapping opinions with a party but that doesn’t necessarily mean that you feel an emotional attachment to the party, though it might be more likely that you develop such an attachment.

It is also possible to feel an attachment to a party without caring much about its ideology. What the literature on partisanship shows internationally though is that as parties shift their ideology, those who feel attached to the parties often follow along in these shifts.

The data that you have at least on this is quite stable. So, this idea of rising polarisation is not necessarily reflecting in higher partisanship?

In the data we have access to, we don’t see any evidence of an increase in the share of voters feeling close to a party. But because there’s a growing population, there is a higher number of partisans. We also see that the share of people feeling loyal to the Congress has gone down compared to those who feel loyal to the BJP now. So, there are more partisans toward the BJP now than before, as the party has been growing. But, we don’t have good measures of the quality, or the intensity, of partisan loyalties over time.

If asked very basic questions or whether you feel close to a party, that has been very stable over time. It is possible, however, that the intensity in partisan emotions have increased, but we don’t have good measures for this.

In our own survey we have tested this by looking at people responding to questions like: “do you get upset if someone speaks badly of the party?” These are questions that are used across the world to get at how strongly people feel about parties, but we don’t have such measures in India for other time periods. Therefore, we cannot say anything about how they have changed.

Are the share of partisans quite different in India versus elsewhere?

The Indian share of partisans is around 30%. That is quite low if you think historically in an international context, but not so much anymore.

In Europe, for example, partisanship has dropped dramatically, as the party system and what voters care about has shifted over time. There had been a de-alignment between voters and parties. In the US, partisanship is still very high, which is not so surprising given that voters basically have two big electoral options.

The idea of partisanship is also that it makes electoral politics stable. It creates these stocks of folks who can predictably generally vote in one or the other direction. But you’ve also worked and studied volatility elsewhere in India, particularly with voters switching parties and candidate hopping. How do those two play into each other?

Well, both of these things are true at the same time. We do see that partisans vote in a more stable manner than others. In fact, if they don’t vote for the party they feel close to it is because the party isn’t on the ballot. Often these voters don’t care much about who the candidate is, they care more about the party. Using the measure of partisanship we have discussed above, that’s 30% of the voters.

But then, in other work, I have focused on the group of people who are not loyal to any party or candidate and that is also a large chunk of the voters. If you look at the National Election Studies, a large share of voters say they have been in touch with a party around the time of the election campaign, but it’s far from all.

Many voters are, in fact, never in touch with parties during the campaign or otherwise. And if you ask whether they’ve seen the campaign on radio or TV or anything, they also say no. Many of these people also have limited or no education, meaning that it’s not likely they’re Googling political information either.

And so, a large part of the population has very limited connection to political parties or politics at all. If you look at voting patterns, the evidence suggests that these are the voters who really jump around from election to election.

Some of this volatility might be in response to handouts or shifting alliances, but many of them also seem to try to use their hope for political change, but often based on limited information. I do think it’s the case that a lot of voters keep trying new parties in the hope that they’ll get more in return. And that is what democratic accountability is about – you try to vote for someone who would give you something more.

So, we see massive vote swings in Indian elections. And I think that’s important to bear in mind when you look at the pattern now of BJP dominance at the national level, or even Congress dominance previously, that the vote shares that bring them to power are never very high.

I know the Opposition has harped on the BJP winning with just 37%. What is more important to keep in mind even when talking about 37% is that this is out of the 68% of voters who turned out to vote. And that percentage is out of eligible voters, which is far from the entire population. So, the share of the population that voted for the BJP in the 2019 elections was actually not that high, but because of the First Past the Post System, one can win the majority of seats with a small share of votes. And the more fragmented the vote is locally, the more you can win with small percentages.

This is not new. If you look at Indian elections historically, it’s never been the case that any party has gotten a massive vote share. The Congress party also won with less than 50% of the vote most of the time. It’s always been a fairly low share of the Indian population that have brought politicians to power.

If you look at the vote shares that go to parties other than Congress and the BJP, those also jump around. I’ve had a project where I looked at polling station level data. If you look at how much parties’ votes swing from one election to the next, it’s quite impressive. It’s not the case that villages generally vote together or the same from one election to the next.

Even if you look at data from the Uttar Pradesh State Assembly elections in 2022, you can see that it’s not the same places that supported BJP in high numbers from one election to the next. Parties generally can’t take their vote for granted in most areas. And so, they have to work hard to string together a winning coalition in any constituency. So, the volatility and fragmentation are present at the same time as you have a chunk of voters who are quite loyal, both things are happening at the same time.

Particularly as you mentioned in a First Past the Post system. In a more proportional system, it might have mattered less, but having that stock of partisans and some volatile voters might play out in different ways versus having a system where you don’t know where the spread of those committed voters are?

How they are spread out probably matters. And the size of constituencies also probably matters. Indian constituencies are huge. Most Indian constituencies are the size of small European countries. In the political science literature, we often talk about how it makes sense in the First Past the Post System for politicians and voters to coordinate around a few main contenders to not waste their votes.

