Latest / 80,000 Hours Podcast / #170 – Santosh Harish on how air pollution is responsible for ~12% of global deaths — and how to get that number down

Transcript

#170 – Santosh Harish on how air pollution is responsible for ~12% of global deaths — and how to get that number down · 80,000 Hours Podcast

• Santosh Harish: The way this mechanism has been set up in India has relied on what are called command-and-control instruments. Basically, there is a standard, and you’re meant to be under that standard. Somebody comes and measures this from time to time. If you’re above that standard, it is a criminal offence, and therefore there’s going to be a lawsuit filed against you. You could land up in jail and pay a fine of some kind. In practice, the legal system in India is backed up: most cases take years and years and years. The actual compliance against these standards is quite poor, the pollution control boards are understaffed, and there is no real mechanism by which they can go after these many industries that are flouting the law. As a result, for the most part, the regulatory regime just sort of fails. So if a particular industry is found to be noncompliant, there is a gentle slap on the wrist, there is some kind of polite correspondence where the regulator writes to the industry and asks them to explain themselves, and it sort of ends with that. There is very little action taken. Rob Wiblin: Hey listeners, Rob Wiblin here, head of research at 80,000 Hours. Air pollution kills and harms people far more than we appreciate, in rich and poor countries alike. It probably kills ten times as many people as malaria. Tuberculosis is the single communicable illness that kills the most people each year and air pollution kills maybe 4 times as many as that. And much of the harm isn’t about deaths, it’s about worse health, brain fog, and stunted potential while people are alive. So I was really excited to get to talk to Santosh Harish of Open Philanthropy, who has taken on the challenge of using philanthropy to reduce the harm done by air pollution in India, something that almost nobody else in the world is attempting to do. We talk about: How bad air pollution is for our health and life expectancy The different kinds of harm that particulate pollution causes, and how strong the evidence is that it damages our brain function Whether it was a mistake to switch our attention to climate change over air pollution Whether most listeners to this show should, like me, have an air purifier running in their house right now Where air pollution in India is worst and why, and whether it’s going up or down Where most air pollution comes from The policy blunders that led to many sources of air pollution in India being effectively unregulated Why indoor air pollution packs an enormous punch The politics of air pollution in India How India ended up spending a lot of money on outdoor air purifiers The challenges faced by foreign philanthropists in India Why Santosh has made the grants he has so far And plenty more Without further ado, I bring you Santosh Harish! Rob Wiblin: Today I’m speaking with Santosh Harish. Santosh leads Open Philanthropy’s grantmaking in South Asian air quality, which he’s been doing from his office in India since the start of 2022. Prior to that, he was a fellow at the Centre for Policy Research in New Delhi, working with the Initiative on Climate, Energy and Environment. Before that, he worked as the associate director for research at the India Center of the Energy Policy Institute at the University of Chicago and did postdoctoral fellowships with J-PAL South Asia and Evidence for Policy Design. Back in his relative youth, he studied a bachelor in technology at the Indian Institute of Technology in Madras, before doing a PhD in engineering and public policy at Carnegie Mellon University in Pittsburgh. Thanks for coming on the podcast, Santosh. Santosh Harish: Hi, Rob. Thanks for having me here. It’s a pleasure. Rob Wiblin: I hope to talk about where the most air pollution comes from, as well as the cheapest ways to reduce it. But first, give it to us straight: How bad is particulate air pollution? Santosh Harish: Air pollution is the single largest environmental and occupational risk factor to public health globally. Per the Global Burden of Disease estimates, it accounts for something like 6.67 million deaths a year, as of 2019 — which, to give context, is about 12% of all deaths globally. Not all of this is particulate matter. Particulate matter is the vast majority of this, but a small fraction of this is what’s called ground-level ozone. But yeah, it’s pretty bad. When I started working on air pollution, which is about roughly a decade back, some of these high numbers were hard to come to terms with. It almost seems implausibly large, intuitively, because you’re like, it’s presumably bad for your lungs or something, but could it really be this bad? So the thing about air pollution, which makes it so harmful — in particular particulate matter air pollution — is that particulate matter is not a single substance. It’s a cocktail of various things that are in the air that just happen to be finer than 2.5 microns in diameter — which is a tiny fraction of how thick your hair is. It is composed of a variety of chemical substances, some of which are relatively harmless, some of which are extremely toxic. So it could be stuff like soil dust, which is naturally occurring, or sea salt — which are likely to be not particularly harmful. And then there is stuff like lead and other heavy metals that are suspended in the air. There are inorganic compounds like sulphates and nitrates which originate from vehicle emissions, from coal power plant emissions and so on. So it’s a variety of different things. Because these particles are as fine as they are, they are able to enter the lungs, enter the systemic circulation — and then basically these various things that have no business being in our body can travel to different organs and cause a variety of different harms. Rob Wiblin: Yeah. I guess we would all have the intuition that smog is not going to be great for your health. But I suppose the thing that would be possible to intuitively miss is that there’s a whole lot of different chemicals, different kinds of tiny particles that end up suspended in the air. We talk about PM2.5 — we’re going to talk about that a lot; that’s particles that are smaller than 2.5 microns. I’m not sure how to make it intuitive, but they’re small enough that they can cross over into the lungs and into the bloodstream, and then out of the bloodstream into other organs in the body. So they can end up in your brain, end up in your liver, end up in your intestines, end up in children developing inside a pregnant woman. So they can do damage to many different parts of your health all at once. My impression is that people in recent decades have come to think that air pollution is worse than they used to believe. There’s been more research pinning down what effect air pollution has on many different health issues, and this has caused people to realise that the issue, the impacts, is far more systematic than what we’d realised. Is that right? Santosh Harish: Yeah, I think so. I think it’s objectively the case that with each passing year — sometimes it feels like with each passing month, or even week perhaps — there is new literature out there that seems to suggest that air pollution is somehow more harmful on a particular health outcome than we had previously thought, or it actually seems to affect something else that we hadn’t even considered earlier. For instance, there is a growing literature now on how air pollution affects cognition, and thereby affects productivity, affects a bunch of academic outcomes. And this is relatively new literature; this isn’t the basis by which the Global Burden of Disease, for example, thinks of the harms of air pollution. So these are new impacts of air pollution. And yeah, I think the literature has been growing, and growing at a fairly rapid click, which is quite alarming. Rob Wiblin: What is perhaps the most outrageous or emotionally grabbing example of air pollution to you? The thing that I have in mind is something that barely benefits the person who’s emitting all of this pollution, but it’s causing massive health damage to people in the nearby area. Santosh Harish: One thing that comes to mind is municipal waste burning that happens in many cities in the Global South. Basically, this is waste that gets collected from people’s homes, and instead of being transported to a waste management facility or a landfill or something, gets burned at some point, because that’s the fastest way to dispose of it — which really points to poor delivery of public services. But this is ubiquitous in virtually every small- or even medium-sized city. It happens in larger cities too, in this part of the world. So I think that’s something that truly annoys me, because it feels like the kind of thing that ought to be fairly easily managed, but it happens a lot. It happens because people presumably don’t think that it’s particularly harmful. I don’t think it saves a tonne of money for the municipal corporations and other local government that are meant to manage it. That’s one example that comes to mind. I find it particularly annoying simply because it happens so often; it’s something that you’re able to smell in so many different parts of these cities. Another which seems downright evil to me is a whole bunch of industries that tend to not use the pollution control equipment that they have in their facilities already. And just basically dump the flue gas, as it’s called — the gas that gets emitted from the various processes in the industry — without the emission controls, in the middle of the night, when it’s not obvious and it can’t be detected as easily as it would in the day. And this is basically to, again, save what I suspect is change in terms of maintenance and operation costs of this equipment. You have the equipment and there are these standards. So that’s I think downright evil on the part of these industries. Rob Wiblin: Yeah. Over the last couple of years, I’ve started getting more and more emotionally incensed by air pollution. I studied economics and I read a lot of economists on Twitter and on blogs, and I listen to lots of podcasts that have economists on them. And it feels like over the last five years, economists have just been losing their shit more and more about air pollution — just getting more and more frustrated with how much damage seems to be getting done by this, and maybe how surprisingly unseriously it’s taken in public policy and just in general, maybe even by voters in some ways. I live in London, and every so often I see a car go down the street that’s obviously not in good nick, not in good condition, and it’s spewing fumes out the back. I just get, “Ah, these people are assaulting me as a pedestrian!” It’s so outrageous what’s happening. Why don’t the police do something? Why doesn’t something happen? I don’t think this happens so much anymore, but when I first moved to the UK, there were lots of boats along the Thames that had, I think, wood fires or just fuel fires inside the boats. And there were also houses that still had wood fires. I mean, this is obviously a very densely populated area in London. It’s crazy that they’re just burning fuel, producing these really seriously toxic fumes that are affecting everyone around. Anyway, that’s my rant about air pollution, of my intuitive reaction to it now. I imagine you probably have worse stories from your own experience. Santosh Harish: Yeah. Don’t get me started on the smoke-spewing vehicles on the roads. Again, it’s one of those cases where that almost certainly is a symptom of a highly inefficient vehicle that’s still driving on the road, and a newer vehicle would likely save a tonne of fuel and also cause much less harm. It seems to be the kind of thing where it’s in everybody’s interest that this vehicle does not exist on the roads. I’m actually optimistic about this: I do think that the number of these vehicles tends to reduce quite quickly, and that these incentives not to be running them ultimately dawns on whoever owns these vehicles. Rob Wiblin: Yeah. OK, let’s just clarify something quickly. You’re working on South Asia air quality, and I guess South Asia includes India, Pakistan, Bangladesh, Sri Lanka, Nepal, Bhutan — but you’re mostly focused on India, because that’s where you spend a lot of time and that’s where you know the most? Santosh Harish: That’s right. I am based out of India. I have primarily worked on air pollution in India. As a grantmaker as well, this is my relative comfort zone. It’s also the case that air pollution work in India is ahead of the countries in the neighbourhood — just a significantly larger number of organisations working on it; the scientific understanding of air pollution is more advanced — so there are obviously more grantmaking opportunities here as well. But yeah, by “South Asia,” we are primarily funding work in India, Pakistan, Bangladesh, and Nepal. These four countries tend to also feature in the top 10 when it comes to the highest population-weighted annual concentrations of PM2.5, or the health impacts of it and so on. So this is arguably one of the most polluted parts of the world. Rob Wiblin: And in terms of different pollution types, obviously there’s particulate matter, so that’s the majority of the harm that’s been caused. There’s also ozone. Are there other things that people should have in mind? Santosh Harish: Well, there are others, like carbon monoxide for example, which is also an incredibly harmful pollutant. But ground-level ozone and PM2.5 really dominate — both in terms of just how harmful they are, but also how widely prevalent they are at levels that are harmful. Rob Wiblin: What are the actual particles? I’d kind of assumed on some level that they were ash or that they’re this kind of thing that you’d get if you burned wood. So that’s got to be part of it. But are there other things as well? Santosh Harish: That’s definitely part of it. So there’s the finer version of the ash that you see when you burn something. But the other constituents of it are things like inorganic compounds, like sulphates and nitrates. So sulphur dioxide is principally emitted by the combustion of coal in power plants, in industries which burn coal for a variety of different reasons; nitrogen oxides are primarily emitted by vehicles burning diesel or petrol. And these then react with other gases that are in the atmosphere to form sulphates and nitrates, which are particulate matter. The other constituents of it are there are a bunch of what are called organic compounds, which again get emitted from a variety of different sources. There are trace heavy metals, which are particularly harmful — so this could be lead, for example, being emitted from the improper smelting or disposal of lead acid batteries and so on. So it’s a wide variety of different things which just happen to be finer than 2.5 microns. And just to help get a sense of this, a micron, or a micrometre, is a millionth of a centimetre — and therefore something like a tenth of the thickness of your hair, if that helps. Rob Wiblin: So let’s maybe come back and analyse a little bit more carefully how large the scale of the damage is being done by air pollution, because on this show we don’t just settle for something’s “really bad.” We want to kind of quantify it a little bit more specifically. So far we’ve talked about the issue that it causes people to die. So it causes a whole bunch of mortality. What are the different ways in which it causes people to die or suffer directly ill health? Santosh Harish: The evidence is strongest when it comes to a bunch of heart and lung diseases. Some of the principal sources of death and illness that can be attributable to air pollution seem to be heart disease, lung cancer, something called chronic obstructive pulmonary disease — which is basically a chronic lung infection, which, once you have it, it doesn’t go away; it cannot be reversed — and there are lower respiratory infections like pneumonia, strokes, type 2 diabetes. I think those are some of the big health harms. So Global Burden of Disease I’m sure your listeners are familiar with, but just in case, this is an effort that’s anchored by the IHME, the Institute for Health Metrics and Evaluation. Basically, these are periodic, massive scientific enterprises that bring together the who’s who of public health researchers to try and synthesise evidence on a bunch of different risk factors. So that’s what they do for air pollution as well. It tends to be a relatively conservative effort, in that they do try and restrict to the sources of illness and death where the evidence is relatively stronger, although they are pretty transparent that there are significant uncertainties involved in these estimates. So until 2017, which is two iterations back, the air pollution harms were basically restricted to just five diseases, the ones that I mentioned earlier: heart disease, lung cancer, COPD, lower respiratory infections, and strokes. They have since added type 2 diabetes, and an emotional one, which tends to get a lot of attention when these reports get released, is they also included impacts on neonatal deaths in GBD 2019. This is basically infants that die within the first 28 days after their birth because of the exposure of the foetus within the womb to air pollution. And they estimated something like half a million neonatal deaths per year that might be attributable to air pollution, which is a striking number. Basically, these are infants that are likely born preterm or possibly with lower birth weight than ideal, and as a result are more vulnerable and therefore there’s a higher chance of death within the first month. Which, again, I think is a striking and, if true, alarming number. Rob Wiblin: Yeah. So you mentioned earlier that air pollution is a major cause, or the major cause of around 12% of deaths globally, which is a large fraction. It’s a very large number. To try to make it a bit more intuitive, let’s say that the air in London went to being perfect: how much more life expectancy would people in London get? And then if, say, the air pollution in Delhi disappeared completely and they had perfect air, how many more years of life or years of healthy life might someone expect to enjoy in Delhi? Santosh Harish: Yeah. So at the moment, going back to the GBD numbers, in the UK more generally, and not just London, the death rates attributable due to air pollution is something of the order of 11 per 100,000. So 11 