Anant Sudarshan on Market Solutions to Air Pollution, Energy Policy, and Ecological Disruption

SHRUTI RAJAGOPALAN: Welcome to Ideas of India, where we examine the academic ideas that can propel India forward. My name is Shruti Rajagopalan, and I am a senior research fellow at the Mercatus Center at George Mason University. 

Today my guest is Anant Sudarshan, an Associate Professor at the Department of Economics at the University of Warwick and a Senior Fellow at the Energy Policy Institute at the University of Chicago (EPIC). We talked about air pollution in India, crop burning, subsidizing electricity, depleting ground water, the impact of the collapse of keystone species and much more. 

For a full transcript of this conversation, including helpful links of all the references mentioned, click the link in the show notes or visit mercatus.org/podcasts. 

Hi Anant, welcome to the show. It’s a pleasure to see you here.

ANANT SUDARSHAN: Hi, Shruti. Thank you for the invitation. It’s very nice to be here.

RAJAGOPALAN: How long have I known you? I feel like I’ve known you before both of us were economists, but I can’t quite pin down a time or a date or even a vague decade to it.

SUDARSHAN: I do think you might have been a debate adjudicator at a Delhi University debate that I was at, at some point.

RAJAGOPALAN: Oh, wow. This goes back longer than I had imagined.

SUDARSHAN: Yes.

RAJAGOPALAN: We have some common friends some of whom have appeared on the podcast. That’s cool. It’s a thrill to have you here. One, I’ve known you a long time. I have been following your work for, I feel like, forever. Most of what I know about air pollution and climate change and its economic and noneconomic consequences in India is really through the work that you and your institute and team have done over the last decade or maybe even longer than that.

Air Pollution in India

I want to pick your brain on all of those themes. First, can you just start us off on air pollution and specifically air pollution in North India? What is going on? Is it as bad as the numbers suggest? At this point, the AQI numbers look made up. It doesn’t even seem like this is within the realm of possibility—why things are so bad in North India. And both of us are from Delhi. It didn’t used to be this bad when we were growing up, but something seems to have changed and things seem to have gotten a lot worse. If you can start us off with that, then we can jump into your research on it.

SUDARSHAN: It’s a good idea to level set first. What’s happening in North India—it wouldn’t be an exaggeration to say that it is the worst air pollution in the history of humanity in terms of sustained exposure in large population centers. Air pollution rises with industrial activity and economic activity, and these are the most polluted places on the planet. It’s a very serious problem, and it’s unfortunate that it’s been getting worse, not better, over time.

We are often told that there’s this environmental Kuznets curve where countries get richer and richer and then they get cleaner. It’s not clear where that bend is if this story is true, but it’s reached levels of pollution that are an active health hazard. Yes, it’s gotten worse over time. We don’t really have very much data. If you went back even one decade—we’re now old enough that when we were children was more than one decade ago—it’s difficult to make a concrete statement about how much worse it’s gotten.

It’s an incredibly densely populated part of the world. It also has certain meteorological factors that make air pollution worse. There are what are known as thermal inversions, which basically means that cooler polluted air settles at the bottom, near the ground. That does not help. 

You have a whole bunch of sources of pollution, many of which come from economic activity that is badly regulated: small-scale industry, vehicles that are often polluting too much and so on. And some of which, at certain times, comes from activities such as farmers burning crops, which we hear about a lot, which is a single and time-specific source, but which is, I’d say, uniquely Indian in this context.

RAJAGOPALAN: Let’s pick this up bit by bit. One of the things you alluded to is in most places in the world, there’s industrial pollution, and there’s this Kuznets curve, right? As societies get richer, they sort of pollute more. China is the classic example in the recent past. Once they cross a particular threshold, they start caring about air pollution. They also develop their state capacity, I presume, and have the revenues to really enforce their own environmental regulation. Then pollution comes down. And, of course, technology is always going to be innovating.

In Delhi, something weird seems to have happened. It seems like most of the pollution in Delhi is not from your standard industrial sources because Delhi seems to have deindustrialized in an important way. This is, of course, thanks to many Supreme Court cases. It’s also because real estate just became more valuable, and more and more of the industries moved out. Here, Delhi seems to be a little different from, say, Kanpur or the outskirts of Lucknow where you still have a lot of small-scale, extremely inefficient polluting industrial activity that’s taking place. 

First up, is this right or wrong, or is this more about different kinds of pollution depending on whether it’s PM2.5, PM10? What’s going on with Delhi’s industrial pollution activity?

Causes of Pollution in Delhi

SUDARSHAN: To some degree, I think Delhi is a piece of land around which we drew a border and decided to call things on one side of that imaginary line Delhi and on the other side, not. In some sense, since pollution does not care about those boundaries, that’s not necessarily, I think, the best way to look at it. 

Now, Delhi did push out a lot of industry, so a lot of them have gone just outside Delhi’s borders. There’s a lot of cross-border pollution from there. There are diesel generators in the greater national capital region, not necessarily within Delhi’s borders. Again, pollution doesn’t care.

I think one of the policy challenges, which, in my experience, tends to frustrate policymakers is there’s no one big single cause that they can knock out. Transportation—it’s an incredibly densely populated city, a lot of vehicles on the road, many of them quite old—is about 15% or 20%. I mean, these numbers vary depending on who’s doing a study like that. Then, the rest is road dust, construction activity, these cross-border sources. I think there’s a whole bunch of causes, and I don’t think there’s anything particularly special about Delhi other than, if you like, the political economy of if you close industry in one geography, they may move just outside.

Addressing Crop Burning as a Source of Pollution

RAJAGOPALAN: Now, another part of it is India is a little bit unique, especially the northern part of India, because the worst pollution comes from the most unproductive or inefficient part of the economy, which is agriculture. I know it’s very fashionable to blame crop burning, but they’ve been doing this for a very long time, and it’s only a couple of weeks in a year, but that’s what gets all the attention.

Can you walk me through what’s happening with crop burning? How much of it is just the actual burning of crop? How much of it is this meteorological thing where places like Delhi, very similar to other places—LA has had this problem, London’s had this problem—where the particulate matter just seems to settle in at particular times of the year? Then there are some wind trajectories, I believe, which also make it blow northwest to northeast, hurting the entire Gangetic belt. What is going on with crop burning, and how much of Delhi’s pollution, especially in the winter, can we attribute to that?

SUDARSHAN: I think crop burning is an interesting problem. There’s a particular time of the year, which is about now, November, when you have farmers in Punjab and Haryana and even in parts of Pakistan who are burning the residue from the paddy harvest in order to prepare fields for the next crop plantation. This creates an incredible spike in air pollution levels, which, if you average it out over the year, is less than 10% of the total. Which is why some people will often tell you, “Okay, crop burning is not the real problem.” But the impacts of air pollution are not just about what’s the average level, but if you’re becoming a complete gas chamber for three weeks in the year, you’re going to cause a lot of damage.

That is what is happening with the burning of this crop residue. You’re absolutely correct. It so happens that the winds carry this smoke from the northwest flowing eastwards across the Gangetic plain, which means the damages done by burning crops in Punjab are largely borne by states other than Punjab, which is one reason why it then gets harder from a political economy point of view to solve.

Yes, meteorology plays a role in thermal inversions. The fact that it happens in winter means that it tends to stay closer to ground level, so this pollution punches above its weight. The bottom line is there’s this massive externality being imposed on the rest of the population from this one portion of the economy basically using a shortcut. There are other ways to clear your field. They’re more expensive. This is the quickest and cheapest.

RAJAGOPALAN: One of the things that I learned—and this is from Ashok Gulati’s work, and I think Pranay Kotasthane and his co-author had written about it in their policy book—[is] the main reason they burn the crop residue is that the window between the kharif crop, which is the rice or the paddy crop, and the rabi crop, which is wheat, is really short.

One of the reasons it is shortened is, I believe, because Punjab has this particular law called the Prevention of Subsoil Act or something thereabouts that prohibits paddy plantation before mid-June because paddy requires a lot of water. It takes about 50% more water if it’s planted in April versus if it’s planted in June.

There is a host of reasons which don’t have that much to do with industrialization or vehicles or air pollution. As someone who is an economist or an environmental economist studying these questions, first of all, is what I’m saying correct? Second, how do you tease this out? The kind of work you do is, you’re trying to figure out some cost-benefit.

Of course, there is a cost to pollution, but there’s also benefit, which is all this economic and industrial activity and our ability to get in a car or get in a bus and drop off the kids at school. You’ve got to trade these things off, but what’s happening in Punjab seems like this really bizarre thing that makes no economic sense whatsoever.

SUDARSHAN: Yes. I think that’s fair. I think there’s two ways to look at it. One is that what’s happening in Punjab is, as you said, a unique source, which is not the standard economic activity that we think of, which creates this kind of horror scenario in a particular part of India at a particular time. If you were to zoom out and look at other Indian cities at other times of the year, they are still historically polluted. In some sense, you could argue that the farmers in Punjab are a little bit of a distraction in the sense that the more time we spend talking about them, the less we are focusing on the fact that industrial regulation is broken and so on. 