The US is the stereotype where most people support the Democrats or Republicans, because if you start voting for smaller parties, then you might contribute to bringing to power a party you don’t like at all. You don’t want to do that. Therefore, you end up voting for the one among the top contenders that you find to be the best alternative, even if you would ideally prefer something else.

Now, in Indian elections, the coordination between parties and candidates has been limited. We generally see a high number of candidates standing for election in each constituency. From a rational perspective, that’s odd because it would make sense for candidates to build coalitions, share seats, and collaborate so that they could beat their main opponents. But for various reasons, coalition building in India has been complicated. Still, it would make sense for voters to try to coordinate around top contenders. And in India, some voters seem to do so, but many don’t. As a result, the vote is fragmented.

Interestingly, when we look at voting behavior polling-station level data, there is evidence of a fair bit of coordination. In fact, Indian voters seem to be coordinating much more than what we see at the constituency level, but often they’re not coordinating on the same two candidates across a constituency. There is evidence of people coordinating on the wrong top contenders. This might be because of regional stronghold or because it’s not so clear who’s going to win an election. I find that fascinating, how there might actually be more coordination in India than what we generally believe.

One methodological question before we get into that insight is about how much more quantitative work there is in India on political science of late. What are the challenges and difficulties of working with data on India?

When I started working on India, I worked mostly with qualitative data. I came to the quantitative study of Indian politics because I realised that I could hear stories of almost anything. India is so big and so diverse, and there are so many different lived realities that almost any story can be true, but I didn’t know how common things were.

Therefore, I got interested in quantitative analysis to get a better grasp how normal it is to behave in this way or think in that way. That’s where surveys are fantastic. But when I started studying Indian politics, I also realised that it was hard to access quantitative data on many of the things I was interested in. I was interested in how many women had been members of parliament over time, and I couldn’t find it. Of course, that data existed – there is a lot of great data on India – but it hadn’t been put together in a systematic way.

And so, me and many others at the same time, started trying to collect, collate, and digitise data related to the study of politics and society. How many women have been in power in the Lok Sabha? How many SC [Scheduled Caste] candidates run for re-election? Or how much time do state assemblies spend on legislative discussions? All of this data was available, but in different languages and often in formats that are hard to work with for statistical analysis. So, there are lots of challenges to answering even basic questions that we feel we should know about any political system.

To go back to the insights you got from the data, NES surveys suggest that there’s this section of Indians who are not visited often by parties. What do we understand about this segment? What does it tell us about how parties are mobilising voters?

I actually still think there’s a lot we don't know about how parties are mobilising people. There is interesting work being done right now and I’m trying to follow people in the field. We know from anthropological literature and some political science literature, where people have followed politicians, for example, that parties work very differently.

Parties don’t have a major presence locally. West Bengal and a few other places have been exceptions, where there seem to be party workers everywhere. But in other places, parties string together support from local leaders, they try to reach out to people who are not very loyal to their party. Or they have party workers who travel and do a whirlwind tour where they drive into a village and have a meeting there for half an hour and then head out. And of course, there’s no way you're going to have a deep engagement with every single voter in that village in a half hour's visit.

What I also hear frequently, when I talk to villagers, is that they are disappointed with politicians who came by once during the campaign and then they never saw them again. I remember interviewing one politician who also said that, “What people care about is that I'm present and I show up.” His take on it was that he could win an election by simply going to people’s weddings and funerals and showing that he cared. So, people have different mobilisation strategies.

What we have seen in these past few years is that the BJP have been extremely organised about how they run election campaigns; for example, using the Panna Prabharis [booth-level party workers] that cover a single page in the electoral rolls.

They also have an explicit strategy right now of trying to reach out to the poor, the young, and women, exactly to mobilise communities that have traditionally been less in touch with parties. I don’t know to what extent they are succeeding in reaching those groups right now. But I hope that there are field studies that can tell us whether they actually succeed in that ambitious plan of reaching out more broadly.

And then, one often hears now that parties reach more and more voters through WhatsApp or Koo or other social media. It is definitely true that they reach many people this way, but at the same time, we must remember that only about one-third of Indians are on social media at all, and even fewer are active users. So, although they reach many people, it is far from all.

Building on or responding to your work, or beyond, what interesting research have you seen taking the field forward over the last few years?

I think it’s a very interesting field of study because there was limited data access before. And as there's more and more new data coming out, we keep learning new things. The Indian politics field, from an academic perspective, is not a contentious field at all. There is so much to learn, and so many things to study, that it’s not a crowded field. Indian politics is too complex for anyone to tell the full story, so we need to learn a lot more to be able to piece together a coherent narrative.

As we have talked about, there are a number of contradictions. We have studies focusing on the importance of ideology while others argue that ideology plays almost no role. We see disempowered women who don’t care about politics, but also, there are studies of highly active women’s networks gaining increasing influence. People say that Indian parties are weakly organised, but then we suddenly see this amazing organisational capacity around election time where parties reach out very broadly. How common are each of these things? Here, we are still trying to fill in dots in a blank map and we need a lot more knowledge to understand more.