deaths per 100,000 that can be attributable to it. And in India, that number is a factor of 10 higher: something like 98 per 100,000. Rob Wiblin: Per 100,000 people per year? Santosh Harish: Per 100,000 people. Not 100,000 deaths — 100,000 people. Rob Wiblin: Got it. Yeah. Santosh Harish: And the difference between the levels of air pollution is maybe like eightfold on average. Maybe slightly less than eightfold — so somewhere between 7x to 8x. It does not scale linearly. One of the things that we have learned in the last few years is that air pollution seems to be harmful at levels that were previously thought to be safe. So much so that the WHO reduced its guidelines for what classifies as clean air from 10 micrograms per metre cubed of PM2.5 to something like five — which is much lower than regulatory standards in any part of the world at the moment. And that’s a reflection of research showing that air pollution could actually still be quite harmful at those low levels. The shape of what’s called the dose-response curve — basically, the dose of your pollution exposure, and therefore the response in terms of your health impacts — seems to be curvy linear. Basically, at lower levels of pollution, a marginal increase leads to a higher increase in relative risks than at higher levels of air pollution, so the relative risks flatten out as the pollution increases. So beyond a point, if I could swear, you’re basically screwed; it doesn’t matter. That seems to be the shape of the relative risks. But there are multiple sources of nonlinearity here when we think about how the air pollution risks manifest. So in low- and medium-income countries, because there are other sources of public health risk — malnutrition, other sorts of pollutants, lower incomes and therefore lesser ability to access healthcare, things like that — they multiply with each other, and therefore it becomes more harmful. I don’t know if I gave a straight answer to your question. Rob Wiblin: Well, I think when I was doing background research for this, I saw some map that was attempting to quantify this. I can’t remember exactly which source it was, but I think it was suggesting that in the UK, in cities, people were losing one to two years of life, maybe, from air pollution in total. Whereas in India, at least in the more built-up areas in the north, where the air pollution is worse, I think it was estimating six, seven, eight years, maybe. I’m not sure. Does that seem plausible, or do you think I’ve misunderstood? Santosh Harish: I think you’re referring to the Air Quality Life Index report that the University of Chicago takes out. It is possible that these numbers are overestimates. At least compared to the Global Burden of Disease–style synthesis of health impacts, this certainly seems to be on the higher side. An alternative estimate, for example, from the GBD’s table, is that Indians might be losing something like three years of their life expectancy due to air pollution on average, and obviously that number is higher in north India. But that’s almost half of what the AQLI report tends to estimate. Rob Wiblin: Yeah, well, these things are hard to estimate. So actually, only having a twofold range between these different sources… Santosh Harish: Yeah, exactly. There are massive uncertainties still. Rob Wiblin: OK. So three to six, maybe, in India could be a plausible range. And then in the UK you would expect that maybe it’s a fifth or something like that. Looking at that map, there were relatively few places where people live where the air was really clean by this standard. I guess in Norway there were some places that were really good, and if you’re far away from population centres, then you’ll be OK. But it seemed like the great majority of people are breathing air that the WHO at least would say is not healthy. Santosh Harish: That’s right. Especially when you compare it against the updated guidelines of the WHO: it basically means that like 99% of the world’s population breathes air that is insufficiently clean. Rob Wiblin: Yeah, if you want to breathe clean air, you have to move to the fjords of Norway. OK, so that’s the classic life expectancy issue. But there’s this whole other cluster that’s been becoming more and more prominent, which is these effects on cognitive performance and mood and things like that. I’ve read that there’s a bunch of studies suggesting, as you were saying earlier, that it harms school performance to have polluted air. Makes you do worse at school, makes you do worse at your work, causes you to drive worse and make more mistakes. Potentially also affects mental health, like makes people more likely to have depression or commit suicide and things like that. Is there any way of summing up the evidence on that? Because at the moment there’s just so many different papers looking at specific different outcome measures, it’s a little bit hard to aggregate it into a simple picture for me. Santosh Harish: Yeah, I can give you my take, which is possibly more sceptical than others. There is a body of work that’s emerging that seems to be pointing in a similar direction. There seems to be evidence to suggest that air pollution harms you and your productivity in a manner that is distinct from its harms on health: it’s not productivity that is getting harmed because you’re falling sick more often, but productivity that is being harmed because of cognitive losses. So the magnitude of these impacts — I don’t claim to be an expert on this stuff, and in some sense I’m a consumer of these results — I suspect, will still have significant uncertainties and publication biases and things like that. In some sense, the reason I am more comfortable with the Global Burden of Disease–type estimates is that it reflects a larger body of work. And although the causal identification with some of these public health papers may not always be up to scratch, they’re relatively transparent about what they do and don’t, and there are just more papers from different parts of the world, at different levels of pollution, which seem to back each other up. And I think on the cognitive and productivity stuff, it’s still a more nascent body of work, which is still pointing in the same general direction, but I just don’t know if there has been enough of it. So probably harmful to your cognition in addition to the impacts on your health — on heart disease and lungs and so forth — but my own sense is that we don’t yet know the magnitude to the level of certainty that we would like. So for example, in our grantmaking, this isn’t something that we actually factor into our cost-effectiveness calculations: so far it’s been restricted to health impacts alone of prolonged exposure, and I imagine that will be the case in the foreseeable future. Rob Wiblin: Yeah, that’s really great, because you’re preempting where I was about to go next, which is probing how confident we can be about these different papers suggesting that air pollution is having these different effects. Listeners to the show will be familiar with the general idea that it’s not uncommon for academic literatures to produce really striking, amazing results, that then over time you end up with more qualifications and you get conflicting results, and then you’re not as sure as you initially were. And then when you do higher-quality studies, sometimes the effects can disappear completely. So I slightly worry that this could happen with some of these more cutting-edge or speculative impacts that air pollution might be having. That said, there is almost certainly something going on here, because one thing that’s very nice about air pollution is that it’s possible to get these quasi-random studies, basically, because of the weather effects: the weather is this external exogenous source of random variation that causes some days there to be much more pollution than others. So inasmuch as you’re studying the direct, immediate effect of air pollution on someone rather than its cumulative effect over a lifetime of exposure, I know that there are these papers that economists have done where they’ll look at people at a school, for example, that is, say, to the west of a road that produces a lot of air pollution because it’s a highway. So they can compare days when the wind is blowing west — so the school is getting exposed to lots of air pollution from this highway — to days when the wind is blowing east, when the air is a lot cleaner because the pollution isn’t being blown over from the road. And they can see from test results, from standardised tests and so on, that the students do worse, they do materially worse on standardised tests on days where the air pollution is higher. Of course things could go wrong, and your result could end up being wrong, but that’s a relatively credible kind of research design. Do you want to comment on that? Santosh Harish: A couple of my priors: one, it’s almost certainly the case that most of the health harms due to air pollution is because of prolonged exposure as opposed to short-term exposure, even to extraordinarily high levels. So that’s probably true. And therefore the short-term exposure, and the studies that estimate its impact on health or cognition, it’s likely to be a relatively small fraction of the larger pie. It is true that there are these quasi-random variations in pollution levels that allow you to study the impacts of the short-term changes. One of my colleagues at Open Phil — who’s with the Cause Prioritization team, who I think has thought about this stuff more carefully than me — Lauren Gilbert, is extremely sceptical of weather as an instrumental variable, and believes that the current understanding in the econ literature on causal identification has basically set aside weather as a reliable source of exogenous variation. Rob Wiblin: So just to clarify what you’re saying. “Instrumental variables” is this method where you use something like the wind as this source of random exogenous variation that’s not associated with other things, hopefully, and then see whether that affects test results. And then you hope that the only way that the weather could affect test results would be via it bringing pollution to the classroom. Are you saying that your colleague is sceptical that this is actually a sound research design, and that economists have become more sceptical of it? Santosh Harish: That’s right. And as a result, some of these papers that are using weather as the principal IV — instrumental variable — may not eventually stand up to scrutiny, and therefore might either have errors that are larger than is estimated in the paper. Either the estimated impacts become more muted, or they’re just unable to make a claim once you correct for some of these measurement errors. So again, that’s the kind of stuff that makes me a little sceptical of the emergent literature. Once again, the same cautionary warning that I mentioned earlier: I do not claim to be an expert in causal identification methods and the strengths and limitations of IVs. I’m just a consumer of this work. Rob Wiblin: But you think we should take it with a pinch of salt? Santosh Harish: Air pollution seems to be a harmful thing. “What can we do about it?” is where my realm of experience begins. Rob Wiblin: OK, great. We’ll stick up some links to people who are analysing this kind of question, trying to figure out how much we can trust these new seeming research results on air pollution, for people who want to look into that more. And yeah, we’ll have to get someone who’s working on that research in particular to dig into it in more detail at some future time, because I find it super fascinating. You mentioned earlier that you had kind of declining marginal harm with air pollution. That is to say that the most damage is done by the first bits of air pollution that you’re exposed to, and then at some point you’re just screwed: having a little bit more or a little bit less doesn’t make that much difference. That’s a bit counterintuitive to me, because often you have this model that the body can handle some level of abuse, but then beyond some point it’s no longer able to repair the damage that’s being done. And if that were the way that things are working, then you’d expect to have increasing incremental harm from each extra bit of air pollution, because the body would be overwhelmed. But are we reasonably confident that actually it goes the other way? That it’s the first units that are the worst? Santosh Harish: Yeah. To clarify, even at high levels of air pollution, there is likely to be some incremental harm. It’s just that the slope of that graph is much… Rob Wiblin: It levels off. Santosh Harish: Yeah, exactly. It’s steeper at the initial levels. It is counterintuitive. It also means a few things. One is that for places like Delhi — which have abominable levels of air pollution throughout the year — for policy to be able to bring it down to levels where the health harms are brought as close to zero, they don’t need to just improve by a little bit; they need to improve by a lot. Because most of those gains are actually at a level that’s maybe like a third or a fourth of the levels that you see right now. So that makes the challenge in many ways steeper. It also leads to this potentially interesting conundrum on where you want to prioritise your efforts. For example, there’s this paper which basically poses a question on whether you ought to be making blue skies bluer: do you actually then focus on cities that are closer to that inflection point, and basically look to make them significantly cleaner and more livable? It seems like a very EA sort of dilemma to ponder over. Rob Wiblin: Yeah, the idea there is that if it’s the first bit of air pollution that does most of the damage, then what you want to do is find places that are reasonably clean, not that bad, and then get them to be really clean. And that is a little bit counterintuitive as the most cost-effective thing to do, because usually the most cost-effective thing to do is to focus on the people who are suffering the worst of something, rather than people who have the least effect. Santosh Harish: Yeah, for better or for worse, right? It is a fact though, that eventually the way policy seems to prioritise where resources are to be spent ends up being places like Delhi, where the levels of pollution are just so high that they’re obvious. It becomes more politically salient and it gets a bunch of media attention. So I don’t know if this is actually something that is actually a dilemma that faces policymakers. It possibly ought to be, but yeah, for better or for worse, I don’t think it’s actually a serious question that government folks tend to tackle. Rob Wiblin: Yeah. I’m planning to interview someone from the Lead Exposure Elimination Project later in the year, so lead is on my mind. Do you know if air pollution in India or elsewhere ends up inadvertently exposing people to lead — which we know is just extremely bad for people’s health and cognitive development? It’s really one of the worst poisons that we’ve ever systematically been exposing people to. Santosh Harish: Yeah. One of the mechanisms by which we get exposed to lead is through air. We are also exposed to it through food, through spices and all that. A couple of points. One is that trace levels of lead would technically be part of PM2.5 — although, because lead is as harmful as it is, regulators tend to measure that separately anyway, and have separate standards for it anyway. There are some common sources of both lead pollution and particulate matter. For instance, improper recycling of lead-acid batteries tends to make both worse due to different dynamics. So there are relationships between the two. I’m unaware of any other literature that seems to suggest that air pollution makes the impacts of lead worse. Rob Wiblin: When I was growing up in Australia, there was leaded petrol. So I think cars when I was a kid were spewing out presumably lead, among other things, and I was just breathing it in. I guess lead petrol is not in India, and not in most places these days, but it might still be in the fumes of industrial uses right? Santosh Harish: That’s right. The reason I didn’t mention that is that fortunately, it seems to be something that’s largely behind us. But yes, it’s possible that that’s still a factor from industries. Rob Wiblin: Yeah. Are there any surprising things that people should know about air pollution, or misconceptions that people often have that are worth clearing up? Santosh Harish: One that is actually a significant hindrance to effective policy in India and similar countries is that air pollution is assumed to be an urban problem. This was certainly true in big industrial cities and so on, where air pollution started becoming visible and salient — thinking of London, Pittsburgh, Los Angeles. In places like India, though, that’s just not true, because rural air pollution can be significant. In fact, on average, rural exposure is not very different from urban exposure. One of the largest sources of air pollution exposure in India, in Pakistan, a whole bunch of other countries, would actually be the household burning of solid fuels: wood and dung cakes and things like that. So it’s actually not at all an urban issue alone, and historically, it has been treated as that. For example, there are no rural air quality monitors in India. It’s sort of a chicken-and-egg thing. Until recently, the response to why there aren’t monitors there was that we know villages are clean; that’s where you go when you have respiratory problems and so on, right? So it neither gets measured — because it’s assumed to be not a problem — and because there isn’t any measurement to suggest otherwise, that never really gets updated. Because there are alternative sources of data now — satellite-data-based estimates of air pollution, for example — I think there is growing evidence that rural air pollution can be substantial, and therefore there has been a growing demand for air quality monitoring in rural areas. Another misconception that in some ways we touched upon is that there are sort of “safe” levels of air pollution; it’s only the truly apocalyptic levels — that one sees, for example, in winters in places like Delhi and so on — that’s what harms you. That, unfortunately, just doesn’t seem to be true. Impacts have been detected at much lower levels that were previously considered safe. An unusual type of misconception that’s sometimes popular in government circles is that air pollution is something that you can build immunity to. Rob Wiblin: Oh, wow. I’ve never heard that one. Santosh Harish: Yeah, it’s something that’s as ridiculous as that