RAJAGOPALAN: Fair.

SUDARSHAN: Now, but that said, you have the capital city, and the capital city becoming a gas chamber for a few weeks in the year—it’s a problem that’s important in and of itself. I think you’ve hit the nail on the head. I have a fairly strong view on this that I am going to try and put the numbers down at some point, but agriculture in Punjab, if you think about this, it’s a very water-hungry crop, paddy.

The reason why the planting season was short is that this law exists in order to move as much of the paddy-growing season into the monsoon rainy season as possible. You’re doing that because, otherwise, paddy is decimating groundwater. This is an unintended consequence.

You are left with this economic activity that has these massive environmental externalities, a problem of cross-border pollution, and so it’s incident on someone else either in space, as in the case of air pollution, or time where the groundwater shortage is going to bite some years down the road, so you can kick the can down the road. Yes, it’s a unique problem, but it’s just horrendously bad economics. It just makes no sense.

RAJAGOPALAN: That’s what bothers me so much about it. Actually, I don’t just focus on it because it’s Delhi. That entire belt which is affected by crop burning actually has about 350, 400 million people living in that belt. This is also very bad for Punjabi farmers and their families because they are the most affected by crop burning in their immediate vicinity. It’s just that there seems to be no clear economic upside to this entire thing that, I think, irritates me so much about the crop burning. 

Regulatory Frameworks for Pollution in India

Now, coming to the industrial part, you have this really nice survey paper where you talk about persistent problems with the air pollution crisis. This is with Greenstone, Harish and Pande. It’s very readable. I encourage everyone to read it. You pin down the pollution crisis as a consequence of the regulatory framework, which instead of setting broad standards and enforcing them very clearly, focuses on command and control, a highly micromanaged framework, which is not just true for industrial pollution laws. This is true across the board in India. This is just how we do things except, here, it has very real consequences of turning Indian cities into terrible places to live.

Can you walk us through how industrial pollution regulation actually works in India? Give us some sense of that, and then we can talk about some of the consequences that you address in the paper.

SUDARSHAN: Sure. Before I do that, the questions you’ve asked underscore the fact that I think the environmental problems we face are about things that, as people interested in development, we’re constantly talking about in other contexts. The Punjab problem is a political economy problem. How do governments deal with their voters? This one is a state-capacity problem, I think, at the end of it.

The way, for example, most Indian industry is regulated pretty much across the country is through the use of, essentially, standards and criminal enforcement. Say, for example, to simplify things, that here is a hard limit on how much pollution, say, 150 micrograms per meter cube, that this plant can emit. If you go above this level, you are liable for criminal penalties. You can be imprisoned. Your plant can be shut down. All sorts of very stringent things. If you are below that level, you are okay. It’s not uniquely Indian.

Prior to the Clean Air Act amendments of 1990, and still in many parts of the developed world, you have a similar kind of regulatory structure, so it’s not something that India dreamed up. What I think is interesting is that you see in India this form of regulation—perhaps this is a little bit hyperbolic, but not so much—it’s essentially completely collapsed across large parts of the country. By completely collapsed, I mean that a very significant share of plants is in violation of these norms and everybody knows they’re in violation of these norms, so it’s one of these open secrets. There’s this really interesting question then of, why is this the case?

I think a lot of my research and other people who have done work on this are thinking about: Why is it that this regulation doesn’t work? Why is it that it seems to have worked much better in, say, the United States and is not working in India? That’s where I think of the regulatory design and the state capacity problem as being the point of entry, if you like, in thinking about air pollution regulation more broadly.

RAJAGOPALAN: I think that’s super helpful because this is true for other industrial regulation also. You look at labor law, and it’s basically identical. The norm is that no one follows the rule. Everyone knows that no one follows the rule. There are these extraordinary criminal penalties for the most minor violations, but obviously, they’re not throwing half the firms in India or half the entrepreneurs in India in jail.

We see this repeated across the board because the amount of regulation, I guess, is so overwhelming and a lot of it is so outdated and absurd that it doesn’t make sense to follow most of it. It might make sense to follow some of it, but we don’t have the state capacity to actually repeal the bad parts and only enforce the good parts.

Creating a Market for Pollution

Should we talk about your Surat paper now? I was very excited to read this. This is a recent paper that you have. Is this out in the QJE? I read something like that.

SUDARSHAN: It’s just been accepted, but it’ll take a bit of time to come out, I guess.

RAJAGOPALAN: Fantastic. One of the things you look at is exactly because of this regulatory state-capacity problem that you point out, you look at, can we use other mechanisms to address the question of industrial pollution? The most famous of these is the cap-and-trade, famous Coasean solution to the problem, which is, let’s create a market for pollution emission. It used to be done for carbon. You can also do it for other kinds of pollution, not just limited to carbon. Can we actually make this work?

I guess the interesting part, connecting the previous point we were discussing, is if it works, it’ll dramatically reduce the amount of load there is on the regulatory system because now it will switch to supervision and monitoring. The market incentives can do a lot of the work. The most important part of the market incentive is firms that are lower-cost avoiders of pollution will actually put in the technology and sell the permit. Firms that are not low-cost avoiders of pollution can actually buy the permit, but the overall effect will be about the same. First, is this how you were thinking about the problem?

SUDARSHAN: I think there’s a lot packed in there, and it’s interesting when we first started this project—I think when you are an economic student or doing a Ph.D. or learning about markets, the first thing that you are told is that markets are efficient because they reduce costs. That’s maybe at the back of your head, and it was at least at the back of my head when we started doing this project.

Then as you get into the nitty-gritty of trying to design a market, for the first time in a country—there is no other cap-and-trade market in India—you suddenly realize that you’re hitting a more basic problem which is, is this going to work at all? I remember when we just started out this effort, which was way back in 2012, 2013—

RAJAGOPALAN: Oh, wow. This paper was 12 years in the making.

SUDARSHAN: Yes. The market was launched in 2019, which tells you, if you like, something about how regulatory innovation happens. I remember in 2013, a very famous professor, whose name I will not reveal, told me, “Why have you got this trivially simple set of market rules with no banking, no borrowing? The permits are traded at the same price; there’s no adjustment for population density. These guys are not hiring economists from MIT and Harvard to write down this set of rules.” As you look back at the end of this, and you frame this as a state-capacity problem, of course, that’s what you need.

Even pulling that off is a huge accomplishment, and it’s by no means obvious that any new regulatory idea would work, just as the existing command and control is cheaper, arguably, than the market for plants that ignore it. This is the key point. We say command and control is more expensive than cap-and-trade systems. Of course it is, provided you’re adhering to the command-and-control norms. If you are breaking the laws, then it’s free.

I think there are a number of different ways to think about this. It’s about reducing costs. It’s about moving away from criminal sanctions to market-based incentives, that is, fines and permit prices. It’s about saying that you need state capacity to do a market. You need a different kind of state capacity. You need automated monitoring. You need someone to run an auction, so on and so forth. You need a few highly trained people. You don’t need an army of inspectors.

It’s not ex-ante obvious whether it’ll work, but it’s worth trying because the status quo is not working. I think as the project proceeded over this extraordinarily extended period of time, our thinking about what we were really changing and why this had potential also moved along the spectrum.

RAJAGOPALAN: I think to the students who are listening, anyone who’s contemplating a doctoral program, can you walk us through exactly how you set it up? I think that would be fun to listen to because I don’t think this 12-year portion is discussed very well in most published papers.

SUDARSHAN: I’m happy to. I don’t think that a 12-year project is what most doctoral students want to make the core of their thesis.

RAJAGOPALAN: No, that’s just the time it sits under review nowadays. I think maybe they’ll benefit from this.

SUDARSHAN: There’s a well-known paper that Esther Duflo and Michael Greenstone and Nick Ryan and Rohini Pande, some of whom are my co-authors on this, wrote in Gujarat, which was published in 2013, which was basically making a point that, look, existing regulation is broken, the data is systematically falsified. They were trying to fix that problem, but it shone a light on the fact that what is in place today doesn’t work very well.

At that time, the minister of environment was Jairam Ramesh, and the chief minister of Gujarat was Narendra Modi, who’s the prime minister now. One of the sort of ideas that bubbled up was, okay, we don’t think the data makes sense. How about we try technology? How about we put these emissions monitors in a plant, and we don’t have human beings reporting? Then very quickly from there, if we’ve got these emissions monitors telling us what pollution is all the time, we could use that to implement a cap-and-trade market.

The key thing with the market is, I need to know what you’re emitting at all points in time. It’s not easy to do if I’m sending human beings in. This idea, Minister Ramesh was interested by it. The states of Gujarat, Maharashtra and Tamil Nadu were brought on board, and I think Gujarat, the most actively, in saying, “Let’s try this two-part problem. Let’s introduce these continuous-emission monitoring systems, and then let’s use that to set up a market.”

I actually finished my Ph.D. from an engineering school. I did an interdisciplinary Ph.D., and I actually joined this project initially to work on, okay, if we are going to have these continuous-emission monitoring systems, they need certain technical standards and specifications and what should that be? The market was eventually launched in 2019, and I can’t possibly go into what happened between 2013 and 2019 in any kind of detail.