I find the academic community studying India to be very constructive and curious, working collaboratively toward better knowledge. And for this to continue, we need open dialogue on what is actually going on. And it is important for our research not to be politically motivated. Political science is susceptible to being politicised and made ideological because it is about politics and we are all members of our field of study. And so honest, open, diverse, deliberative, scientific discourse on what's going on politically is extremely important for understanding more.

One of the things you mentioned at the top was your interest in India in a comparative sense. Has making the case for India within a comparative view become easier over time?

When I started studying India, people said, “Oh, that's so narrow.” And I said, “It’s not narrow. I study five different topics on Indian politics. That's not narrow.” But it was often hard to convince people why it was interesting. I do think that’s changed.

As the field of the study of Indian politics has become really good at communicating and speaking to literatures and debates elsewhere, and as better and better data access has made it more statistically sophisticated, the interest in what's going on in the subfield of South Asian politics has increased dramatically. We see a lot more research coming out now than we previously did, and I feel one has to fight much less to explain why this is interesting.

What I often try to argue and show is that not only is it interesting in itself what a large section of the world is up to and what’s going on in countries outside of Europe and North America, but also I find that one can learn a lot theoretically and methodologically by looking at countries like India, which often challenges some of the assumptions in mainstream political science theory.

For example, Pavithra Suryanarayan and I have a paper where we try to show that it’s because of switching candidates that we don’t see evidence of economic voting in India. Once you look at the subset of places where you have the same partisan candidates running for election, you do see evidence of economic voting. And that’s a typical example of how literature on economic voting in the western world hasn’t really thought about the difference between party and candidate so much, because it hasn’t been so relevant to them.

And in the Indian context, when candidates switch parties, we can see some voters will stay loyal to the parties and some will stay loyal to the candidates. Voters can get confused about who to reward and punish. And that makes it clear that party-candidate loyalties and linkages is a core assumption in economic voting theory. In this way, the study of Indian politics can help clarify theoretical thinking elsewhere.

The same is true with things like state capacity. If you think of voters behaving rationally, and if you’re in a village where the state is hardly present and you get nothing, why would you bother to coordinate very much and care about who wins the election? You maybe don’t care so much. I think state capacity, for example, is a precondition for seeing the rational voting behavior that we should expect.

Going into the current elections, what about this current moment are you curious about? What are the big questions about the moment that you'd want to look at from a political science lens?

I find it really interesting, for example, this question that we talked about: How are parties actually out mobilising right now? How do they work? I am very curious what will happen to mobilisation and the workings of the Opposition if the BJP wins, as is now predicted. Will new Opposition alternatives spring up? I think there is a potential for it, given the local level volatility and the low vote shares that actually bring ruling parties to power.

I am also curious about what will happen to the BJP. From a political science perspective, we should expect it to have internal splits. It’s very hard to keep such a large party together. We might not see those splits as long as Modi is at the helm, but maybe later? Also, what happens if the BJP thinks for sure that they will win, do they mobilise people as much and in the same way? Still a lot of interesting developments that could take place.

Finally, do you have recommendations for those who are interested in this subject?

When I think of my bookshelf and what I'm very fond of, there's one book here called Marginal Players in Marginal Assemblies: The Indian MLA that really shaped how I thought about Indian politics. It’s an old one by Vir Chopra, before he went into Bollywood. It really shows how we talk so much about these politicians, but in the end, they feel they have very little power, and they spend very little time in the state assemblies and end up deciding very little. I find that really fascinating.

Another book I really love is Jen Bussell’s book, Clients and Constituents: Political Responsiveness in Patronage Democracies, on showing how many politicians work non-contingently. There’s a stereotype of politicians always giving goods to their supporters. And I have work myself showing that politicians tend to give more to their supporters, but she also finds evidence that they often don’t. And whether this is strategic or just because there are too many people contacting them, she shows really brilliantly at many levels how different politicians strategise in different ways. I also really appreciate Adam Auerbach and Tariq Thachil's work on the urban poor: Migrants and Machine Politics: How India's Urban Poor Seek Representation and Responsiveness.

Then, there is one book that I’m really in love with right now. It’s a very detailed study of how voters and vote mobilisation happen in Pakistan called Crafty Oligarchs, Savvy Voters; Democracy Under Inequality in Rural Pakistan by Shandana Khan Mohmand. It's about how voters try to negotiate benefits from different vote mobilisers. I think it's fascinating to reflect on how stuff in India and Pakistan can look really different, but also very similar.

And I know the list is getting long, but I have to add that I am also very excited about a book that's just come out, Accelerating India’s Development: A State-Led Roadmap for Effective Governance, by Karthik Muralidharan, which looks really useful and empirical and based on lots of data and information.

Francesca R Jensenius is a Professor of Political Science at the University of Oslo.

Rohan Venkat is the Consulting Editor for India in Transition and a CASI Spring 2024 Visiting Fellow.

The interview was first published in India in Transition, a publication of the Center for the Advanced Study of India, University of Pennsylvania.