sounds. Rob Wiblin: I guess it’s like if you go out in the sun a lot, maybe you get a tan, which slightly helps you to not get sunburned. But I don’t imagine your body can do that with particulate pollution. Santosh Harish: Yeah, so there is this insistence by I certainly don’t think all regulators, but by some, that Indian lungs, for example, are just better able to handle air pollution than, for example, American lungs. Because we’ve been exposed to it over a period of time. Rob Wiblin: You’re used to it. Santosh Harish: We have gotten used to it, so we don’t routinely fall sick when the levels of air pollution are high. And when somebody’s visiting Delhi or something, you feel it — you feel the difference quite viscerally. So that clearly points to Indian lungs being better adapted. That’s of course nonsense, but it’s unfortunately a persistent myth that has been hard to shake off. One of the implications of that also has been a general scepticism towards the public health impact estimates from sources like the GBD here. So there’s an insistence that we need more indigenous research, studying health impacts in Indian populations, and that these extrapolations from other parts of the world are just not reliable. There is a truth to it. I mean, it is an understudied topic. There just isn’t enough high-quality cohort studies and so on in the public health literature on air pollution at elevated levels of concentrations. There is some that’s emerging from China, but those are still levels that are lower than what one sees in north India. So there is some truth to there being a significant research gap. I am, however, not as persuaded that that then means that the absence of evidence is evidence of absence. I think I’m saying that right. Rob Wiblin: Yeah. I mean, people do differ around the world, but I think one fortunate thing about humanity is that — from a biological, medical, basic-physical-requirement point of view — humans everywhere are more similar than they are different. And I guess if we were finding that air pollution was doing people a lot of damage in America, I would tend to assume by default that could also be true in India, rather than saying, “Who can say? We need to prove it here as well.” But I don’t know. Santosh Harish: Yeah, I think those are my favourite misconceptions. There might be others, but the immunity thing I think is a particularly fun one. Rob Wiblin: An argument that I came across when doing background research for this was the observation that in recent decades, the main environmental campaign, or the most prominent environmental campaign — at least in rich countries, or at least in English-speaking countries that I’m familiar with — I think by a decent margin, has been preventing climate change. And it’s kind of famously hard to get people and countries to do stuff about climate change, because if you emit carbon, the damage that does is spread globally, it’s spread everywhere, not just in your country. And it’s not just this year; it’s spread over decades or centuries. And people have been pointing out, given that air pollution seems to do so much harm to people’s health — even in the US or the UK, it’s still really quite substantial impacts on people’s health relative to the things that we tolerate elsewhere — that it’s funny that we’ve focused on trying to get people to do this thing that’s a really hard sell because of its global public good nature, when we could have instead been focusing on particulate pollution. Which is kind of the reverse, in that you benefit very immediately. Because the air cleans itself up, particulate pollution falls out, and you could be healthier really quickly if you closed up a coal plant nearby you. So it’s really concentrated in time and it’s much more concentrated geographically, so that a given political unit that’s voting on these decisions could benefit itself, and doesn’t have to worry about the benefits spilling out to other countries. And simultaneously, if you were closing coal plants to stop particulate pollution and replacing them with renewable energy or something else, then that would also have happened to have helped with climate change — a very large amount, potentially. So maybe, even if all we cared about was carbon emissions, we would have done better to set that aside in terms of the public messaging and continue the environmental campaigns that were extremely prominent in the ’60s and ’70s around clean air and clean water — and sell that to the public, and sell that to voters and politicians as a way of encouraging people to switch over to cleaner power. I know this isn’t your specialty, but what do you make of that line of argument? Santosh Harish: A couple of thoughts. One: in some ways it’s understandable that this happened — the focus of the environmental advocacy groups pivoted from local air pollution to climate — simply because places like the US and countries in Europe have made significant progress on air pollution over the last few decades. The levels are objectively lower. Therefore I guess it’s harder to animate folks into thinking about the health harms, which is where the updated literature now becomes increasingly important. If it actually turns out to be the case that what we thought were safe levels are actually not safe, then there’s clearly an argument to reinvigorate that strand of environmental advocacy. In places like India, though, I think civil society groups have actually been sensitive to this this entire time. The fact that air pollution and climate change share several similar sources, that similar types of actions can lead to benefits along both those dimensions, I think that’s something that folks have grasped and have actively been working with. So I don’t necessarily think that that’s a correction that, for example, environmental advocacy groups in India ought to be doing. In fact, even in terms of the literature on this work, there’s this expression called “co-benefits.” There are folks — for example, Navroz Dubash, whom I used to work with in the past while I was at the Centre for Policy Research — who strongly advocated for countries like India to adopt that lens of prioritising actions based on co-benefits such as improved health, improved quality of life in cities through better public transport infrastructure, and things like that, and not necessarily think about mitigating global greenhouse gases. And therefore using that lens even in international negotiations and so on. So I think in the West it’s on the one hand understandable why this pivot happened. I think there’s also increasingly a case to pivot back towards thinking of it from a local health harms lens. And in places like India, I think we’ve actually been approaching it from this lens. Rob Wiblin: People have already been alert to this. Speaking of health harms in richer countries, or I guess countries that have made a lot of progress on air pollution: The majority of listeners live in the US, Australia, UK, and some in continental Europe; how should people think about and approach air pollution in their own lives? How worried should they be? For example, we’ve had air purifiers in the office here and in my living room for the last couple of years. When I was preparing for this interview, I bought another one for the bedroom to clean up the air that I’m breathing while I’m sleeping. Is that a sensible move for people to take in general, selfishly? Santosh Harish: Yeah. I think defensive measures against air pollution is an eminently sensible thing to do. I don’t necessarily think it’s a selfish thing at all from a public policy lens. This is actually an area that I’ve sort of struggled with, as to what governments should do, especially in places where the pollution levels are much higher: How do you think about encouraging people — typically the elite and higher-income folks — to install air purifiers, and in some ways look for private solutions to the problem of air pollution rather than the harder politics of mitigation. But yeah, I think it makes tremendous sense for you to purchase that additional air purifier. Air purifiers work great. It’s tech that works indoors. Rob Wiblin: I wasn’t saying “selfishly” in a bad sense; I just mean from a personal point of view, is it worth the cost? And of course, even someone who can afford as many air purifiers as they would like eventually has to go outside. So it’s not a very systematic solution to the problem, and it’s not the most efficient solution to the problem, I think, given that we have lots of options potentially for reducing air pollution. One thing that I’ve noticed is that there’s all these websites online where you can just look up air pollution and what’s the quality of air in your area right now. And there’s quite a lot of monitors in London. Something that I’ve noticed when checking that regularly is that it’s incredibly local: that in London you can have one suburb that’s really dirty right now, while another suburb adjacent has really quite clean air, and it will just change, flip between them. And that it also just varies enormously from day to day: some days the air is super clean, some days it’s super dirty. What’s going on? Why isn’t the air mixing more between different suburbs? Santosh Harish: So if you look at an annual average basis, the prolonged exposure standpoint, I’d be pretty surprised if some of that variation actually shows up. It’s definitely the case that within short-term windows, there are local hotspots. This is true in virtually every city that local sources — significant traffic junctions and so on — can make a significant difference to local air pollution. So hotspots are definitely a thing. But I think from a prolonged exposure standpoint, unless there is something weird like a polluting industry, I’d be surprised if there are actually very large variations locally. Rob Wiblin: On average, you mean? When you average over the year? Santosh Harish: On average, yeah. Exactly. It’s also one of those things where weather and climate play a huge role in terms of both variations over the course of the day and variations from one day to the next. And potentially, if there are cities which have significant microclimates, it’s very likely that will also translate into spatial heterogeneity in air pollution. Rob Wiblin: I suppose that there has to be something going on where, just depending on the weather, the air can kind of get trapped in a particular area — where it’s not moving upwards, maybe because there’s some temperature differential that’s preventing that, and it’s not blowing east or west. So the stuff’s not getting removed, and it can just build up in a local area. There were a couple of hours over the previous weekend… To give people some sense of the scale, ideally you should have below 5. The measure is micrograms per metre cubed. But I just think about it in terms of the number, because micrograms per cubic metre means nothing to me. So ideally the air pollution levels of PM2.5 would be under 5. I guess under 10 would be OK; under 30 would still be kind of all right. There was a while in my part of London where it was over 150 over the weekend — really very severe — and you can kind of see the smog out the window. And then it just crashed down to 20 again a couple of hours later. I don’t know, people might be surprised. One thing is when I got this air purifier, it actually shows that number on it, because it’s got a sensor that causes it to turn on or off. And that’s caused me to have this sense of how the air quality in the area is changing over time, over the day, and just how much it varies — which is something that I had no intuition for previously, and it’s made me a lot more alert to air pollution. I’m not sure that’s entirely good, because there’s only so much you can do about it, but I wonder whether that’s something that would be helpful. If people had all of these monitors, if they were seeing these numbers regularly, I wonder if it would be a much larger political issue, because people would appreciate what’s going on, that would say that it’s not just steam in the area — it’s actually pollution that’s harming their health. Santosh Harish: Yeah, I think London in many ways has been something of a leader on this. The density of the low-cost sensors that London has is extremely unusual, and is becoming a bit of a model for other cities to replicate when it comes to thinking about how one monitors air pollution. Traditionally, you basically had a few monitors — these very sophisticated, expensive monitors — in a few carefully selected locations. But since we’ve learned that there can actually be the variation from one neighbourhood to another, from one road to another, and so on — and that might actually be a reasonable thing to target the sources of that local variation, even in order to make a larger impact on the air quality in the city — there has been a growing call for how to integrate things like low-cost sensors, which didn’t exist even 10 years back. In a sense, we weren’t sure about the quality of the data that they produce. We have made significant progress in the last decade or so, have a reasonable handle on their strengths and limitations. So yeah, there’s this call to have hybrid monitoring networks complement the more expensive stuff, with just a very large number of low-cost sensors to try and measure this variation. I’m curious, did you figure out what led to that spike? Rob Wiblin: No, I couldn’t figure it out. It did cause me to Google “Are there fires in Europe?” and maybe the smoke from the forest fires is blowing up, but no, there wasn’t. I think it must have been something to do with the weather trapping pollution in the area. But it’s a little bit of a mystery. But I have noticed when I’ve checked it many times, that it’ll show you a graph of the increase and decrease over the day, and it is just super variable over a period of days; it will fly up and down all the time. Though that was the most extreme that I’ve ever seen. Santosh Harish: Yeah, 150 is objectively bad. Rob Wiblin: Yeah, it’s worse than most of India. OK, that’s about air pollution at a high level. Let’s zoom in and talk about India and Indian cities, which is the place you know the most about, and the thing that you’ve been specifically tasked to work on. Is there a way of describing how bad is air pollution in India in particular, and where is it the biggest problem? Santosh Harish: Yeah. So going back to the number that you just mentioned, WHO recommending that the safe level is 5 micrograms per metre cubed: the population-weighted average in India is something of the order of 80. And the most polluted parts of the country, in terms of annual average, will have levels that are closer to something like 120, 130. And the times of the year where air pollution in India makes headlines globally, we are talking about 24-hour averages that are of the order of 400, 500. So that’s one way of thinking about the levels of air pollution. Alternatively, something like 80% — 76%, I think, was the number that one of the papers had estimated — of Indians breathe air that is worse than the national standards. The national standards, by the way, are at 40 micrograms per metre cubed, so there’s a reasonable chance that the national standards are actually extremely lenient and that even if a city was compliant with it, that there was still breathing air that’s pretty harmful. But even relative to those national standards, more than three-fourths of the country breathe dirtier air. This is in particular a problem in Northern India. In fact, if you zoom out a little bit and look at South Asia, there’s something called the Indo-Gangetic Plain — basically, the plains that are in some ways defined by the rivers Indus, which is in Pakistan, and Ganga, that flows through India — basically spanning from the eastern Punjab province in Pakistan all the way across north India, covering several states — Rajasthan, Punjab, Haryana, Delhi, Uttar Pradesh, Bihar, West Bengal — and parts of Nepal, the plains in Nepal, the southern part of that country and all of Bangladesh. So this contiguous region, which is called the Indo-Gangetic Plain, has particularly high levels of particulate matter. In some ways it’s a large-scale version of what we just discussed, in terms of climate and topography and so on playing an important role. Essentially, the Indo-Gangetic Plain is sort of this basin that’s basically bound by the Himalayas in the north. And with winds from the Indian Ocean sort of pushing and keeping the air in that, therefore it becomes sort of a sink of air pollution, where in particular times of the year, it’s very hard for the pollution to escape, easily leading to the very large numbers that we just went over. This is also an extremely densely populated part of the world. Rob Wiblin: So if people look at a map of air pollution in India, they’ll see it’s an issue everywhere, but in the south, it’s not nearly as bad as in the north. There’s this kind of strip, as you’re talking about, this basin that runs from around Pakistan in the west towards Bangladesh in the east. It’s like this big stripe. And roughly how many people live in that area? Santosh Harish: About 700 million people live in this broader region. And yeah, Bangladesh is a densely populated country. Uttar Pradesh almost completely falls within this region, and it would basically be one of the largest countries in the world, if it weren’t a state of India and was a country. So I think that principally is what makes this a significant public health crisis. Rob Wiblin: An expression I hadn’t heard before, which I encountered preparing for this, is “airshed.” It’s an air analogy of a watershed. It’s this area where the air just keeps circulating within this circle, and air pollution in any part of it can potentially spread to any other part of this broader airshed. You’re saying this is created by the Himalayas in the north, blocking the air from moving north, and then you’ve got the winds coming up from the sea and from the south of India, and you’ve got this pocket here where the air gets stuck. So pollution that’s produced by industry or farming or whatever can just circulate there for quite a long time before it manages to leave that area. And unfortunately, that also just happens to be the area where there’s maybe the greatest concentration of people in India, more or less. Santosh Harish: That’s right, absolutely. One of the consequences of this idea of an airshed also means that for individual cities and states to be able to make progress, it can actually be quite hard, because they’re substantially affected by the regions around that city in a particular