RAJAGOPALAN: Let’s just say it takes time to set the project up in a way that it makes sense.

SUDARSHAN: I think that’s fair. It takes a lot of time to set the project up in a way that it makes sense. I think with something new, you often hear this question about whether this idea will work in the Indian context. This is a self-fulfilling bureaucratic fear about a new thing maybe not working. You are pushing against that throughout this process. These things can then take a lot of time. This market launched in 2019.

RAJAGOPALAN: Sorry to interrupt, but in all fairness, it’s not just bureaucrats. It’s also all of the elite academy, which tends to think that market institutions are for the developed world, and for the developing world, it’s going to be best practices and all this industrial regulation. 

Part of the reason the Global South has a lot of this bad regulation is it’s an export from the past. It’s what the Western world was doing 100 years ago or 150 years ago that’s gotten exported to these countries, except now we have better technology, we have better information about what works and what doesn’t work. There seems to be this kind of bias that cap-and-trade and sophisticated transaction systems are only for the Western world and not for the developed world. Would that be right?

SUDARSHAN: I agree. I think there’s this bias, which is widespread within the government. It’s within academics.

RAJAGOPALAN: It’s everywhere.

SUDARSHAN: It’s everywhere. I think it’s unfortunate because the assumption is that if the United States moved from X to Y and the United States is so rich, then India has to stick with X, which is the status quo command and control until we get richer. I think that’s just flat-out wrong. It’s not the case that I say that because the markets seemed to have worked well in Surat. I’m saying that because it’s a way of thinking about the problem that kills regulatory innovation. You are not going to try experiments if you are convinced that it’s not going to work in a developing country because it sounds sophisticated, whatever that word means. It’s unclear what is so sophisticated about a market.

RAJAGOPALAN: Buried in your experimental design is also the core idea that ideas pertaining to regulatory innovation and also technological innovation diffuse much faster than state capacity. It’s much, much harder and it takes a lot longer to build up state capacity and a really good industrial inspection system, relative to using sensors that we use somewhere else in the world and saying, “Okay, it’s going to be a bit expensive, but why don’t we buy up these sensors, maybe do some jugaad to make them work in Indian conditions and actually try it out.” I think that is buried somewhere in there, which is an underrated or, I think, underappreciated point overall.

SUDARSHAN: No. I think you are absolutely right, and I think the state capacity is a big problem. The first question we should be asking is capacity to do what and why. I think, perhaps, because I was once an engineer, the point of technology is that it can change how you do things and, therefore, place different demands on capacity. You need to answer these questions at the same time. You obviously need to improve state capacity, but probably worth asking: Can we change the goal a little bit? Can we change the rules of the game a little bit? Maybe then it’s easier to get there.

RAJAGOPALAN: Now you’ve set this up. You’ve obviously persuaded the Gujarat Pollution Control Board and, in particular, Surat. I believe there are just over 300 industrial plants that come within your catchment to actually try the system. Then you randomize which plants get included to be part of the trading market, the emissions-trading market, and which industrial plants belong to the old command-and-control industrial regulation system. Is that an accurate description of the design?

SUDARSHAN: Yes, pretty much; 50% of these 300-odd plants were monitored by a market. And incidentally, this has never happened anywhere in the world, so let’s give them credit to decide to —

RAJAGOPALAN: Exactly.

SUDARSHAN: —do that. And to both the bureaucrats and politicians there at the time. Because it is very uncommon to say, “Let’s try a regulatory RCT.” That’s essentially the design, and so we compare what happens in these two groups.

RAJAGOPALAN: Now, walk me through what you find because I’m so excited about what you find. I think I’m going to mess up the description.

SUDARSHAN: We found that in the treatment group—let’s use the air pollution outcome first, although I don’t think it necessarily captures the whole story—in the treatment group, plants reduce their pollution emissions by between 20% and 30%, depending on how you calculate that, relative to control plants. This is data coming from these continuous-emission monitoring systems that are meters sitting in the smoke stack.

RAJAGOPALAN: This is pretty big. 20% to 30% is no small amount.

SUDARSHAN: This is no small amount. These reductions did not happen at the expense of higher costs. We look at costs from plant-bidding data, and we look at costs from a survey, and there’s no evidence that costs went up. In fact, the market seems to have cut abatement costs by about 10% for any constant level of emissions. Now, the reason pollution went down though is that the cap in the market, the limit in the market, was reduced over the status quo. One way of looking at this is this is entirely unsurprising. You made a rule, you made the rule stricter, so obviously, pollution went down.

RAJAGOPALAN: That’s unsurprising only to people who don’t work in the political economy of developing countries.

SUDARSHAN: Exactly. In some sense, one of the most important findings here, which I think does not necessarily need an RCT to do, is to say this actually worked. The rules were adhered to, enforcement happened, and plants complied. Since that’s not a given, these other things emerge from there.

RAJAGOPALAN: Now, that’s just one part of it, but I think the second part, which is super interesting here, one, there’s lower pollution, which is obviously what you were looking out for. But the second part is you can also have lower pollution by just shutting down all the industrial activity. We saw this during COVID. Suddenly, the sky was blue and the birds were chirping.

The point is, this is done at relatively low cost, and the second part of it is done by firms that have a lot of knowledge about the demand elasticity for their own good in the market, which then tells them, is it better for me to buy the emissions trade from someone else who’s a lower-cost avoider, or is it better for me to become the lower-cost avoider and just sell it? I think, again, that insight gets buried in such a complex experiment, but that’s exactly what you need in most parts of the world, right?

SUDARSHAN: Exactly. I think what’s cool here is that we see these plants in Surat making plenty of trades—permits are bought, permits are sold. Plants do not all end up doing the same thing, even though they’re largely textile sector SMEs in Gujarat. So you might say, “Okay, this is not the most heterogeneous sample possible.” But even within that, there’s a lot of variation in how much plants cut and what they do. You can only extract all those options using the market.

One interesting thing we found, which I think is worth dwelling on, is that the cost to cut pollution was extraordinarily low in this market. Permit prices were very, very low. The reason it was so low is that if you want to set up—there’s another aspect of the License Raj, if you like, in India. When you want to set up a factory, you need to check certain boxes on what equipment you have. It’s obviously a lot easier to test compliance with that. All of these plants have air pollution control equipment installed. You’ve got 90% of the way there, but it’s expensive to run, but much less expensive than buying it new, and you need to maintain it and operate it and so on.

If you think about this setting, what the market did in the Surat setting at least, which is going from a very high level of pollution to slightly lower, is it just let these plants turn all this stuff on, if you like, to simplify it. Looking back, it’s not the case that anything was expensive. It was all very cheap. 90% of the heavy lifting had happened already, and so all you are saying is that by opening up this flexibility, you unlocked that last 10%, which punches above its weight in environmental terms.

RAJAGOPALAN: We all know RCTs are not generalizable. Even though your experiment design is so—I don’t want to say simple—but it is, in a sense, very, very clearly put down what needs to be done, what needs to be followed. Do you think something like this can work across Gujarat in other cities? Would you then consider that it works in Bihar, or is that still too low in terms of state capacity to even have the License Permit Raj component of, oh, all firms do have the air pollution monitor? They don’t turn it on. In Bihar, that may not even be the case, right? Do you think this can be extended to other parts of India, or [do] we need to pick 20 or 30 high-state capacity, relatively highly industrialized and polluted cities, and that’s where you start this kind of a project?

SUDARSHAN: I think to the question of can this spread, we know the answer to that because it’s already spread. Gujarat has expanded this market across the remainder of the plant since Surat, there is a cap-and-trade market now in Ahmedabad and the chief minister announced a broader scale-up across the state.

RAJAGOPALAN: Nice.

SUDARSHAN: Maharashtra is currently working with us to implement a market in SO2, which is a different pollutant—

RAJAGOPALAN: Yes, sulfur.

SUDARSHAN: —than particulate matter, sulfur. That is coming, again, out of seeing that Surat could run this market, that it worked. At least the rules of the game were followed and the pollution control board was able to implement it, and so now other states are interested. I think, ultimately, emissions markets are useful in highly industrialized settings. Bihar is not a major industrial state—

RAJAGOPALAN: Yes, yet.

SUDARSHAN: —yet, so I think this question of, where’s the best place to use it and who has the highest capacity? Well, it’s really Gurajat, Maharashtra and Tamil Nadu that are at the heart of India’s manufacturing. You would think states like those are also somewhat higher capacity. There’s a nice match there, if you like, between where it’s most useful and possibly who’s best suited to use it.

I think the interesting thing is Maharashtra was one of the states that started out as a possible location of this pilot, and it never happened there. There was obviously much more concern about whether it will work or not. In some sense, the great value of these pilots is not just what’s the impact, but if you’re trying something new, could you even do it? Then people can discuss, okay, maybe if I tweaked it this way, it would’ve had a greater impact, but you’ve broken that first— 

RAJAGOPALAN: Yes, you have proof of concept, which is really difficult in a developing country to start with, right?