state. They might actually be dependent on actions in the neighbouring state — in the neighbouring country, even — making this a particularly tough nut to crack. Rob Wiblin: Yeah. So it’s not as bad as climate change, but there’s still a significant public good externality issue here, where one city might pollute, and maybe most of the harm of that pollution is done in other cities — possibly other states, depending on where you are, conceivably even another country. I think one thing that you mentioned in a presentation is that in Delhi, roughly 40% of the pollution that someone might be suffering from outdoors in Delhi comes from Delhi, but 60% might be coming from elsewhere in this airshed. So it has to be this common project, otherwise it’s very hard to make major progress on it. Santosh Harish: Correct. Delhi gets a lot of attention, but there are a bunch of cities that are downwind of Delhi, which in many ways are affected by pollution from Delhi, which as a result end up having even higher levels of air pollution. And sure, they’re not as populated as Delhi is, and they’re not as large a city, but in many ways are substantially dependent on Delhi getting its act together. There are parts of the Indo-Gangetic Plain where, compared to the 40% or 50% that Delhi is responsible for its own pollution, the corresponding numbers might be as low as maybe 30% — and therefore completely dependent on actions outside. Something that the larger policy and regulatory infrastructure, at least so far, is just not well equipped to do. Rob Wiblin: Yeah. So at a high level, where does air pollution in that airshed come from? I’m thinking there’s cars, there’s coal plants, there’s industry, and so on. Is there a way of breaking it down? Santosh Harish: There are multiple sources. One way I like to think about this is that these are 19th, 20th, and 21st century problems coexisting with each other. So from the 19th century, there is the question of poor energy access for cooking and heating and so on, and therefore a very large number — we’re talking about hundreds of millions — of people being dependent on solid fuels for daily uses like cooking and heating. So that continues to be widely prevalent, and that is a significant source of air pollution that is likely to be the largest source. From among the 20th century sources, there are things like waste burning, which we briefly touched upon earlier, a widely prevalent practice. So there’s this waste burning. There is road dust, which is basically roads being poorly maintained, and as vehicles go over it, it breaks down a little bit and it actually adds up to quite a significant fraction. And then there are the more standard sources that folks would be familiar with: things like vehicles, industries, power plants that burn fossil fuels — which results in greenhouse gases, but also leads to a bunch of different pollutants which then react in the atmosphere, eventually forming particulate matter. So particulate matter as a class can get directly emitted from some sources — construction dust; road dust from, to some extent, combustion — but it also is something that gets formed in the atmosphere because other gases that have been emitted react and form particles that are less than 2.5 microns. Rob Wiblin: Is it possible to give kind of rough percentages for what fraction of it is coming from each of these different categories? Santosh Harish: Let me start with a city like Delhi. Delhi is relatively well studied; there are multiple “source apportionment studies” that try and attribute the contributions of different sources. For a city like Delhi: waste burning, vehicular emissions, industrial and power plant emissions, and a variety of residential and commercial sources — mainly in the preparation of food — all of these are, give or take, 20% plus or minus 5%. As classes, they are roughly similar to each other. Now, each of these are actually fairly broad categories. Like among industries, you have power plants and other large formal manufacturing plants and things like that. But then there are also a variety of smaller enterprises, informal industries like the battery recycling type things, or e-waste recycling. Again, these things can add up to quite a bit. So there are multiple individual sources that are part of these larger classes. With vehicles, again, you have some subcategories like trucks: there are a small number of vehicles on the road that can have disproportionately large contributions, while the number of cars and two-wheelers might actually be much smaller relative contributors. So for a city like Delhi, you almost have this equal-ish split between these four or five broad categories of sources. At a national level, there seems to be evidence that suggests that household burning is the single largest source, accounting for something between 25% to 30% in terms of the ambient air quality exposure — this, by the way, is separate from the impacts it has indoors, and we can come back to this point later — about 10% each is from industries and power plants, and then you have smaller and smaller shares. So at a national level, vehicular emissions actually might potentially be in the single digits, and you have industrial and residential actually accounting for the vast majority. Rob Wiblin: I guess an unfortunate bottom line from that is that there’s no one thing that, if we just targeted that, would really solve the issue. Unfortunately, it’s spread somewhat evenly across quite a lot of different sources. So in order to make big progress and reduce air pollution by half or two-thirds or four-fifths, you’re going to have to deal with five different categories in a pretty big way. Santosh Harish: Exactly. That’s one of the things that also makes the air pollution in developing countries hard, and also sort of different from air pollution as was seen in the US or UK in the first half of the 20th century — where it was primarily dominated by industrial pollution, and to some extent the use of coal for heating. There’s just a much wider variety of sources which fall under different government agencies, which operate at different scales, and therefore making progress becomes harder to do. And you would have to make progress on each of these sources, because they all contribute enough. Rob Wiblin: One category you haven’t mentioned yet, which gets a lot of coverage in the media, is stubble burning, which is, I think, the burning of leftover crops as part of the growing cycle. Where does that fit into the picture? Is that a big source or a relatively small source? Santosh Harish: So stubble burning is an important episodic source, in a sense. In the months of October and November, when air pollution levels peak in north India, stubble burning actually is a significant contributor. It’s possibly even the primary contributor in those few weeks. Much of the burning of crop residue happens over the course of three or four weeks. This is basically the residue that remains after paddy has been harvested, and as the farmers get ready for a new cropping cycle of wheat, basically they have a narrow window in which the stuff has to be cleared out and the field is made ready for a new crop. And burning it is the fastest, cheapest way of getting rid of it. And because it coincides with the time when these other changes are happening to the weather, the stuff that gets emitted basically stays in the air for longer durations; it doesn’t get cleared away very quickly, leading to apocalyptic levels of air pollution. So it’s an extremely important source during those few weeks, not just in Punjab and Haryana where this happens, but really the pollutants travel, and it’s a significant contributor across north India. There is some evidence to suggest that it actually may travel hundreds of kilometres, to even southern India and contribute a small share. So it gets a lot of attention because it is the principal source when air pollution makes news. Seen over the course of the year, it’s not an unimportant source, but it’s not nearly as important as the amount of airwaves it receives — whereas stuff like household burning and so on are largely under the radar, despite how important they are. Rob Wiblin: Maybe one reason it gets attention, at least in the foreign media, is that it’s so strange. As someone from here, my question would be: Why do Indians put up with this? Why don’t they ban people from burning crops that’s causing what you describe as “apocalyptic” conditions for them? I guess we don’t want to get too caught up on this, because, as you’re saying, as a percentage over the year of all exposure, it’s not so large. But it is very striking. Santosh Harish: It is striking. And it does lead to significant short-term heightened exposure, and therefore a significantly larger number than usual having to be admitted to the hospital because of breathing difficulties and so on. So it does have significant impacts on health. I guess part of the answer to you is that it’s in fact banned: this is not something that is allowed by law. It just happens at a scale that’s so large and by a group that is sufficiently politically influential that there is very little that folks can do in terms of enforcement. It’s hard for governments to go around fining everybody who indulges in this, but in fact, these are banned in the strict sense of the law. Rob Wiblin: OK, maybe we should come back to enforcement as a key issue later on, or the politics of it all. Is there any really important data that we wish we had about all of this, that we don’t have? Santosh Harish: One of the things that we don’t have is better measurement of air pollution in rural areas. It’s the kind of thing that might also help with the monitoring of stubble burning and so forth. But the complete absence of air pollution monitors in rural India is a significant problem. So there’s a very large part of this country, and the region at large, that we don’t have enough visibility and insight into how pollution changes. Many cities in India are still under-monitored. Delhi actually is a pretty well-monitored city at this point. But this isn’t the case with virtually every other city. Compared to Delhi — which may have something like 40 to 50 monitors operational on any given day — most cities actually have two. That remains a problem. With the source apportionment — like the numbers that I gave you on, for example, the contributions of household versus industries and power plants — most of these are based on the results of models. So these are not measured; these are estimated from a bunch of atmospheric chemistry and dispersion models relying on something called emission inventories, which are basically gridded estimates of emissions from different sources. The emission inventories are likely to be very wrong. So we just don’t have a good measurement of how much is getting emitted, and that becomes the basis for all these subsequent estimations of the relative shares of different sources and so on. We have substantial gaps to fill on the emission inventories. And those are pretty fundamental, because without that, it’s hard to be sure of how much different sources contribute. These estimates of how much of air pollution in Delhi is because of the neighbouring states: all of these are model-generated; all of these are reliant on being reasonably correct about where the emissions come from. That’s a big gap. There are actually a whole bunch of fairly fundamental gaps in our understanding of air pollution. Rob Wiblin: In terms of the broader global picture, I think India has among the worst air pollution in the world. I think air pollution is getting worse in Africa at quite a rapid pace, and there’s some countries in Africa that are now giving India a run for its money. I think DRC might now be a very large country that has among the worst air pollution in the world. On the other hand, China has managed to reduce its air pollution levels by more than half over the last 10 or 15 years. So they managed to make a lot of progress, and it’s a less severe issue there now. But what’s the trend in India over the last couple of decades? Is it just getting worse as the country industrialises? Santosh Harish: At least until a few years back, the trend had been almost a monotonic increase in the levels of air pollution across India. There were parts of the country that had virtually negligible [pollution], or at least that had significantly cleaner air, that had become much worse as industrial and mining activity sped up. So in general, if you looked at a map of air pollution in India, in many ways it has grown a deeper shade of red, and that region has expanded over time with new pockets of high levels of air pollution. In the last few years, there’s been something called a National Clean Air Programme that has been largely restricted to cities. It’s in many ways an imperfect programme: it’s insufficiently funded, for example, and in many ways it’s really a start more than anything else. But I think there are reasons to be cautiously optimistic that levels have at least plateaued a bit, maybe even reduced by 10% or 20% in some cities. Not nearly the kind of progress that cities in China seem to have seen over the last decade. I think the estimates that between, I’d like to say 2010 and 2020, China on average saw like a 30% reduction in PM2.5 levels. That might very well be higher now. So they have in fact managed to make significant progress, whereas I think Indian cities are at an earlier stage of that process. Rob Wiblin: So in terms of explaining why this northern Indian airshed has among the worst pollution in the world, one thing is just that it’s incredibly densely populated. Another is that it’s industrialised enough that there’s quite a lot of pollution, but it’s not maybe rich enough and developed enough yet that people feel wealthy enough to do a whole lot in order to reduce it. And also, it’s just an area where air really gets trapped, so air pollution that gets produced there doesn’t just blow out to sea so it can really build up over time. Are there other kinds of possibly policy mistakes, or other key factors in terms of India in particular, or the politics, or the direction that the country’s gone that has made air pollution worse? Santosh Harish: I think at a high level it’s less about policy mistakes than it has been about policy neglect. So the Air Act in India dates back to 1981, and in many ways the regulatory apparatus that resulted from the Air Act reflects the understanding of that time. There are pollution control boards at the state level and at the central or federal level, and these have been primarily set up to deal with industrial pollution, because at the time that was the general sense, that this has got to be an industrial emissions problem. Over time, in the mid to late 1990s, primarily because of civil society advocacy and the interventions of the judiciary, the Supreme Court of India, it expanded to include vehicular emissions in big cities. But until maybe the middle of the last decade — around 2013, 2014, 2015 — in many ways air pollution once again started becoming more visible in Indian media, and therefore it became more salient, and there were a bunch of these source apportionment studies that got commissioned and got publicised. That was it; that was basically the extent to which the regulatory apparatus was readied. So these other sources, like waste burning or household burning, were completely neglected, right? Household burning, for example, I think we’ve still not recovered from that. It’s assumed that the household burning of solid fuels is something that leads to indoor air pollution, and that it really doesn’t have much of an impact on pollution outdoors. That’s just not true as per the literature. So that’s been completely neglected, and treated as a distinct problem. Because the regulatory apparatus has been set up for industrial pollution, the pollution control boards are not really well equipped to deal with the updated understanding of where air pollution seems to come from, and therefore what you ought to be doing about it. So if you consider, for example, waste burning: this is something that really falls under the jurisdiction of the municipal corporations — the local government. The municipal corporations have never had to think about air pollution ever, right? This simply isn’t something that they think of as something that falls under their mandate, until at least very recently. And therefore the agencies that are supposed to be doing something about air pollution — the pollution control boards — don’t have the jurisdiction on this source; while the agencies that do have the jurisdiction have not had the regulatory experience or the capacity to deal with this. Therefore it has sort of fallen between the cracks a bit. In terms of policy missteps, though, when it comes to industrial pollution, the regulatory regime in many ways has not been designed in a manner that is flexible enough and sufficiently in sync with the challenges of regulation in the field. So with industrial emissions, typically the way regulation functions is that all of these industries that pollute have chimney stacks from where the flue gas and the other pollutants escape. For the longest time, across the world, regulation was basically about the height of the chimney stack: the assumption was that you make that tall enough, and the impacts are not felt in the immediate vicinity. Which is not untrue, except that pollution can travel, and over time all of this adds up. So the next generation of regulation in many ways was to set standards for what the concentrations of pollutants in these chimneys could be. There is a mechanism by which the pollutants are measured in the chimney. You compare it with what the regulatory standard ought to be. In order to comply with these standards, the industries basically install a bunch of pollution control equipment — scrubbers and filters and things called cyclones and so on — which are meant to clean up the air. And regulators basically measure this from time to time. If you’re found to be above the standards that have been given to you, there is some form of punitive action. The way this mechanism has been set up in India has relied on what are called command-and-control instruments. Basically, there is a standard, and you’re meant to be under that standard. Somebody comes and measures this from time to time. If you’re above that standard, it is a criminal offence, and therefore there’s going to be a lawsuit filed against you. You could land up in jail and pay a fine of some kind. In practice, the legal system in India is backed up: most cases take years and years and years. The actual compliance against these standards is quite poor: based on some data from a few years back, something like 50% of industries in the state of Maharashtra, for example, were not in compliance with the particulate matter norms. So there’s widespread noncompliance, the pollution control boards are understaffed, and there is no real mechanism by which they can go after these many industries that are flouting the law. As a result, for the most part, the regulatory regime just sort of fails. So if a particular industry is found to be noncompliant, there is a gentle slap on the wrist, there is some kind of polite correspondence where the regulator writes to the industry and asks them to explain themselves, and it sort of ends with that. There is very little action taken. The policy misstep, I guess, is that the evolution that the regulatory framework should have had over time — from being reliant purely on these extremely rigid, some may say even sort of draconian, command-and-control type regulations towards a wider variety of more flexible tools that allow the regulator to levy fines without having to file a criminal lawsuit and so forth — that evolution just did not happen. And as a result, noncompliance became widespread. The amount of industrial activity in the country increased; the pollution control boards were never really able to keep up with it. And the one source of air pollution that ought to have been regulated well also did not see much progress. Rob Wiblin: I’ll try to sum up. So there’s a whole lot of things that have potentially gone wrong. One is that the legislation in this area only looks at industrial pollution, primarily. Agricultural sources, household burning, waste management, I guess roads falling apart — that stuff is kind of ignored, because that wasn’t really on people’s radars in the ’80s when this was written. So that’s a massive problem. And then even in the area that the legislation is focused on — industrial pollution — the mechanism of enforcement is that occasionally someone will show up and try to measure whether you’re in compliance. But presumably they’re not super well resourced, so the checks are not very frequent. And then even if they do find that you’re noncompliant, the enforcement mechanisms at that stage are so impractical — taking someone to court, which might take years to even get a hearing — that if you’re an industry player and you don’t want to follow the rules, you can just break the rule, hope that you don’t get caught, and then even if they do catch you, then you just keep them stuck in correspondence and ultimately, no one’s ever really going to come and shut down your factory or force you to change your ways. I mean, in that picture, it’s almost surprising that things aren’t worse, because it seems like there’s very little, in a way, being done to prevent air pollution. Santosh Harish: That’s exactly it. And thanks so much for summarising what I assume took several minutes. Yeah, I think what the regulators have therefore done, and I guess minimised the extent to which things can go horribly wrong, is that they’re sort of focused on industries that are known to be egregious polluters — are like known problem cases — and largely restricted their attention to them. And the vast majority of noncompliant industries were just ignored for the most part. I mean, arguably, Rob, air pollution in India is about as bad as it could possibly be. Rob Wiblin: There’s not much worse place to go. So in a sense, this is nice, because it means that there’s a lot of low-hanging fruit here. It seems like it’s obvious what changes you could potentially make if the government were really up to the task. You could create regulations around air pollution that covered a far wider range of sources. You could properly resource the agency that’s meant to enforce it. You could give them many more gradations of possible punishments — like you’re saying you could use fines that they’re just allowed to impose without having to take anything to a court. From what I’ve heard, I think you mentioned in a talk that, oddly enough, in India, when air pollution comes up, city governments can turn to outdoor air purifiers, for example, as a possible solution. Maybe because it’s very visible and it kind of looks cool, but it is incredibly expensive and incredibly ineffective, as you might imagine, sticking an air purifier outside: there’s only so much that is possibly going to do if you haven’t done anything to control the source of the air pollution. What is going wrong there? Why aren’t people able to fix these problems, given that it seems kind of obvious what improvements there might be? Santosh Harish: Right. So starting with the outdoor air purifiers, I guess the charitable way of seeing it… I mean, I’m a fairly optimistic person, so I guess one way you could see this is that this, in some ways, is a manifestation of the public demand for cleaner air going up, and governments at least being forced to do something. And smog towers, as they are called, you could imagine the case for them: that they’re sort of plausibly useful, they’ll do something; they’re physical, visible manifestations of the intent of governments to clean up the air. And perhaps equally importantly, it leaves nobody worse off in the near term. Rob Wiblin: Except taxpayers. Santosh Harish: Yeah, the taxpayers, but it’s not as visible. Rob Wiblin: Spread very widely. Santosh Harish: Exactly. That’s one reason why smog towers are so attractive. Most regulation has winners and losers. And here, except for the taxpayers who might not be noticing it, nobody’s really left worse off, and therefore it is politically very viable. But yeah, let’s be clear: this is an absolute waste of resources. They will do absolutely nothing. I mean, sure, they may clean up the air a couple of metres away from wherever they are stationed, but it’s highly ineffective. Part of the problem here seems to be that the sources are visible, sure. For some of them, there are obviously good longer-term actions that you ought to be taking. So vehicles are a problem: you need to reduce the number of private vehicles on the road; you need to reduce the number of dirty vehicles on the road. So you can have a bunch of policy actions that try to clean up the fleet, that could potentially improve the public transport infrastructure in cities and things like that. But in the near term, which most governments try to optimise for, one of the challenges is that we don’t have a menu of easy-to-implement, scalable solutions. I think that has been one of the challenges. I think there’s a legitimate uncertainty. You know, if you were a municipal commissioner in one of the Indian cities, or you were the secretary in the Department of Environment at a state level, and you had a pot of money to be able to deploy, I do think that there is a certain gap in terms of saying, “OK, here are the top 10 things that you ought to be doing; here are the most cost-effective interventions that you ought to be investing in.” I think that there’s actually a significant gap in the literature. It’s not sufficient to say that you need to have more buses on the roads: that might not be under your mandate, or that might be much more expensive than you can afford in the near term with the constraints you’ve got. I think that’s been one of the challenges with being able to make progress in India. As a result, for the most part, what the cities have been doing is basically dust-control type measures: having these mechanical street sweepers clean up the roads. They’ll do something; we don’t necessarily know how much they actually improve the air pollution even on these roads. That’s, for example, some of the stuff that we’ve just funded, to try and get a handle on how much of an impact this might truly have. It’s not obvious at all that these are cost-effective things to be putting your money behind, but it’s the kind of thing where it’s not expensive enough for the government agencies not to be able to procure them. Again, it’s the kind of capital investments that the corporations can do more easily than some of the harder improvements — in terms of how you operate, how waste management in a particular city functions. That’s a systemic thing, right? It’s much easier to purchase 10 of these street sweepers or something, put them on the road, and hope to God that it makes a difference. And unfortunately, that’s what they’ve been doing. Rob Wiblin: Yeah. Isn’t there also an issue that dust particles tend to be relatively large, which means that they’re less likely to cross the blood barrier into the lungs and into the rest of the body? Like actual soil dust, I think, is rarely under that PM2.5 threshold that makes things particularly dangerous. Santosh Harish: That’s true. I guess, in their defence, there is a fraction of the dust particles from roads, from construction activities, that are actually finer than 2.5. But your point is well taken. I think that’s exactly right, that by focusing on dust control, you’re probably spending your resources on larger particles that are relatively innocuous. Rob Wiblin: Yeah. Something that’s very visible but not quite as dangerous. Santosh Harish: Exactly. Rob Wiblin: OK, so a key issue, or a key barrier, potentially, is that governments don’t have a clear menu, a clear smorgasbord of “Here’s the cheapest way to abate air pollution per dollar,” or “Here’s the easiest way to do it” and then they just start working down. Instead, they’re grabbing things somewhat scattershot that are very visible and don’t create identifiable losers who are going to campaign against it. So maybe let’s dive in and try to think about that menu. What are some of the best object-level interventions or changes that could be made at a technical level — to industry, to transport, and so on — that would do a lot to reduce dangerous air pollution at relatively little human and financial cost? Santosh Harish: I guess part of that preface on not having that list of actions also means that I don’t necessarily have a great [list]. I don’t think anybody does. And that’s just one of the challenges. That said, some of the almost certainly low-hanging fruits or no-regret measures in some ways are basically anything that improves energy efficiency: anything that makes vehicles more efficient; anything that makes industries more efficient. Most energy efficiency type interventions tend to pay for themselves over time, and will almost certainly reduce the amount of fuel that gets burned. The less fuel that’s burned, the less pollution there will be. So as a general class of actions, anything that improves the combustion, the efficiency of the combustion of coal, the use of fuels. The next set of things will likely be on household air pollution. This is a slightly harder source to be sure of. Like I mentioned, there are tens of millions of households that rely on the use of solid fuels to some extent. Some might be exclusively dependent on them; some might be using them for like 50% of their cooking needs. The cleaner fuel to be using here would be liquefied petroleum gas. Perhaps the cleanest one would be induction cookstoves and electric cookstoves. But that depends on you having reliable electricity and so on, which is another problem, which in many ways precludes this as a serious option in rural areas. Historically, liquefied petroleum gas (LPG) used to be subsidised. A couple of years back, the subsidies were removed. There was a large government programme at the federal level to provide something like 80 million households with connections to this cleaner gas. But over time, the subsidies themselves for continued usage of the LPG were removed, discontinued. They have recently been reintroduced, but at modest rates; they’re likely to be insufficient. This is perhaps, at a policy level, the most important intervention that will have large improvements in air quality associated with it and so on. It’s also incredibly expensive. Just from a fiscal standpoint, this will likely cost something like 10 billion a year, perhaps more. Much of the LPG also is based on imports, and therefore there is a larger energy security question associated with that as well. All of this has made the federal government sort of reluctant to reintroduce subsidies at a larger scale. So while it is likely to be pretty cost effective for a government programme — by which I mean a back-of-the-envelope estimate I had was that this may cost something like $800 per DALY averted, which would not be a GiveWell top charity–level thing, but for government policy, that might actually be actually incredibly cost effective — because of the sheer fiscal costs that entails, it’s been hard to make progress on this particular action. Rob Wiblin: It’s something that was tried, but then I guess just the cost was so great that governments didn’t want to continue these subsidies indefinitely. Which I guess makes some sense in a kind of lower-middle-income country like India. OK, the first cluster was energy efficiency. I suppose a general stylised fact that people claim is that, around the world, businesses and individuals and households don’t invest enough in improving energy efficiency even in order to maximise profits. Maybe because it moves all of the costs upfront, and the benefits come over a long period of time, people tend to delay investments in new equipment that would improve energy efficiency. That’s something that could be relatively cheap over the long term, that could reduce emissions a whole lot. And not only would it reduce burning of fuel, but in many cases it might mean getting new equipment, which presumably might be better on emissions as well, simply because it was new and the most updated car or whatever. One thing that you pointed out in one of your talks is that household waste burning is much worse than you might think. Not only is it probably the largest single source of particulate pollution over this airshed, but it kind of double-dips. So we’ve been talking about outdoor air pollution mostly today, but if you talk about air pollution as a whole, I think that the standard measures, from the World Bank and World Health Organization and so on, are that roughly two-thirds of the health damage from air pollution is done outside from outdoor air pollution, and one-third is done by indoor air pollution. So you’ve got roughly 6.6 million people dying from outdoor air pollution, and roughly 3.3 million dying from indoor air pollution. Now, when someone burns wood or something in their house in order to stay warm or to cook food, they firstly suffer the indoor air pollution in the first place, because they’re right there next to this burning thing that’s producing ash. And then it goes outside: almost all of it leaves the house and it goes and then affects everyone else across the entire airshed on an enduring basis. So because it’s occurring in the household, rather than in a factory that people aren’t living in, it really gets this extra penalty. So that’s one reason why indoor air pollution from fuel burning has been an issue that I’ve heard about for 10 or 15 years as an underrated issue across the developing world. Unfortunately, it sounds like there’s not a really cheap solution to this, because of the alternative cleaner fuels, electricity is not that reliable and natural gas is expensive. You’re saying it’s liquid petroleum gas. Is that different than the gas that I’m familiar with? Santosh Harish: It’s the same, yeah. Very similar. Rob Wiblin: OK, same thing. And I imagine that given that gas prices have gone up a lot over the last 18 months, it’s probably even worse, it’s even more expensive now to try to get people to switch. Just to clarify one thing: what are people burning in their houses for fuel? Is it wood or charcoal? Coal? Santosh Harish: It depends on which part of the country, and expanding beyond India, it depends on what’s easily available around you. So if you have access to firewood, that’s what you would use. If you are part of a region where coal mining happens, where coal and charcoal are relatively easily accessible, that’s something folks use. If cow dung, for example, is available in plenty around you, folks pat them down and make these things called dung cakes, which are also used plenty. Folks use kerosene, but that’s once again relatively expensive compared to these. These tend to be the non-clean sources, if you will, of cooking energy. Rob Wiblin: So I guess the argument in favour of focusing on fuel-burning inside houses is that it has this extra penalty in terms of the health effects. I suppose it’s a huge source, and very scalable to try to get people to stop burning the dung cakes and instead get them to switch to some other source of cooking energy. It seems like it’s not super-low-hanging fruit: it’s not going to be trivial to persuade the government to spend lots of money on this. And maybe this is something that would get solved over time as the country got richer and could afford middle-income countries to rely on electricity or other sources more. One thing that occurs to me is maybe the key issue here is the electricity reliability. If that’s why people in rural areas are not willing to just rely on electric cookstoves, which I don’t think are super expensive, then maybe what we desperately need to do is get them a reliable source of electricity. And conceivably, even if that electricity was reduced by burning coal, maybe that would be a significant improvement from an air pollution point of view. Santosh Harish: I think that’s true, and that’s exactly right. A couple of things though. One is that people’s cooking habits are one of those things that is pretty sticky. So folks have been burning firewood and other solid fuels for cooking basically since the beginning of time, right? It’s the kind of thing where you may have specific preferences for food cooked in a particular way — that rotis, bread, taste better when cooked on a solid-fuel-based cookstove. I don’t know if this is true, but it’s possible that there are these very specific preferences that folks may have in terms of how food tastes and how food ought to be cooked. Even among higher-income urban households who actually have reliable electricity, getting them to move away from gas and use electric stoves instead is not a trivial transition to happen. Rob Wiblin: Well, it’s not trivial in the United States or the UK either. People like their gas stoves here as well. So that’s something very relatable. Santosh Harish: Exactly. So that’s one of those challenges. Now, if the cost of gas increases and it becomes less and less