SUDARSHAN: Exactly. It may not work. I’m sure as you scale it up, there will be other problems that come up. For instance, the stakes for data falsification obviously go up a lot when a market is at a large scale. The resources available to stop it also go up. Now, who wins and how that turns out is something you can only figure out as you try and scale these things up.

RAJAGOPALAN: That makes sense. The other question I had, when I was reading the design of the experiment, my first thought was, “Hey, it’s just 162 firms in the treatment group. Is this market going to be thick enough?” Was that ever a concern for you? That’s the number one requirement for a functioning emissions market, that a lot of places don’t even attempt it because the markets aren’t thick enough.

SUDARSHAN: Yes, that was. We judged without having the data to prove this, that 150-odd plants would give us enough heterogeneity to trade. Part of this is because there’s a similar sector. We did some engineering modeling before launching the market. We had some guesses at how permit prices might pan out. That’s one reason why it’s tough to do in Bihar. 

RAJAGOPALAN: Exactly. There aren’t enough firms.

SUDARSHAN: Exactly. Now, for a pollutant that spreads all over the place like sulfur, you might think of a cross-state market. That would solve the thickness problem. It would create a different problem, which is the political economy of regulation, but the economics of a thicker market might be solved that way.

RAJAGOPALAN: That makes sense. I am very excited about how this gets extended to other places. I think the overall insight of how markets and market prices are basically knowledge-generating devices, this is an old-school Hayekian idea. A low-transaction cost device, I think, somehow that gets missed out in how we do most economics. Somehow the idea that you could just bring that simple insight back and make this work, just—I loved reading the paper. This was awesome.

SUDARSHAN: It’s an extraordinary bargaining tool because we used to have these monthly meetings when this market was happening. Each time the cap was lowered or even initially when the cap was set, the industry association would come and say, “This is going to be too expensive.” The regulator was able to say, “If it’s so expensive, then why are permit prices so low?” You had a signal that instantly resolves this bargaining, which otherwise is based in absolutely nothing. You have a concrete objective piece of evidence that tells you whether something is too expensive or not too expensive. You are exactly right. That information content is very valuable.

Vehicular Pollution and Rationing

RAJAGOPALAN: I want to ask you a couple of things that are related to air pollution, but also not industrial pollution. India’s air pollution, the way you alluded to it at the head of the conversation, comes from farm activity like crop burning, which might be the dumbest place from where we are doing this. It may come from construction, it may come from vehicular pollution. Vehicular pollution has exploded in Delhi.

You also ran a study on the odd–even vehicle number rationing that the AAP government tried out in Delhi and so on. What is a good way to think about vehicular pollution in India? Is it a problem of too many vehicles? Is it a problem of too many old vehicles? Is it a problem of too many vehicles, but too much urban sprawl, and therefore too much construction, which is the bigger issue relative to too many vehicles? How does one think about this?

SUDARSHAN: That’s a very difficult question. Let me take a shot at it. There’s a lot of vehicles, there’s a lot of road dust, there’s a lot of construction. First, it’s all of those things put together. There are a lot of cars in other parts of the world, more so than in India. 

It can’t just be because of the cars. My guess, is that our ability to regulate how much old cars pollute is quite poor. We have a smog test system in place in most states in India. There’s just one interesting factoid: hardly anyone ever fails. You can take that for what it is. It isn’t proof of anything by itself. If hardly anyone ever fails, and you have a large number of three-wheelers, two-wheelers, older vehicles, then you might start to worry that there’s something wrong with how these mock test regimes are in place. I think it’s a really thorny problem to solve.

I think an estimate I saw recently was about 15% to 20% of Delhi’s air pollution, for example, comes from transport. What does that mean? That means even if you could think of something and cut it by half—which is extraordinary, this is not going to happen very easily—if you cut it by half, you would see reductions by about 10% on average in the city. Most people would not even notice.

That gets to the political economy of why it’s a difficult problem to solve because to do that reduction you have to work at all of these different 15% policies at the same time. But any one of them taken by itself is going to generate a huge amount of opposition from people who want to use their cars or want to use their old cars or whatever. It’s not going to turn into a win for a government in a four- or five-year term that they will be able to point to and that we can agree with it. I think that’s why it’s a tough problem to solve.

RAJAGOPALAN: I think it may be even worse than what you’re saying. Of course, there’s this timeline problem, but there’s also a problem of what we can point as the proximate cause and what we can point as the more persistent cause. If we now start tracing back, let’s just take vehicles, for instance, if we start looking at why are vehicles in India so much worse than vehicles in other parts of the world, one simple thing is our standards are lower.

The other part of it is our tariffs are what, 125 percent on cars? Even a Honda, which is a very upmarket car in India—the model they run in India is the 30-year-old Civic that no one would drive in the United States. Would this problem get solved if we reduce tariffs and allowed better quality cars? Sure, but how much impact would that have on air pollution in cities? We can’t exactly measure in a precise way. Same thing with roads. If you reduce the amount of urban sprawl, are people more likely to carpool or walk or take the metro or other things? Yes. 

Again, for instance, Delhi, maybe not other cities, just decimated its bus system. This is over a decade. The Supreme Court just systematically destroyed Delhi’s bus fleet. Now you can’t be surprised that people want cars, because Delhi’s got this horrible climate. It’s either too hot, too cold or too rainy most of the year. It makes no sense for anyone to walk or bicycle. The distances are too large.

Which of these is [the] approximate cause that we should actually devote our effort toward? Is that the biggest problem? Because to me, as an outside person looking into this, that seems to be the biggest problem, even at the point of which problem do I pick up to study next?

SUDARSHAN: I think you mentioned two things. The first was, what sort of cars are we selling and using in India? It’s actually not the case that the standards, say, in a city like Delhi are low. If you buy a new car in Delhi, at least the efficiency and pollution standards are pretty high. It’s just that these old cars still remain in the market. Delhi now has a rule that says cars older than 10 years can’t be driven in Delhi.

There’s a flourishing market outside Delhi for cars that are older than 10 years. We had one at home, and it’s now being used by a relative of ours in Bangalore. These older vehicles still keep polluting. They just do so elsewhere. One is the stock versus flow problem, where the flow may be getting cleaner, the stock is not. I think with electric vehicles, and especially if India is willing to slash tariffs on China, which is sort of something the whole world will have to confront as being the only possible way of getting cheap— 

RAJAGOPALAN: Is a political nonstarter, or at least today, but who knows in the future.

SUDARSHAN: It is currently, but that’s the only place cheap electric vehicles are being manufactured, so—

RAJAGOPALAN: Exactly.

SUDARSHAN: —it’s got to come from there at some point, but you’ll still have this stock versus flow problem. Old cars are not going away. I think the public transport bit is really interesting because it also raises a bit of a chicken-and-egg situation. The Delhi government, myself and some of my co-authors have been trying to work on a congestion pricing scheme in the city, which is, say, a more market-based version of the odd-even driving bans that you alluded to earlier.

RAJAGOPALAN: Where people are just buying older cars to solve the old-even problem. It’s actually perverse, right?

SUDARSHAN: Yes, you’d pay a price for driving through entering the city at particular times or potentially driving through intersections at a particular time. You say that people in Delhi care about pollution; I remember the government organized a conference early on when we were presenting this idea, and I presented the idea and got lots of pushback from the audience, who were people with cars basically saying, “Okay, this makes no sense. It may be something London is doing, but London also has a great public transport system.”

There is this question, a little bit of a chicken-and-egg situation. If you have no disincentives for driving, then the people who have political voice are not going to be using your buses and so your buses are not going to improve. I think these two things need to go together. Unfortunately, in the political discourse, they can actually be pitted against each other. You first do this, and then we’ll think about that. 

RAJAGOPALAN: Delhi has enough numbers that if the government allowed private bus fleets, and [the] Supreme Court didn’t shut them down every time private bus fleets were allowed, that there could be a flourishing market. Now, we have massive amounts of data. The post-Uber world where everyone has a smartphone is a very different world. The same data you would use for congestion pricing, they would use to figure out a bus and a jitney fleet, right?

SUDARSHAN: Yes. You could potentially think of that happening, say, with small electric 12-seaters, 14-seaters. In fact, Uber has launched this in New York. You could think of that being a decentralized transit system that sits somewhere in between the car with one person and the bus with 50 people that can be introduced by the private sector. I think it’s a cool idea. We should try it. I don’t know if it will work, but it certainly seems, I agree with you, that it might work and it might get us some of the way toward what we need.

RAJAGOPALAN: After the podcast, I will connect you with some people who have some of this data, and we can chat more about this later. 

When you did study Delhi’s odd-even vehicle rationing—and vehicle rationing is the worst thing, but it’s also the easiest thing to do. The odd-even seems very fair. Most people don’t pick their own number. It can either be odd, or it could be even, it can’t be both, which makes life simpler. It’s very clean and easy to enforce. In those terms, it was a sensibly rolled-out plan. For a few weeks of the year, they did roll it out when Delhi is an absolute gas chamber. Different people say it had different effects. 

You estimate it had somewhere between, say, 8% to 13%, something thereabouts. It lowered pollution. Other people say it had zero effect. Now, what’s going on with the data measurement here that we get such vastly different estimates of whether the odd-even scheme worked? I think this should be one of the easiest things to get clarity on.

SUDARSHAN: The odd-even scheme was basically a ban on vehicles with either an odd number plate or an even number plate on alternating days. You could only use them on particular days. The Delhi government tried this a few times, and compliance dropped over time. The first time they tried it, everyone was very excited about it. It was the flagship program of what was supposed to be a sequence of steps. You’d expect that this was going to certainly reduce the number of cars on the road, and so it should be intuitive that it will reduce pollution by a little bit.

It’s not that easy to prove it because you think about pollution that spreads over space and time, where you have a limited amount of monitors and where the first order determinant of pollution is all sorts of other things such as whether farmers are burning crops, or so on. If you look at a time series of pollution in Delhi and overlay the odd-even scheme, I think that time series actually saw increasing pollution, which is not surprising because it gets more and more polluted in the winter. That’s why the scheme was there.

If you just said, “Okay, what were the 10 days before the odd-even like, and what were the 10 days in the odd-even like?” you might conclude that it had no effect, or in fact, that it was even worse. Now, we ran a definitive estimation that we thought was carefully done. Other people approach this problem in different ways. This is an unfortunate part of empirical social science that there are different ways you can try and answer a question. We have an opinion on what was actually delivering causal inference. If you do something like a before and after, you will come up with a different number.

I think, so yes, it did reduce pollution per our analysis. The amounts by which it reduced pollution pass a sanity check, given how much you think these cars might contribute in the first place, but it underscores the political problem, that if you can’t see it, it’s very hard for a politician to sell it. It’s not so easy to just see it.

RAJAGOPALAN: Yes, and this is such a classic externality problem, and it’s literally where the wind blows that will change the monitoring or whatever is getting picked up by these sensors day by day, so it’s very, very hard to point to who is the blameworthy cause or the blameworthy party, and everything will unravel from there. 

A question on trying to connect the problem of vehicular pollution and crop burning, which is mostly, we’re thinking about the problem from, say, Punjab to Benaras. That’s the extent of this. Is there a way to think about cap-and-trade when it comes to crop burning, such that the farmers who can most efficiently clear the field can actually sell their permits to those who can’t? Is this totally crazy because this doesn’t even have that minimum amount of heterogeneity that exists when it comes to garment manufacturers or someone else?

SUDARSHAN: I don’t know about a cap-and-trade scheme. I think it doesn’t lend itself very well to this sort of setting, and monitoring is a problem. There is an interesting paper though, by Rohini Pande, Namrata Kala and Seema Jayachandran, who are environmental economists in the United States, and they’re trying to see if a payment for ecosystem services approach might work in Punjab where you use satellite data to figure out whether farmers have burned their crop or not, and if they didn’t burn their crop, you pay them essentially to do this.

RAJAGOPALAN: It’s a carrot instead of a ban, basically.

SUDARSHAN: It’s a carrot instead of a stick. You could argue that’s economically efficient. The Coasean solution, even though it seems unfair, can also work just as well by paying farmers not to pollute. I think that’s an approach to using a financial incentive that may have some promise. I’m not sure. I think a market may be both difficult and overkill to solve this problem because you’re not really worried about a whole lot of heterogeneity in costs.

How Subsidies Complicate the Pollution Problem

RAJAGOPALAN: That’s true. No, and the other thing is, I have a bigger problem with the Punjab farmers growing paddy in the first place. It seems like the only reason they do it is because of massive amounts of output subsidy. It is not the lowest-cost place to grow paddy, and no one would grow it unless the government had put in so many subsidies.

In an ideal world, if the subsidies disappear, the paddy will disappear and the crop burning would disappear. It’s not clear to me we should actually keep the subsidies and create a market for crop burning if we can just solve the root of the problem. That’s even harder to solve than air pollution, I think, agricultural subsidies.

SUDARSHAN: No, you’re absolutely right. If you think about Punjab, you’re producing a crop which is dependent on government support prices, so the market value is lower than even the low support prices. Pretty much all the inputs are subsidized. The electricity to pump out water is subsidized. The environmental externalities are massive, both from this air pollution problem and from water depletion.

At some point, forget being inefficient. It’s not even clear this is a positive social value activity. You’d have to crunch through the numbers. More and more, as you have rice that’s worth less and pollution that’s causing massive damages and a state that’s going bankrupt paying for electricity to deplete groundwater, you begin to wonder whether we might all not be better off if we took all this money and didn’t have so much agriculture in Punjab in the first place, at least paddy agriculture. That’s entirely plausible. It’s also seemingly impossible to actually make happen all at once.

I think if you read Ashok Gulati’s written piece in the Indian Express where he alludes to similar things, this is not a novel observation, but the question of how we get from here to there.

RAJAGOPALAN: Yes. It’s always a transitional gains trap.

SUDARSHAN: The transition is difficult.

RAJAGOPALAN: I think it’s Ashok Gulati who wrote a piece on how we should pay Punjabi farmers to just not grow paddy and put in a solar farm or something like that. They’ll become net green energy positive; instead of depleting energy resources, they’ll hopefully replenish their groundwater and then stop the crop burning. It’s many solutions in one. 

You just mentioned the point about getting electricity subsidies which makes farmers pump more groundwater, and you have a different paper on this, where you look at groundwater in Rajasthan, again, where electricity like Punjabi is highly subsidized. You make an interesting observation. Most of the people who look at subsidized electricity say that groundwater depletion is happening in India because the electricity is so subsidized. What you find is, because of the subsidized electricity and the way India’s DISCOMs are structured—and we can get into that a little bit later—it means that electricity companies at the state level don’t have enough electricity to actually provide. It’s a huge supply crunch then. So they start rationing the electricity. Even though the price is much lower, the supply of electricity to these farmers is just a few hours in a day. They don’t actually end up using anywhere close to the amount of groundwater, as is suggested by the literature.

First of all, did I get that about right, that they’re not terribly overusing groundwater, but there are still other problems you discussed in the paper which we can get to in a minute?

SUDARSHAN: I think that paper was written about Rajasthan, which has got the most extreme groundwater depletion problem I think in India. The way rationing works is it’s not that distribution utilities have an electricity shortage; it’s that I cannot possibly stay afloat if I let someone who has free power use as much as they want. When I have a zero price because I don’t have an unlimited budget—and I think this is sometimes forgotten when we talk about price subsidies—I’m going to have a quantity instrument in place instead.

Across India, you have[a power ration to farmers, which is the only check on how much groundwater they use. Now, as a state gets more and more depleted, you need more and more energy to pull out a small amount of water. That small amount of water is incredibly valuable. What we were showing is that in the worst places in India, like in Rajasthan, on average, you are not necessarily using too much water, even though electricity is free because you have these power rations.

Now, of course, we are using it inefficiently because what you want is for the inefficient farmer to not be using any at all, and for the efficient farmer to be able to use more. On average, maybe it works out. This starts to bite as things get really bad. Some people have interpreted this paper to mean, “Oh, free electricity pricing doesn’t matter because you’ve got these rations.” That’s not quite what we are saying because when you started—

RAJAGOPALAN: I know

SUDARSHAN: —with lots of water, then you have a small amount of electricity, incredibly inefficient use and a groundwater externality. When you get to a complete desert, like Rajasthan is, maybe then the quantity ration is enough. The other development effect of that quantity ration, though, which is worth mentioning, is that it’s going to affect household supply as well. This has been a massive problem for Indian state governments across the country, that if you are cutting power to farmers because you can’t afford to let them use as much as they want, both for electricity, water purposes and for your budget, then you are also, maybe, cutting power to households and to villages.

RAJAGOPALAN: In those areas.

SUDARSHAN: Yes, and so on. You’re also killing the opportunity for structural transformation. There are ways out of it, but that’s been historically further side of all this.

RAJAGOPALAN: No. I’ll tell you what I got when I was reading the paper. Two things. One is, I think the idea that there is an important distinction between overuse versus misallocation is underappreciated. Because typically when we are thinking about subsidized or free electricity, we are so focused on the overuse, we never quite get to the misallocation. Here you are able to show that effect very clearly because you are more keenly studying the outages as opposed to the overuse. Does that make sense?

SUDARSHAN: That’s absolutely right. We complain just now about Punjabi farmers, but they’re going to be farmers there who are doing amazing stuff. We want them to use as much electricity as possible. There’s definitely both things going on and the quantity instrument really helps, which is being used across India.

RAJAGOPALAN: This could be something fairly simple. There are a lot of farms only engaging in farming, but there are a lot of farms in India, smaller farms, which also have dairy farms attached to them. Dairy farms require cold storage. They require access to fridges and better electricity and so on. You are actually going to kill dairy farming, which is a much higher-valued use and also a more sustainable activity in a place like Rajasthan versus growing whatever else they’re growing. I don’t think they’re growing paddy in Rajasthan, but whatever other stuff that they’re growing in Rajasthan is not quite ideal.

SUDARSHAN: They’re growing wheat. You are absolutely right. The question is really how do we get from where we are today where a bunch of people have free electricity and it’s either rationed but badly misallocated to what you are talking about where the productive user can get more? One of the things that myself and some other researchers have been trying to do is to say, can we buy out the distortion, so to speak?