viable, that’s one way in which you can persuade more folks to make the transition, but it’s not an easy one. Solving the reliability thing has its own set of complexities involved. Providing reliable, affordable electricity across rural India likely means you’re probably doubling down on coal in the near term when it comes to electricity generation. So there is a bit of a tradeoff that probably exists out there. But even if you solve this reliability question, there is the matter of getting people to get the right kinds of stoves, the right kinds of appliances that can run on it, like the induction stoves, be willing to make those changes in their cooking habits and so forth. That might potentially be hard. Gas in some sense is easier, therefore. I mean, it’s a more aspirational fuel. It is obviously cleaner. At least the women who end up doing the vast majority of cooking in most of these homes would vastly prefer the gas over solid fuels. They have to bear the brunt: it’s highly inequitable how the burdens of indoor air pollution in particular are shared across genders. In some sense, gas is the easier lift. If the government was persuaded that 10 billion a year or whatever it is was worth it, it’s more amenable to a top-down, sort of one-time thing. And those types of policies are easier to implement than stuff that is more nuanced, and requires millions of people. Rob Wiblin: OK, I think we’re getting a bit of a picture of some of the challenges here. The picture I’m getting of governance around this issue in India is that people aren’t paying a super high amount of attention to it, necessarily. And when people do pay attention to it, they don’t necessarily have a great idea of what the cheap options are for addressing it. And also, this is an area where there’s a lot of people who are still really quite poor. They’re living in poverty, and they can’t afford to use gas, necessarily. They can’t afford to switch the equipment that they’re using, because these rural areas are really quite poor, and the government doesn’t necessarily have the tax revenue to make changes that might seem sensible in a somewhat richer country. I feel like we have a decent grasp now of the household fuel issue. Is there another category of change that looks promising, that you’d like to learn about more, or maybe the government should be considering making changes in? Santosh Harish: One general class of pollution sources, where in some ways there has been more movement than others, is on the vehicle emissions side. If you have to reduce air pollution from vehicles, first you need to get the dirtiest vehicles off the road: move the fleet at large towards cleaner and cleaner vehicles. As a general urban transportation policy, try and incentivise or otherwise move folks from relying on private vehicles to public transport, and in general from motorised transport to non-motorised transport — although I think that’s a heavier lift: India is a pretty hot country; it’s in many ways hard for folks. Rob Wiblin: What’s non-motorised transport? Like bikes or pushbikes? Santosh Harish: People walking and biking for their commute. So getting folks who’ve gotten used to driving motorbikes to start cycling in a hot city is probably harder, and also requires significant infrastructural investments and changes in just the way the cities are designed. It’s a nontrivial thing. Now, one of the areas where we’ve actually managed to make a lot of progress is in moving the fleet at large towards significantly cleaner vehicles. For example, a few years back, primarily because of judicial intervention, India moved from the Euro 4 equivalent vehicle standards to Euro 6. So these are cleaner vehicles running on cleaner fuel with better pollution control equipment installed inside — so the vehicle exhaust is much, much cleaner, by a couple of orders of magnitude, in terms of some of the pollutants that get emitted. So that was a big win. For that to lead to tangible improvements in air quality, basically you need to wait it out. Most vehicles, depending on the type of vehicle, whether it’s a two-wheeler or a car, get replaced every five to 10 years. So that particular policy action will bear fruit, really, over the course of a decade. Another way in which you can reduce vehicular emissions is to have a robust vehicular inspection programme. Basically all the vehicles get themselves tested from time to time, and there’s a certificate that these meet the standards. This is the kind of thing where the Indian state is unlikely to be particularly effective. This is an easily gameable inspection: you can pay a very modest bribe — less than a dollar — and get one of these certificates that say that you are clean. It’s very hard to improve that infrastructure then, because there are a very large number of operators who are certified to give these clean certificates out, and the vast majority of them are susceptible to very small bribes. And that’s the kind of problem where it’s hard to tighten the vehicle inspection thing. So the best course of action is to internalise that reality, in terms of what state capacity is and what can and cannot be implemented. And yeah, the moving vehicles to a cleaner fleet strategy is something that, not just now, even in the past, has been an effective way for advocacy to function and ask for. So that’s likely to be an area where we’ve already made a significant improvement. With the introduction of electric vehicles now, there’s a strong case for more and more people to shift. It’s very likely that my next car is an EV, right? And that’ll be true for most people out there. It’ll just make more and more sense. The infrastructure to charge these vehicles will only get better over time. So I think in general, with vehicle emissions, there is a lot of promise for things to organically get better, and for incentives to be sufficiently well aligned for that to happen. The challenge in the near term is what you can do about identifying the most harmful vehicles on the road. Rob Wiblin: OK, so transport is this other big category. There’s quite a lot of things that could potentially be tried. I suppose one thing would be, ideally, I suppose Delhi and other cities in India might have wonderful underground metro systems, but that’s hard to build anywhere, and potentially quite a challenge to scale up. And it’s difficult to get people to go on their bikes or to go walking places because it’s incredibly hot. That’s borderline sadistic. So what on Earth are we going to do? One option would be to try to find all of the old gas-guzzling cars that are producing tonnes of smog — the kind of thing that drives me insane on the streets of London — and pull them off of the street. But unfortunately, that is very challenging, because there’s just too high a level of corruption, and people who want to keep driving a car like that will almost certainly be able to work around these regulations and just keep going until the car breaks down. So the more practical idea in that vein is instead to have higher standards for what cars can even be sold in the country, and say, look, we’re never going to get these cars off the road, so anything that we sell is just going to keep driving for 20 years, potentially. So we need to get ahead and change the rules so that the cars that we’re selling now are cars that we’re happy to have on the road in 10, 15, 20 years’ time. Presumably, these cars that are better in terms of air pollution, better in terms of energy efficiency, are somewhat more expensive. Do you have any sense of how expensive that is relative to the improvement in air pollution? Santosh Harish: I think this is likely a very cost-effective policy action. In a sense, sure, the upfront costs for most of these will be higher — and I guess it depends on the specifics of the type of vehicle on how much more expensive it may be. But these are costs that are borne privately in most instances, and are also likely costs that will pay for themselves in terms of the reduced fuel expenditure and so on over time. So these are actually likely to be pretty cost effective. Identifying the most gas-guzzling vehicles on the road, which are visible polluters, that’s a harder one, right? Identifying them if it’s not visibly the case is hard, and I guess there are different shades of smoke and so on that one can’t possibly identify. In fact, this may also look very different in different cities. So Delhi, because of court interventions and also general efforts by the state government and the federal government, compared to most other cities in India, has significantly more stringent rules on which vehicles can ply on the roads. As a result, you’d actually have a hard time finding very obviously, polluting vehicles on Delhi’s roads. Diesel vehicles, diesel cars that are more than 10 years old are just not allowed to ply there anymore. So this is hard to do at a national scale. And unfortunately, what often happens is that the vehicles that get off Delhi streets don’t necessarily get scrapped, but really show up in other cities. Rob Wiblin: They just go somewhere else. Santosh Harish: Yeah. And those cities then really get messed up, because they may not get one of these restrictions possibly ever, or at least in the foreseeable future. So that’s been a bit of a challenge. Rob Wiblin: Let’s just talk about this court intervention for a second, because something that was quite striking when I was doing background research here is the role that courts have played in air pollution in India over the years. As someone who’s not from India, it struck me that courts were playing a borderline legislative role. Maybe in the absence of actual politicians doing anything, courts felt that they had to step in and kind of legislate from the bench in order to try to make some of these changes. So there’s lots of cities across this region, but in Delhi specifically, I can’t remember what court, but some courts said, “These private buses and public buses in India are too polluting; we have to get new buses, and they have to use natural gas. That’s how we are going to interpret the law.” And then this didn’t happen for a while, maybe because it was too expensive or people weren’t willing to spend their money on it. And then ultimately, the court just insisted that they come off the road, even though new buses hadn’t been purchased — so this whole bus system, this whole private bus system kind of disappeared. And unfortunately, it wasn’t replaced by cleaner buses; it was replaced with people driving cars or using other ways to get around. So it really backfired, and probably made the problem worse and possibly made things worse. It’s kind of an interesting example of maybe courts don’t necessarily have the flexibility or the understanding to be doing what you might expect a government agency to be doing — like setting different priorities, and figuring out where they want to spend money, and how to most cost effectively reduce air pollution. Is there anything else you want to say about the role of courts in this overall picture? Santosh Harish: Yeah. Like you said, the judiciary has played an outsized — in many ways, a completely unintuitive — role in air pollution regulation in India. Basically, there’s this instrument of public interest litigations (PILs) that have been instrumental in leading to these court judgments. Some of these PILs, by the way, are still active. So they’re cases that were initiated in the mid to late ’90s that are still running today. Rob Wiblin: You’re saying that the court cases are still ongoing? Santosh Harish: The court cases are still ongoing. In a sense there have been multiple orders and judgments, but it hasn’t concluded. And the court continues to play this quasi-executive, sometimes quasi-legislative, role in designing policy. Some of those interventions, I think, were ultimately good and resulted in improvements. Some of those judgments have actually been pretty poor. The courts are simply not the places for some of these decisions to be made. Rob Wiblin: Yeah. In general, we don’t expect courts to be a good place to be doing cost-benefit analysis and setting budgets and so on. It’s not the strength of lawyers. Santosh Harish: That’s exactly right. And it’s also not the place where you necessarily have a democratic reconciliation of the various people who are affected by the judgments. I mean, the case that you mentioned of getting the buses off the road and this mandate that buses can only run on compressed natural gas. It’s not obvious if the state government or the federal government would have made that call, because it’s not obvious that it passes muster in terms of who’s left worse off. It did paralyse public transport in Delhi. It’s also true that we can be reasonably confident that Delhi’s air quality actually improved for almost a period of five to eight years as a result of that ban. Like you said, the long-term impact was that more people then relied on private vehicles, because the number of vehicles in Delhi boomed. And that was the reason there was an uptick again in pollution levels, which eventually resulted in increased attention and increased acknowledgment of the problem in like 2014, 2015. The catalyst in 2014, 2015 was, unfortunately, again the courts. So that problem of policy neglect has manifested in the executive — and to an even larger extent, the legislature — completely ignoring the problem and the judiciary having to step in. The judiciary is sort of limited to fairly blunt instruments. The courts are not the places for cost-benefit analysis, as you put it, and that has unfortunate consequences. That’s increasingly less and less the case, though. The courts have become less activist-y over time. The smog towers in Delhi, by the way, was a direct result of a court judgment. The outdoor air purifiers came as a direct result of the courts basically demanded that Supreme Court demanded that these be set up, because something has to be done for Delhi’s air quality. Rob Wiblin: Yeah, right. This is a little bit embarrassing, but I don’t know: Does Delhi have a metro system or an underground, or is that common in cities in India? And is it possible to expand them? Is that an approach that we should be considering? Santosh Harish: It is possible. From a cost-effectiveness standpoint, urban transport folks tend to stress that buses and a good bus infrastructure moves more people at significantly lower costs — and also sooner, in the sense that metro systems take a while to build. It takes years and years. Rob Wiblin: Yeah, I think in the UK, it takes five, 10, 15 years to build a station. Santosh Harish: Exactly. Buses, in that sense, it’s much more modular. You can increase your bus fleet. It can reach the last mile in ways the metros can never really do. But yeah, the Delhi metro is the largest metro system in India, and virtually every other city is getting a metro of some kind. It’s sort of become this aspirational piece of infrastructure for a city to boast of; like a city is better developed if it has a metro system. But it’s not obvious that was the right step from a cost-effectiveness standpoint. Rob Wiblin: Have you looked into buses, and maybe advocacy around improving buses as a method of public transport as a possible way to reduce air pollution? Santosh Harish: From a grantmaking standpoint, we’ve actually done very little on transport emissions, partly because it’s a relatively mature field: there are a larger number of organisations; there is funding from among the climate foundations, for example, to do some of this work. So it’s not as neglected as some of the other sources of air pollution. The other reason is that if you had to justify these grants from a purely air pollution standpoint, and not think about what it means in terms of increased incomes because of greater mobility and improved quality of [life], things like that — which would be the case for me — it becomes harder to justify the public transport stuff somewhat. Rob Wiblin: Well, let’s set that aside, because I’ve got lots of other questions on other topics. One thing that jumped out earlier is this issue that there’s no effective regulation of industrial pollution, from both stationary power plants and from actual non-electricity-generation industry — that the regulations aren’t really fit for purpose, and then people can potentially flout them. Now, one thing would be to say, you know, Indian state capacity here isn’t really up to the task of regulating these industries very well. Evidently the laws on the books are not terribly good, the agencies aren’t very well resourced, and maybe there’s some degree of corruption involved as well. And so maybe we just have to live in that reality. But an alternative approach would be to say, no, this is actually key, and this is the thing that we need to fix: we need to get this agency properly resourced, we need to change the legislation so that they can impose fines, we need to make sure that they can’t wheedle out of it and just buy time. Have you looked into that as a general approach to cost effectively reducing air pollution? Santosh Harish: Yes. Making progress on industrial emissions is likely one of the more tractable things to do here. And you have a regulatory apparatus; it should be possible that from a state capacity standpoint, these agencies are able to plug some of those gaps, hire additional staff, and be able to regulate industrial emissions better. We have a couple of grants which are governance support in nature, providing fairly broad-based assistance to state governments. And in some instances they’ve been working on industrial emissions. Unfortunately, it’s an area where we should be able to do more grantmaking and just haven’t done enough so far. Rob Wiblin: Well, you’re being harsh on yourself. It’s early days, you’re only in year two. Santosh Harish: Sure, that is true. But at the same time — and we can talk a little bit more about what the portfolio looks like at the moment; the grants that we’ve made, and why we’ve made those as opposed to others — but I think it is the case that as a researcher I used to stress the need to prioritise industrial emissions better for all of these reasons, which as a grantmaker I don’t think I have followed up on nearly as much as I should have. Rob Wiblin: So what’s appealing about focusing on industrial emissions? I suppose one thing might be that you could reduce emissions a lot at relatively low costs to the industries, inasmuch as you’re asking them to buy new equipment or properly use the stuff that scrubs the exhaust from the factories and so on. It’s also quite potentially concentrated, so presumably, unlike