If you use less power than a given amount, I’m going to pay you some of the cost of that power you didn’t use, which is essentially saying I’m going to share the subsidy burden with you if you will just stop using this electricity. What that’s going to do is then you are only going to use the power if it’s worth more than that subsidy. Which then shifts consumption toward these more efficient uses. That’s one of the types of ideas that you might think about in trying to figure out how do we get from here to what you described, where the misallocation is lower.

Pigouvian Subsidies

RAJAGOPALAN: We talked about the Coasean market, the cap-and-trade market that you talked about in Surat. This is the other option. In classic, environmental Econ 101, we either talk about Pigouvian taxes and subsidies, or we talk about a Coasean market solution, where property rights are well defined and you can trade. Now coming to the Pigouvian subsidy, don’t you face the same knowledge problem that one would expect to face in every Pigouvian subsidy? To some extent, the household or the farm will reveal that they are the higher-valued use, so therefore they cut back on their electricity, and they actually accept the subsidy. Isn’t it very difficult to figure out what the amount of subsidy should be precisely for that reason?

SUDARSHAN: One, it’s difficult with Pigouvian taxation to figure out what a subsidy or tax should be because you’re trying to identify what the environmental damages are. You can start this sort of scheme by framing it differently. I am paying for your power, so I know how much I’m spending. I know the difference between the procurement cost of electricity and what you are paying, which is maybe zero. Let’s just focus on that. Let’s divide that buy-up differently. There was a small pilot we did in Rajasthan where we said to farmers, “Okay, here’s a target level based on your pump set capacity and so on and past use. If you consume more than this, you’ll pay the full price. If you consume less than this, we’re going to transfer to you the full cost of that power.”

What that effectively means is that if I cut my consumption, the government doesn’t necessarily directly save any money. What it’s doing is it’s taking the subsidy that was earlier being used to pay for your electricity and sending it to your bank account. Then you are not using that power. Somebody else may be willing to pay to use more. I think the way you are getting around it is you’re saying, “I’m not even going to think about the full social externality. Let me start with the difference between the price you pay and the price of power that it costs me and split that pie.” Now, it may not be enough. If the cost of power is low enough, that may not be enough to drive behavior change. That’s an empirical question.

RAJAGOPALAN: There’s this other really nice paper [by] Shoumitro Chatterjee, Rohit Lamba and Esha Zaveri. I think it’s recently come out in Nature. They talk about the problem of declining water tables and subsidies. There they don’t just focus on the electricity subsidy. They’re looking at all output subsidies so that’s more focused on the MSP. Is this one of the problems when you study this piece of the puzzle only with the subsidy that the state government and the DISCOM provides every household or every farm?

Because that’s a tiny portion of a subsidy that is overall incentivizing a farmer to overuse groundwater because then that gets them access to fertilizer subsidy and so many other things. 2.25% of India’s GDP is agricultural subsidies, right? This is just a tiny portion of it. Is that why this is harder in the context that you’re looking at?

SUDARSHAN: First, it’s not necessarily a tiny portion of it. I think in Punjab, you are now spending, if I remember right, $16 billion a year on paying for this electricity. 

RAJAGOPALAN: No, but compared to MSP.

SUDARSHAN: The relevant question with MSP is the difference between the market value of the product and the MSP price, right? The whole thing is not a subsidy. You’re absolutely right, if your output market is distorted, you’re going to have too much agriculture. And if your inputs are distorted, that’s a problem too. Now, the MSP doesn’t work very well in many parts of India. Although it exists on paper, I’m not sure that it binds everywhere. One of the places it does bind—

RAJAGOPALAN: Is Punjab, Haryana.

SUDARSHAN: —and we keep coming back to this is Punjab, Haryana, the northwest of India. It’s certainly both, but the MSP subsidy bit is not the same as the MSP bill.

RAJAGOPALAN: Fair enough. I overall mean that electricity subsidy is not the only reason someone would overuse groundwater.

SUDARSHAN: Absolutely.

RAJAGOPALAN: Right?

SUDARSHAN: Yes.

RAJAGOPALAN: Which means that immediately, if that’s the piece of the puzzle we’re looking at where the offset is the average cost of the DISCOM company versus the subsidy that they are trying to offer, which the state government either pays late or never quite pays at all, then that is going to capture a very small part of the aggregate overuse, which is where I was getting to. Do you see that problem when you look at Rajasthan?

SUDARSHAN: You’re absolutely right about that. The reason when we ran this pilot and talking about it, that I focused on that very small bit, is the accounting benefits. Because the entity that is implementing this program can be the same entity that is paying the subsidy. The math just adds up on their books. It’s really just a convenience in terms of how you would start this, but you’re absolutely right. The economics of this are larger. The distortions are larger, and you could pay more to buy out this activity while still improving efficiency.

Electricity as a Right

RAJAGOPALAN: Coming to electricity, you also have a really nice paper. This is in the Journal of Economic Perspectives with co-authors called “[The Consequences of Treating] Electricity as a Right,” right? This is the big question. This is part of just how I think modern welfare discourse goes: Air, water, food—everything eventually becomes a fundamental right. An important aspect of electricity becoming a fundamental right in developing countries is that it automatically means that the right will be covered or provided through subsidies and the theft and other kinds of distributions. People don’t always think that they are supposed to pay for it and very quickly it becomes a norm where paying for electricity or at least paying the market price for electricity is not the norm.

For most of us who live in the developed world, when we think of electricity as a fundamental right, we think of it within the context of market allocating supply. We’re thinking of it in terms of how do we get to a situation where markets can provide, or the supply can adjust to, how much electricity people actually want to use. It flips in developing countries. In India, if I understand the Indian electricity market correctly—and I think we have a paper [by Dinkar and Surana] on this on our 1991 Project, on how DISCOMs are totally messed up in India—basically, there are multiple things going on. Industrial use and commercial use are basically burdened with cross-subsidizing households and agriculture. Agriculture seems to be the worst offender. Households and the middle class come in next and so on.

Then what you find in the paper is eventually DISCOMs have no option but to limit supply and so the terrible consequence of making electricity a fundamental right or a welfare entitlement that everyone should get is that very often most people, especially in the lowest rungs of society, the poorest or those who have the least political voice, are not going to end up getting it. How does this look in India? Does it look different in different states? 

SUDARSHAN: Yes, I don’t think electricity should be treated as a right, but it’s clearly very important. When we say electricity not [be] treated as a right, that is to say, you could think of ways in which you give people money so that they can pay for electricity. That’s a different statement from saying that electricity itself should be given away for free because it’s a right.

I think a lot of the political discourse in India and the implied fact that distribution [of] utilities in India, from households and agriculture, of course, is often free. Households—often you’re recovering maybe 50% or so of the charged tariff. In other words, I have a subsidized tariff, and then I’m only getting 50% of that back. You have widespread theft that you’re then turning a blind eye to. In fact, what you’re saying is you have a right to this electricity whether you pay for it or not. It’s a complicated question to solve because, alongside that paper, we had a documentary movie shot about that project, which was called “Aayi Gayi.” It won a couple of awards. It’s online, but I use it to teach.

RAJAGOPALAN: I haven’t watched this. This is so cool.

SUDARSHAN: Yes, so I use it in teaching somewhat. In that movie, it was basically conversations with people about why they don’t pay their electricity bill and what does electricity mean for them? You have these heartwarming stories about how crucially important electricity has been in their lives. I sometimes use this in my economics classes, and I’ve used it in Ph.D. econ classes where students will then turn around and say, “Yes, seems reasonable electricity should be a right. You’re an absolute monster for saying that people should have to pay for it and that you should punish theft.”

It’s a very powerful narrative, but it has destroyed the Indian distribution utility market. You will occasionally hear Indian politicians say, as a mark of achievement, that India is now power surplus, by which they mean that there’s more power that they can generate through power plants than utilities are buying, and that’s actually just an incredible indictment of the system.

RAJAGOPALAN: That’s so stupid. 

SUDARSHAN: What it’s saying is that in this energy-starved country, utilities can’t sell their power. They don’t want to buy it, even though you have power plants that could produce it. It’s an emotive issue, and I can see why people treat it as a right, politicians. If you ask people, should healthcare be a right—healthcare is a different good—but many people will say, “Of course, we should have the NHS in the U.K.,” which is free healthcare, “and we’ll pay taxes for it.” It doesn’t work with electricity because of the type of good it is, but much of the same intuitive reasoning does go through. It’s a problem we have to solve.

RAJAGOPALAN: Where do you think the intuition is broken? Is it that people just don’t understand supply elasticities? That supply is not a fixed amount? It will be more forthcoming the more people are willing to pay for it and the thicker that market is, and the more contestation there is between different DISCOMs. Is that where everyone’s brain checks out?

SUDARSHAN: Yes, I think so.

RAJAGOPALAN: Because this is so obvious to me as an economist that I can’t fathom why people would not understand this.