household wood burning or household charcoal burning, we’re talking about a much smaller number of actors here. And there’s also maybe an agency that is in principle kind of on your side, where you could go and speak to these bureaucrats and say, “We want to get you the legislation that you need and the support that you need.” There’s a clear actor who you could ally with. Are there any other kind of benefits or challenges with this approach in particular? Santosh Harish: I think those are all exactly right. Industrial emissions is one of those sources that are pretty well understood. You have clear, good global best practices: the same scrubbers and cyclones and things like that that folks would use in any part of the world would be applicable here. The technology is well understood; you know what you’re going to get at the end of that process. There are emissions standards already. The pollution control boards already have the legislative mandate to be able to tighten it in regions which might be more polluted. So all of that exists, all of that foundation exists. I guess what was missing was that flexibility. Noncompliance to industrial emission standards was a criminal offence and so on, as I talked about earlier. Recently there was a legislative amendment that has sort of decriminalised noncompliance with these standards. It’s very early days, but yeah, it’s one of those windows of opportunity of trying to figure out what does it then mean for how a regulation could work differently than it has done in the last several decades? There is also better technology for monitoring. So one way of measuring the emissions from the chimney is that somebody actually physically climbs up a ladder, there is something called a porthole through which you collect a sample of the flue gas, take that to a laboratory, and measure what the concentration of pollutants in there is and so forth. But you also now have continuous emissions monitoring systems that are basically able to measure these in real time and relay the data. So there have been those sorts of technology improvements. Sensor devices are installed in most highly polluting industries already; the data coming out of several of them is kind of garbage, but it’s something that you could sort of fix. Basically, these devices have to be calibrated better, have to be maintained well, but that’s a solvable problem — and solvable at fairly low costs, so it’s likely to be a fairly cost-effective policy. And providing support to organisations that can help pollution control boards to do this are likely to be fairly cost-effective grants as well. Rob Wiblin: Can we talk for a minute about the political economy of air pollution in northern India? You can imagine a typical person perhaps might be really inclined to vote for a politician who says that they can reduce air pollution one way or another. But I guess a typical person doesn’t know a tonne about the issue, or doesn’t necessarily have a great sense of what policies would be most effective. I imagine you have some industrial actors who are not really keen to have the government increasing their costs or having continuous oversight of their pollutions. Maybe they would lobby against that. It sounded like there’s subsistence farmers or small-scale agriculture where maybe they don’t particularly want outsiders coming in and telling them how they can engage in their farming, especially as they’re on a very low income, so they can’t necessarily afford the fancier equipment that would mean that they didn’t need to do stubble burning anymore. Who are the key actors and the key interest groups here? Santosh Harish: Right. So as you can imagine, just from that list there are multiple sources, which means there are multiple sets of actors. At a high level, air pollution is still, unfortunately, a fairly niche, elite kind of a concern, right? So that’s still sufficient in many ways, for example, for the judiciary to take notice, or for air pollution to even make it to the election manifestos in Delhi and elsewhere — because the elite advocacy groups are able to at least have that level of heft to be able to push things along. I guess the challenge is that it’s not more broad-based than that, and that’s something that politicians are well aware of. When it comes to, for example, the power plants, there is a very, very influential industrial lobby. And the Ministry of Power at the federal level, for example, just has much more weight than the Ministry of Environment, Forests and Climate Change. So the pollution control board and the Environment Ministry may set whatever standards they want to, but if the coal power plants are not keen to actually comply with that stuff, they can and they have persuaded the Power Ministry to help them push back. So some of these issues, for example, with the coal, it is a hard problem. India is a developing country. Electricity demand will only grow. The governments do have both an incentive and possibly even an obligation to provide reliable, affordable electricity. So that’s a hard one. With stubble burning, the farmers — not all of whom, by the way, are subsistence levels; some of them are reasonably well off and should be able to afford it — it’s just that nobody wants to pay. Nobody wants to pay costs that they don’t have to. They know that they could potentially get away with it. So for the most part, policy has actually been trying to subsidise this equipment, and provide carrots as opposed to taking more punitive action, because they just are an influential group more generally. An alternative way in which the political economy concern sort of manifests itself is that even if the political parties want to be seen as progressive on air pollution, and be seen as folks who are taking this seriously, you have an incentive to try and do the kinds of stuff that are more visible in the near term than stuff that are necessarily useful. So that, in many ways, creates the incentive for at least the mechanical street sweepers — which may have plausible impact — and at its extreme, invest in the outdoor air purifiers that have no impact but are visible manifestations of intent. There’s a bunch of political economy literature that looks at the Indian state and what it can and cannot deliver well. One of the conclusions there has been that when it comes to routine delivery of public services — which in many ways are less legible, less visible — there isn’t a strong enough incentive to make some of the larger systemic changes and reforms there. The Indian state is much better at one-time, potentially even complex things. For example, getting the entire country, or a large fraction of the country, vaccinated for COVID might be the kind of thing that the Indian state can actually pull off reasonably well. Or run elections for a country this large in a smooth, sort of streamlined way. That’s the kind of stuff that the Indian state can pull off quite well. But you have your best officers taking charge of that. So there are ways in which one-time things with an exit, in a sense, are easier to do. Improving waste management, making the roads better, not having potholes and things like that, better designed and better maintained — that’s the kind of stuff that’s harder to do, but that may also provide an entry into solving air pollution. Rob Wiblin: I would have thought that it was a big quality of life issue in Delhi and other places, that this would be a higher-level issue for voters. It sounds like it’s not in the top five issues. Is there an explanation for that? Santosh Harish: Partly it is that air pollution — if it’s not at those apocalyptic levels like we have in November — is largely an invisible thing. Rob Wiblin: You just get used to it, this is just the way things always are? Or you’re saying it doesn’t even smell that bad, necessarily? Santosh Harish: Exactly. It doesn’t smell that bad necessarily. Unless you buy into this literature, and you’re poring over the Global Burden of Disease or something like that, the impacts are really over a large period of time. So folks might be dying of lung cancer or might be getting diabetes more, or getting these chronic lung diseases and so forth, but it’s not easy to attribute to air pollution. That’s the challenge. It’s hard to make it more visible and salient to people in their daily lives. Rob Wiblin: In prepping for this, I made a list of possible barriers to reducing air pollution in India, and I broke it down into: people not paying attention or appreciating that the issue matters much; incompetence and opting for solutions that are daft and don’t really make much difference; political power of polluters meaning that it’s hard to get up sensible solutions that would interfere with their interests; it just being too expensive for a country at India’s level of wealth and development to pay for the solutions; and then there was a fifth one, which is that we’re mistaken, and actually some of these things shouldn’t be done because the costs would exceed the benefits. It sounds like at least the first four are quite significant factors: that there’s not that much attention; when attention is turned to it, dumb stuff gets done reasonably often; polluters have a lot of potential to block changes; and also some of the things that we might like to do are just too expensive right now, like getting lots of people to switch to LNG for house fuel. Santosh Harish: I agree. All those four are true. I’m somewhat more charitable in my read on the incompetence angle. Sure, that’s probably there. But in many ways, I think one of the limitations from the side of civil society and think tanks and so on has been not providing those solutions in a manner that’s sufficiently sympathetic of the incentives and constraints that these decision-makers face. And therefore, the slightly more nuanced lobbying or advocacy or whatever you’d like to call it, I think that’s something that’s been missing. Part of it is that there’s some legitimate knowledge gaps. So I don’t think the incompetence is necessarily solely on the side of the government here. Rob Wiblin: On that last one — that maybe something shouldn’t get done — of course, as a country gets richer, often air pollution goes up. But also people get more reliable access to electricity, they have better access to transport, GDP goes up and health improves in other ways. For all these reasons, is it possible that some things that I might be inclined to push for would in fact on net be bad, because they would have too damaging an economic impact? There’s lots of pollution that presumably we think should happen. If we tried to reduce air pollution 80% overnight, it would actually be a catastrophe, because all sorts of really important stuff would have to stop. What’s the risk of things backfiring? Santosh Harish: I think the key is the speed of this. So if you actually wanted most of these changes to happen very soon, you are then necessarily restricting yourself to a set of fairly coercive, borderline draconian, if not draconian, actions — which might end up being fairly harmful. So a case in point is that because construction dust is a significant share of Delhi’s air pollution, one of the things that happens every winter is there’s something called a Graded Response Action Plan — basically a set of escalating actions in response to escalating levels of air pollution, some of which are preemptive in nature. So construction activities are basically banned in Delhi for a month, maybe more. And some of it is unpredictable, so it’s not like folks can plan actively around it, and, say, from the first of December to the fifteenth of December there won’t be construction. These are usually bans that are announced a few days in advance or something, which can therefore be extremely expensive for the folks developing these projects, has significant income loss for the labourers that work on these sites. So yeah, I think the key is the speed at which you want [changes]. So if you’re looking for abrupt, drastic changes in air quality, that will likely end up being expensive. However, I’m reasonably persuaded that for the most part, and certainly at an average net level, the benefits are going to far exceed the costs. And if this emergent literature on productivity and so on is true, and the benefits in terms of improved productivity and improved cognition might be on par if not even maybe larger than the health impacts, then the case is pretty solid that cleaning your air will also boost economic growth. Rob Wiblin: Yeah, that’s one reason I think economists have been getting more and more incensed by this issue, is that they think that so many of the ways that we might reduce air pollution would, from a country level or an economic point of view, pay for themselves many, many fold over. Because having us all be sick is very bad for wellbeing and extremely expensive to deal with. And then it means people can’t go to work. Then if this other stuff — that having air pollution reduces worker productivity, makes people unable to work intelligently — is true, then these effects could be really enormous. So we might be willing to pay really quite surprising amounts, much more than we currently feel like we should be willing to pay, in order to reduce air pollution. And what’s going on is simply that the benefits are not so visible: we can’t connect all of the benefits that we’re getting from reducing air pollution, while the costs are extremely obvious and immediately visible to governments and people having to buy new equipment and so on. Santosh Harish: Yes, absolutely. Rob Wiblin: OK, let’s look at this from another angle now. We’ve been talking about what object-level, on-the-ground changes to equipment and methods and transport and so on might be cost effective. But you have to look at things from a somewhat more meta level, where you’re saying that whatever we’re ultimately trying to have happen, you have to think about grants that would cause the government, in almost all cases, to implement these changes or cause other actors to change their behaviour. And I guess the dream is that you could get a lot of leverage there, where relatively small grants might be able to shift policy across an entire state or a city or conceivably even the entire country, if you really changed what people believed and what was politically feasible. Do you have any indications about how practical it is to drive policy change as a philanthropist, basically, or even just as any kind of private actor in this space in India? Santosh Harish: Yeah, I’m definitely fairly bullish about the possibility of progress through the work that individual researchers and think tanks and policy advocacy groups can do. I think there are certain areas where the possibilities of leverage are pretty high and therefore small grants can have outsized impact. One constraint as a grantmaker here is, because I represent an international foundation and the foreign funding rules in India are fairly restrictive — and this is especially the case when it comes to environmental work, which I think has often been interpreted as being adversarial when it comes to industrial and economic development in India — that does restrict the types of grants that one could make. So for example, we talked about how the judiciary has had a huge influence. I am fairly ambivalent about the overall impact that the courts have had. I do think that some good things have happened as a result of it, but I also think that they are not necessarily the places where this decision making should be done. But either way, that seems like a potentially important institution to be able to influence or work with. But as a grantmaker, that’s something that I can’t touch, or at least have chosen not to touch. There is following the letter of the law — there are certain things that you can and cannot do — but there’s also the spirit of the thing. And we have been fairly careful. Media is another important agent of change. Arguably one of the things that ended up leading to this increased pace of activity since maybe 2015, 2016 was that media outlets took notice, and there were just a lot more stories. And over a period of time, these stories became more and more sophisticated in their understanding of the signs of air pollution and what governments ought to be doing and not doing and things like that. So that had an important role. Again, that’s something that, as an international foundation, we can’t make direct grants to. Rob Wiblin: Oh really? It’s just not permitted? Santosh Harish: It’s not permitted. You cannot fund journalists to write stories, because it could be interpreted that this is basically a foreign actor trying to influence the media discourse, and therefore the general narratives within a country. So some of those are the types of things that could plausibly lead to changes and have very high leverage, but stuff that we cannot and should not do. Rob Wiblin: It’s off the table, unfortunately. Santosh Harish: They’re off the table. As an Indian citizen, I understand where this comes from. Some of it seems to me as being potentially defensive, but this really feels above my pay grade. I think it’s entirely reasonable to be compliant with the spirit of some of these restrictions, even if that means that you do have some opportunities off the table. I wish there was more domestic funding in India that was trying to go behind these opportunities and be engaged in air quality. That, unfortunately, has not been enough of the case. It’s a fairly neglected area, even from a domestic funding standpoint. So we’ve had to therefore restrict ourselves to certain types of grantmaking opportunities. But I do think that there is still plenty of highly cost-effective opportunities on the table. Rob Wiblin: What sort of options are you most excited about that you can engage in? Santosh Harish: One way to classify the types of grants that we could make here is that there is a policy framework in place; there is therefore an opportunity to fund organisations that can work with government in providing technical assistance, project management assistance, and governance support of various sorts of hues and shades. So that’s something that has been the largest chunk of our grantmaking so far. And they sort of fall along a continuum. Some of them you can think of as effectively consultants sitting within government, essentially being responsive to priorities that their partners within government express, and just helping them think through these, helping them implement some of these things or monitor some of these things. So effectively, we are therefore paying for consultants who would be sitting