SUDARSHAN: First, of course, with healthcare, you can make the vaccine free. I’m only going to take one vaccine. Maybe I won’t even take that if I’m the right sort of person, but in the case of electricity, you make electricity free. I’m just going to use more and more and more, right? There’s no limit to demand, and that’s why you can’t have it be completely free and completely unrestricted at the same time, given any budget constraint.

At an intuitive level, people do understand that. I think it’s not necessarily true that if you’re very poor, if making capital investments to buy new appliances and so on is beyond what you can imagine yourself doing, I think then this calculus changes a little bit because you’re never going to be this person who uses large amounts of power. I’m just trying to put myself in the place of someone who is saying I deserve free electricity because you can see these very high returns to your initial units of consumption. You know you’re not buying an air conditioner and so you know this is hypothetical. In your case, you could argue that some of this doesn’t break down as quickly as it can.

The problem is that when you implement this through turning a blind eye to theft and through subsidies—and in particular theft, because sometimes tariffs do account for different levels of consumption—if you let theft happen, then it’s not just this very poor person with this arguable case to make who is using power. 

RAJAGOPALAN: That is still politically connected.

SUDARSHAN: Yes. I work with many distribution utilities. Nonpayment rates do not vary really by the amount of consumption. Most of the power that is not being paid for is being used by people who are consuming a fairly large amount. In other words, these are not the poor. These are presumably better-off people. The rights discourse lets those guys through. Even though the emotive power of it and, if you like, the analytical basis for it is really defined by much poorer people. 

RAJAGOPALAN: This is not the grandmother in “Swades” who is the worst violator of what’s going on.

SUDARSHAN: Yes. In fact, I don’t care. If it were just the very poor not paying for their power, it would mean nothing. They use very little. There are probably large positive externalities. It’s no cost at all. It’s just that it doesn’t stay there.

RAJAGOPALAN: Is the problem just at the DISCOM level? What is the root of the problem? Is it that state electricity boards are so deeply involved in this? Is it that we’ve decided to subsidize electricity to farmers and certain other kinds of activities? Where does the root for this lie?

SUDARSHAN: Let me put it a different way. Let’s suppose we are where we are now, and you ask why aren’t we changing it when all of this makes sense. You end up in the same place as we started earlier with those industries. Let’s say 50% of your population isn’t paying their bill. What are you going to do about it? You can’t actually disconnect 50% of people. Most Indian utilities find their hands tied. They can target a few people, but for the most part with a very large number of people not paying, everybody knows nothing is going to happen. In fact, without something changing about technology or otherwise, we can talk about those technological improvements, but just in the status quo. If I’m going to disconnect you, that’s a very harsh punishment. You could potentially even be sent to jail, but I can’t possibly do it with millions of people not paying.

Everyone knows this. When the U.K.—as energy prices were rising a year ago before the government put in place an energy price cap, there was this petition that was catching on like wildfire, which was the “Don’t Pay” petition. That’s what it was called. Basically, people signing up to say, we will not pay our electricity bill. What are they going to do about it? That’s a very powerful idea because there’s nothing they can do about it. If enough people don’t pay their electricity bill, it’s very hard to do something about it. You’re not going to let poor old ladies freeze in their rooms, in their houses because they are not paying for their electricity bill.

RAJAGOPALAN: What they’ll do is everyone is going to get 60% less electricity.

SUDARSHAN: Exactly, but there’s a tipping point. The problem for India now is now that we’re here, how do we claw our way back to an equilibrium where only a small amount of people are not paying? Then I can enforce on them with certainty. How do I move there? There are ideas people have, but it’s not a solved problem. You can see why it’s not an easy to solve problem either.

RAJAGOPALAN: Yes. It’s actually even worse. The paper I was telling you about that my colleagues, Ankita Dinkar and Samrudha Surana, have written—this is basically the DISCOM problem. One of the things that DISCOMs do is manufacturers and commercial users have faced a higher tariff relative to households and farmers. Because they face this higher tariff, but they also have these unscheduled power cuts, manufacturers and commercial enterprises want more stable electricity, and so they want to check out of the system and get their own captive power plants.

Because too many of the high payers are going to leave the system, now the government starts making it very difficult for them to get a captive power plant. They have crazy number of rules associated with it. We’ve turned electricity into a market for lemons. Where people just can’t check out of this system anymore. Then all the good folks check out, you have no one left to cross-subsidize the folks who are either stealing or subsidized too. It’s hilarious as an economist because the solution is just so obvious and staring one in the face. When I read this stuff, I’m like, “Wow, who thought of this crazy system?”

SUDARSHAN: There’s one view which is, is India rich enough to say, “We will first pay for lots of people to have power. We will not worry about theft. We will let the consumer surplus and the economic activity from electricity rise, and we will keep taking the hit on government budgets.” That’s a strategy that de facto many states are now essentially implementing. It’s not a strategy that, say, many countries in Africa could do, but it’s the sort of thing that India is on the boundaries of potentially being able to do, which is to just say we are going to power through with this.

RAJAGOPALAN: Pun intended.

SUDARSHAN: Yes. Now when it starts hitting these other factors that you’re talking about, that’s the question about how bad it can get.

RAJAGOPALAN: Here, this is also how countries end up in that low- and middle-income trap. They just stop growing. Here, one of the biggest things is because of these unscheduled power cuts, which have to be distributed in some form across the board, your most productive activities are going to have a lot more cost to the unscheduled power cut relative to your least productive activities.

They’re all going to get very similar unscheduled power cuts. I think there was this paper that estimated about, I want to say 9% to 10% of India’s export competitiveness reduces in manufacturing because of these power issues. That’s an extraordinary amount. That’s what most poor countries do. I know what you mean by India now has the tax revenues and the sort of democracy and vote system to make this a possibility, but it’s not clear to me how this is going to end well.

SUDARSHAN: One of the things that made Narendra Modi so popular when he first became prime minister was this extraordinarily capital-intensive solution that he thought up to improve household supply. Which is basically to say, I’m going to build out a whole different parallel grid, which will ration power to farmers for their pumps but not ration power to households.

That is a super inefficient, awkward way of getting around some of it, but it still runs into the fact that you can spend tax revenue on paying for some people’s electricity or you can spend tax revenue on roads and schools and so on. Arvind Subramanian had this op-ed, that he published a couple of years ago, which had some horrific number—2%, 3% of India’s GDP being directly or indirectly spent paying for the private consumption of electricity. Then this is mostly households and agriculture. Industry pays above the marginal cost for the most part.

RAJAGOPALAN: Yes, well above and they still get power cuts.

SUDARSHAN: It’s just a staggering problem. The magnitude of it just doesn’t register on most people just how bad it is.

RAJAGOPALAN: Other than an enormous supply-side shock, India embraces nuclear power plants or something like that, which dramatically reduce the cost of producing electricity, I don’t see how one gets out of this. Do you know a way India gets out of this?

SUDARSHAN: The only way we get out of it is enforcement. I think there’s broad agreement that you have to have enforcement, especially on large users of how much they pay. The theft has to stop. One thing that the government has put a lot of eggs in [is] this basket of trying to use smart meters. Basically, the idea being that I can remotely disconnect you using a smart meter, so I can send that enforcement cost close to zero. Now I could potentially enforce when I have millions of nonpayers. There’s a huge effort by the central government in India to push these smart meters through and to have agriculture metered and so on. It butts against a political economy problem, which it’s the state governments that are going to have to start disconnecting people.

RAJAGOPALAN: Exactly.

SUDARSHAN: They’re the ones who are fighting elections.

RAJAGOPALAN: And they have no incentive.

SUDARSHAN: They have a more nuanced incentive, let’s say. They would like to get more money out of people, but they would also like to win an election. The challenge in India is this kind of center–state debate. You see it with agriculture; it’s got this second layer because agriculture is not even metered. It’s this black box to which you can attribute whatever consumption you like.

You said a lot of this consumption is agriculture, but in fact, if I wanted to give you as a large industry or a mall a free run on power and take a bribe from you, well, the power went in, someone has to use it. It’s very convenient to say it’s all these farmers on whom I have no meters who get power for free anyway. They’re the ones using it. There’s several layers of this that pull into place, but I’m pretty optimistic. I think eventually the numbers are just so large—

RAJAGOPALAN: That it’s unsustainable.

SUDARSHAN: —that it’s unsustainable. There is a technology solution. I think we will, in an inefficient way, get there. Yes, you will be paying a lot of tax revenues along the way, but that 9% will become 8%. It’s a solvable problem, I think.

RAJAGOPALAN: Yes. In the beginning, at least, let’s start allowing captive power plants and more competition among DISCOMs. Most states have one state-owned DISCOM, which has the worst incentives, completely backstopped by the government. We could even do some minor tweaks to make the problem a little less severe in the short run. About that, I am definitely optimistic.

The Near-Extinction of Vultures in India

Before I let you go, I want to discuss your vulture paper, which is one of the most extraordinary things I have read in recent times because it sounds so unbelievable when you read it the first time, and you’re like, could this be so crazy? This is a paper called “The Social Costs of Keystone Species Collapse: Evidence From Vultures in India,” and this is by Frank and yourself. It’s in the AER. You basically look at the enormous collapse of India’s vulture population, and you managed to narrow it down to one particular drug, which is diclofenac. This is used as an anti-inflammatory muscle painkiller across the world. It’s been used forever.

Except you find that it started going into veterinary use in the early ’90s. Because of this, vultures who are basically nonhuman scavengers eating these dead animals suddenly got diclofenac in their bloodstream and died of kidney failures in such alarming numbers that they almost went extinct. 

Then we’ll come to the second part: Because of this species collapse, there is actually a huge second unintended consequence on human beings because now there’s all these rotting dead animals all over the place, which means there’s a public health and sanitation problem and humans die in larger numbers. 

Let me just start with what kind of a crazy person gets this idea to test this and how did you land on this?

SUDARSHAN: Well, I am afraid, Shruti, I’m not going to take credit for the first part. That kind of crazy person is definitely not us. The stuff that we do in the paper that’s new is figuring out the timing of when this happened and what the impacts were.

This is an incredible story. The detective who worked to figure this out comes from a Nature paper written in 2004, 10 years after the vultures had basically started disappearing. They figured out why it was. It’s most definitely not us who came up with the reason why the vultures disappeared.

There’s a beautiful essay that Arundhati Roy has in her book, “The Ministry of Utmost Happiness,” that is about this. She talks about the dramatic increase in us having ice cream and Kwality Walls and so on as we were growing up as kids. Suddenly the ice cream world went from very boring to super interesting over a short period of time, and that’s exactly the same period of time as these birds started disappearing. The story of how the vultures died is not our contribution in the paper.

RAJAGOPALAN: That’s not the only crazy part of the paper. Who would imagine this has such a huge impact on human beings? Why would you even think that? Any normal person would’ve just assumed, “Okay, vultures are dying. We need to really save the vulture.” The idea that this could have such a huge impact on humans that is definitely another crazy idea.

SUDARSHAN: I think it’s a crazy idea for economists, for maybe most of us. I think the ecologists who study vultures know what they do. I don’t think the numbers are staggering. Once you do it, you can sort of see, of course, you have 500 million cattle, and if all these carcasses are not being disposed of efficiently by these birds, a lot of bad stuff is going to happen.

I think, yes, it’s one of those papers where if you went to an ecologist, they’d say, “Yes, vultures are doing these useful things, and of course dogs spread rabies and if the birds aren’t there, there are going to be more dogs and more rats and more carcasses.” You have that factoid. If you want to figure out what it actually means—is this just a tangential side thing that happens or is it first order—you need to go through the numbers. I think that’s sort of the contribution that you have.

RAJAGOPALAN: It is staggering, right? You say it led to a 4.7% rise in human mortality. That’s pretty big.

SUDARSHAN: No, no, it’s a staggering amount. There’s a paper in the QJE, I think, that has an estimate of the impact of climate change heat deaths only. It’s comparable to that, which you think about it, it’s just one species. It disappears and you have this staggeringly large impact. We ran our paper in 2005, but there are still no incinerators in India. You still have carcasses thrown in water bodies and so on. You have a rat population. You have a lot of rabies. It’s not clear that these damages have suddenly disappeared on their own.

I think there are some economists, Partha Dasgupta, for instance, who [have] been making this point for a while about how we are underpricing our natural endowments. Ecosystems, in particular, are stuff that we set the value off to zero because we don’t know how to calculate the value. I think this story is just stunning for the size of the damages.

RAJAGOPALAN: I had to read this a couple of times just because your paper was also widely reported in the news. I was like, this is impossible. Like some stupid journalist has read this and they don’t know how to do math, and they have probably misplaced the decimal point, because it was just unimaginable to me that it could be almost 5% rise in human mortality. Then, of course, using the calculations of the value of a human life, which are contestable, you look at close to 70 billion annually, right? That’s a huge cost.

Couple of questions on this. Human medicines going out of patent or becoming extremely cheap and generic form and ending up in the veterinarian pool, this is a problem we already know not just in the developing world, but everywhere. We have a huge problem of antibiotic resistance. Most of it is a consequence of antibiotics being given to cattle by dairy farmers. We know this problem. What was it about diclofenac that this could not have been estimated, or the people who started using it just didn’t imagine that one runs kidney tests on vultures? How does this even happen in the first place?

SUDARSHAN: I think the latter. It’s a widely used drug even today for humans. It wasn’t doing cattle any harm. One of the points we make in this paper is that if you want to take these spillover effects on nonhuman species seriously from an instrumental point of view, and I think the paper suggests that you absolutely have to, then one of the things you want to do is when you introduce new chemicals, you have to look at the lifecycle of who’s going to be exposed to this. Now, obviously diclofenac will not be used on cows if it kills cows instantly. It’s not that it’s not tested on anything. It’s tested on the cows certainly.

There are many species out there. The one point we make in the paper is that ecologists have talked about and identified what they call keystone species, which essentially are species that do something in the ecosystem for which there’s no redundancy. Sometimes people will say, “Okay, well, there’s so many birds and animals and plants, we can’t possibly test it on everything.” We’re saying, “No, you can’t test it on everything, but perhaps we should test chemicals on the keystone species and not just human beings.” That seems like the sort of policy that you could potentially implement.

RAJAGOPALAN: That’s a very economist way of expressing the problem, right?

SUDARSHAN: Yes.

RAJAGOPALAN: Like, it would be lovely to save everything, but we’ve got to do some cost-benefit and diclofenac is extremely useful and helpful and increases economic benefits and values, but we’ve got to have some offset and truly understand the cost.

SUDARSHAN: Yes. If you had done the offset, you wouldn’t have used diclofenac because there are substitutes [and] because the vultures are disappearing. In general, a keystone species you would probably want to keep around always. One of the problems is you can’t wait till it’s extinct to discover what they were doing because it’s too late then. I don’t think it’s asking a lot to say that maybe human beings should expand what we care about from our pets and our livestock to the keystone species, which might be an earthworm, which might be a vulture, because there’s no coming back from their extinctions.

RAJAGOPALAN: One of the reasons you go through this enormous exercise, and it’s a very sophisticated paper when it comes to measurement, to illustrate exactly how costly it is and what are the consequences. How does one use this for policy? Is the immediate impact just, “Hey, let’s test across the board for the keystone species”? Is it, “Let’s price diclofenac higher such that it’s not viable to use it for cattle”? Is it to ban it all together? How does one even wrap their head around what could be a policy solution to the problem?

SUDARSHAN: Diclofenac has been banned in India once it became apparent that’s why it had wiped out the vultures, as a conservation initiative. Then the motivation for whether you ban it, how much money you spend on it, how much you care about it, is a little lower because it’s mostly about, “I don’t want these birds to disappear,” rather than the instrumental motivation of “it’s costing me billions of dollars and so many additional human deaths.”

RAJAGOPALAN: Including the farmers’ families and their children. It comes back to them.

SUDARSHAN: Yes, it’s switching the frame a little bit from, say, the project tiger. The tiger deserves to exist. The vulture deserves to exist. It’s switching that to a more instrumental framework.

RAJAGOPALAN: We need the vulture.

SUDARSHAN: We need the vulture. It’s not just about us doing them a favor. I think it’s going to be tough to get the Indian vulture back quickly. They can live for up to 70 years. They mate for life. They have one egg at a time. They have a gestation period of several months. It’s a bit like human beings. They’re not going to breed like rabbits. Having said that, we have not built out incinerators to deal with this problem in decades either, but there are vultures elsewhere in the world. There are vultures in Africa doing much the same thing. In parts of Europe, where it’s less of a problem.

The thing in India is that we don’t eat cows. There are many countries with livestock, most of them use the meat. Africa has livestock that are still around. Even the vulture, it’s relevant going forward. The broader implication, I think, is this comment on we need to have systems in place that examine the impacts on nonhuman species. We can argue about which ones—

RAJAGOPALAN: Which ones, yes.

SUDARSHAN: - whether it is keystone species, but that’s a separate argument. The impacts can be very, very large and essentially irreversible. At least not in any simple way. We are often used as economists to thinking in terms of marginal effects. Of course, the marginal effect of a bird disappearing is zero. Extinction is one of few things—

RAJAGOPALAN: Extinction is different. Exactly.

SUDARSHAN: —that is not marginal.

RAJAGOPALAN: You’ve hit the nail on the head. That’s the extraordinary thing about the paper, because we don’t expect things to collapse so quickly such that now you’re no longer worried about this getting slightly costlier on the margin. It’s just you either have this protection from the vultures or it just completely disappears in particular places. Technically, the paper’s quite lovely. Looking at places that would’ve never had very high vulture populations in the first place with those that always did and then comparing them before and after the diclofenac. This was super fun for me, Anant. Thank you so much for giving us your time.

SUDARSHAN: Thank you, Shruti.

RAJAGOPALAN: I had lots of fun reading all your papers again, and I hope you’ll come back and chat more. I don’t see any of these problems going away anytime soon, unfortunately. Hopefully, you have all of your work cut out for you, and you’ll also come back and chat with us a little bit more.

SUDARSHAN: Thank you so much for the invitation, Shruti.