within government, helping with whatever the government agencies are interested in. Further along that continuum are organisations that essentially have their own strong core areas of expertise, which might be restricted to one or two sources of air pollution. Folks who specialise in transport or industries or whatever, and are essentially pitching potential intervention ideas to government, or building capacity within these organisations to be able to intervene better. So providing technical support, but from the standpoint of having expertise in these areas. There are a few grants of that kind. We have a grant, for example, to the World Bank, which in some sense is providing technical assistance, but the World Bank also brings the prospect of potential loans to the table. So there’s a higher leverage that one could potentially have here, that not only are these folks trying to recommend certain types of interventions and get budgets allocated within government for that, but also help the government raise resources at the state level, for example, to do this. So that’s one kind of technical support or governance support grant that we’ve also tried. Something to note here is that it’s not obvious to me how many of these are going to work. From a hits-based grant giving standpoint, it’s not obvious what good governance support would look like here, and which organisations are going to be particularly well placed to actually pull off the stuff that they claim they will. So I’m new as a grantmaker; this is my second year of doing this. It’s been a little bit of a learning curve as well to try and figure out. We’ll take our chances with a few promising organisations, but we can’t be sure what success will look like in any of these cases. Rob Wiblin: Yeah. Hits-based giving is this philosophy where you say, “I’m going to accept that maybe nine out of 10 of the grants that I make are not really going to go anywhere, but one in 10 will be a massive success — it’ll be a big hit, and that will pay for the rest.” A little bit like a venture capitalist-style of investment. And whenever you’re trying to influence government policy, I think you basically have to engage in hits-based giving, with that in mind, because there’s no simple, reliable thing that allows you to slightly improve government policy by an equal amount for every $1,000 that you spend. You can’t just crank the handle. Santosh Harish: Low probabilities of success, but then if it comes off, it can potentially have outsized impact. Rob Wiblin: Exactly. Yeah. So it sounds like you’re trying to cultivate experts and consultants and think tanks that can work either almost directly within government bureaucracies, or can be consulted by government bureaucracies when they want and provide expert advice. And have kind of a policy menu that they’ve thought about and developed, and they’ve tried to answer many of the questions that we’ve been talking about earlier — like, What changes would be most cost effective? How would you change the legislation in order to make this work well? We’re at the stage where that groundwork, as you’re saying, hasn’t been done to a surprising extent. There’s just a lot of questions that would be great if academics or think tanks were looking into — that then this would be really useful for people in government or legislators who wanted to take action on this at some point. That’s the basic picture? Santosh Harish: That’s right. I wouldn’t say that this hasn’t been done at all. It has been done in other spheres of governance, not as much on air quality type things. They have been done by other foundations, but perhaps not at the scale and maybe level of risk-taking that, for example, maybe Open Philanthropy is more comfortable with. So I think that’s something that we’ve tried to bring to the table here. Rob Wiblin: I know one reason that Open Philanthropy decided to go into this area is that they thought it was surprisingly neglected, relative to the very large number of deaths that’s being incurred, like 1 or 2 million people dying in India every year due to air pollution. There were surprisingly few people focused on it. Is there a way of kind of quantifying or explaining how neglected the topic of air pollution in South Asia is? Santosh Harish: I guess one way to think about it is the amount of funding that has gone into it. So one of the more careful analyses of the funding landscape basically came to the conclusion that in 2019 there was something like $7 million a year going into all things air quality in India. Rob Wiblin: Is that from outside the country, or including philanthropy inside the country, or advocacy inside the country? Santosh Harish: I think primarily from outside the country. But then from within the country there was very little. And that’s still the case, so I don’t think that would really change the estimate very much. Rob Wiblin: It’s extraordinarily low. It’s almost hard to believe that it can be that low. One feels that something has to be getting missed here. Santosh Harish: Yeah. I completely agree. And even that, I think maybe the last couple of years that might potentially be shrinking a little bit. Because a significant chunk outside of Open Phil, several of the foundations that were funding air quality work in India were primarily coming at it from a climate standpoint. I mean, since Paris, since the COP — actually not just Paris, maybe even just a couple of years back — folks have now pivoted back towards things like electric vehicles and just transitions and so on. And in some ways the funding for air quality has shrunk a little bit compared to those numbers. So I think that gives you a sense of how neglected it is compared to the scale of the problem. One reason could be that I think there’s a legitimate question of tractability here, which is: Are there enough organisations that would be able to absorb the funding? Actually, even as Open Phil entered this area, that was an uncertainty that was still there. In many ways, I think that uncertainty exists even now. We’re not sure how much folks would be able to absorb funding to utilise in a cost-effective manner. Rob Wiblin: What’s the impediment to absorbing funding? Is it just that there are not really obvious places that you could send the money to that can scale and do much with it? There aren’t people who want to enter this area? Santosh Harish: Yeah, there aren’t enough organisations with a scale — in terms of their own staffing or the experience or expertise they have in-house — to be able to absorb grants of even a few hundred thousand dollars. So I think that’s probably still a bit of a challenge. And not just absorb, but also do useful things with it. I guess compared to the neighbourhood — compared to Pakistan, Bangladesh, and Nepal — there are more organisations. The field is, to my estimation, maybe five to 10 years ahead of where the other countries are. But still, there just aren’t enough organisations covering enough cities, enough states in India with a reasonable depth or breadth of experience in different sectors. Rob Wiblin: So that would leave you in a kind of field-building stage, potentially, where you’d say, we can’t just fund an existing infrastructure here; we can’t find all of these existing experts who could make use of millions of dollars. Instead, we have to gradually train these people and create programmes at universities to create new experts — who then, in five, 10, 15, 20 years’ time, will actually be able to make use of substantially more funding. So it would kind of be a talent bottleneck situation, where there just aren’t enough people qualified to take senior roles in the area? Santosh Harish: That’s exactly right. This is especially true of policy work and think tanks and advocacy groups. One low-hanging fruit, one set of opportunities that were on the table and were sort of neglected from a funding standpoint, was: if you were to buy into the premise that there are substantial knowledge gaps that have impeded action, and that better and more applied academic research could fill those gaps substantially, you do in fact have many academic groups in the country that can actually absorb funds for their research projects — improving monitoring, improving our understanding of sources, and so forth. That has actually been a second significant sort of substrategy for the programme. Rob Wiblin: I see. So this approach would be trying to set up lots of monitors to notice where air pollution is higher than people think, and I guess, conceivably, could you try to find some way of finding out which factories are producing far more pollution than they ought to be, without necessarily having to visit them? Is that the kind of picture? Santosh Harish: That’s the kind of thing. And are there sources that we have basically completely missed so far, that actually play important roles at a larger regional level or perhaps even in specific cities? Things like that. With monitoring, I guess there are two ways of thinking about it. One is actually directly filling that gap, and providing visibility into, for example, what rural air pollution looks like. But then also potentially setting the stage for government to be deploying low-cost sensors on their own at a larger scale. But because that’s not something that they have done so far, then being able to pilot the use of these monitoring technologies, that are sort of new and offer new advantages and so forth, at a scale that is sufficient to then inform the use of these approaches by government in the future. Rob Wiblin: Can you basically fund the creation of a think tank or advocacy group that would push for the air quality boards to be able to just impose fines on anyone who violates the legislated pollution thresholds? That’s something that jumped out to me as a very natural thing that you might do. Santosh Harish: Some of the governance support grants would effectively be potential opportunities to do things like that. At the moment, most of our grants are somewhat broad and source-agnostic in scope. In a sense, they follow from the priorities that the government partners they’re working with have expressed to them — and therefore, in some instances, they’re in fact working on industries. But this is exactly the kind of opportunity I think we have not explored enough of, where you might end up having a lot of leverage there. Rob Wiblin: Could you describe in a little bit more detail some of the grants that you’ve made since you started, or some of the grants that you’re looking at and considering making now? Santosh Harish: Sure. So the very first grant that I made was in fact to a university professor, an academic group, to basically set up a very large, low-cost sensor network across rural Uttar Pradesh and Bihar, two of the most populous states in India. Uttar Pradesh is the largest state, and like I said, rural air pollution was completely neglected. So what this project entails is basically setting up something like 1,400 monitors across these two states by a fairly prominent academic who’s seen as a credible, trusted expert by these governments, and in general has done excellent work on the use of low-cost sensors and so forth. So yeah, get him to basically spearhead this effort at covering these two states, in some ways helping improve the measurement of rural air pollution, but also almost testing the use of sensors for a use case like this. Because the low-cost sensors have many strengths and have many advantages that they can offer, but if they are not calibrated well and maintained well and so forth, they’ll basically spit out not-very-reliable data. So not just deploying sensors, but setting up the processes where you can be confident about the numbers that these are reporting, involves a bunch of uncertainties at the moment. So that’s what in many ways we have taken a bet on this academic to try and solve. And if he’s able to demonstrate to these state governments that you can actually do this well, and it costs a tiny fraction of what the more traditional monitors do, and therefore there is a way in which you could use your own resources better to have hybrid monitoring networks, I think that could be a great win. If he’s able to persuade them that rural air pollution is a large and real problem, and that you should be making much more progress, that’ll be a huge win. Rob Wiblin: I see. So the situation is that, up until now, governments have mostly relied on quite expensive centralised monitors that are I guess more accurate, but they’re much more expensive. Here you want to trial and demonstrate that something that’s much lower cost, you could scale up to many thousands, tens of thousands even, across the country, you can overcome the accuracy issues. And that could potentially cause people to realise that we can monitor it at acceptable cost at a much larger scale. And then that could have a whole lot of effects, but it could make the whole issue much more salient, and people much more able to complain that air pollution in their area is unacceptable. Santosh Harish: That’s exactly right. Yeah. Rob Wiblin: Are there any other grants you could describe? Santosh Harish: I briefly outlined the various governance support type grants, so I can give a couple of examples to give some concreteness to it. The World Bank grant, for example, what they’ll basically be doing is a bunch of technical assistance. So they work with different state governments to try and persuade them on how they could approach planning and state-level prioritisation of different actions. One of the things that the World Bank, I think, has been doing pretty effectively in India is making the case for moving beyond city-specific or city-centric governance to working at the airshed level. Now, an airshed is an intuitive enough concept to acknowledge, saying air pollution does not really recognise administrative boundaries; that these pollutants can actually travel tens, hundreds, sometimes thousands of kilometres — and therefore you want to be operating at the scale at which these pollutants seem to move and get created and so forth. But in practical terms, what this means is that you would need to be able to work with neighbouring states, have almost a common plan of action, and in many ways hold your neighbouring states to account. If you were to actually truly recognise this as a South Asia-wide thing, it means that there is cooperation between Pakistan and India and Bangladesh and Nepal in trying to solve air pollution. To put it mildly, Pakistan and India don’t really see eye to eye a whole lot. So figuring out how to operationalise this concept, and persuading, starting at the state level, that itself is actually a fairly large transition in terms of how governance is being done for air pollution: being able to think of this as not an urban problem anymore, but as something where you have a common plan for rural sources and industrial sources and urban sources. So that’s something that they’re doing. They’re also — in ways that are perhaps unique to a multilateral agency like the Bank — trying to institutionalise cost-effectiveness analysis for this prioritisation, which I’m pretty excited about. I think there hasn’t been enough systematic conversation, systematic attempts even, at cost-benefit analysis in prioritising different actions. So that’s something that we are taking a bet on the World Bank for. They’ve been doing good work, and we also thought of this as a nice learning opportunity for us, in terms of figuring out ways of intervening and trying to influence decision making, and helping move towards the next generation of air quality governance. Rob Wiblin: Yeah, I should have asked this earlier, but the air control boards, they’re at the city level, primarily, is that right? So this tends to be governed at a fairly local level, rather than at the state or let alone national level? Santosh Harish: So given the variety of sources that we have, there are actually multiple agencies at the federal — or central, as we put it in India — level, the state levels, and at the levels of local government. So it really depends on the source. The state pollution control boards are in many ways the principal environmental regulators in the country. There is a central pollution control board which basically acts as, on the one hand, an expert institution that sets certain standards that the state pollution control boards then follow, but they also have some sort of direct regulatory responsibilities. But for the most part, it’s the states. However, they are primarily focused on industrial sources of pollution. When it comes to other sources, for example, with transport — where you really have a much wider variety of possible actions — there is a state transport department, there is the municipal corporation, there’s a traffic police, which in many ways acts as the principal on-road regulator for visibly polluting vehicles, and so on. So yeah, you basically need coordinated action across all these different agencies to be able to do anything. Anything on waste is with the local government; anything on power plants to some extent is at the state level, but primarily it’s at the federal level that you have it. So depending on the source, the agency and the scale of governance is quite different. Rob Wiblin: Yeah, OK. Let’s maybe do a third grant or a third approach that you’ve been interested to look into. Santosh Harish: All right. So this is one where there has been the most scepticism internally within the organisation, but I am extremely excited about it. It’s super cruxy in that sense. So I’ve made a grant to this consortium of primarily academic partners to develop something called a “reduced complexity model.” Let me try and explain what this is and why I’m excited about it. One of the challenges with policy analysis on figuring out what kinds of interventions will work well and so on is that translating a particular action to the delta in emissions from that particular source is relatively straightforward. But going further, and saying, “This action which will have this delta of emissions from this source will result in this improvement in concentrations” — which is what we care about most — is significantly harder to do. Because depending on the source, the pollutants may disperse to various scales; they may react with various other substances in the atmosphere. So capturing this dynamic requires a very specific form of expertise that allows you to model dispersion on the one hand, but also atmospheric chemistry. So it requires a specific kind of expertise and it requires enormous computing power and time. Consider, for example, coal power plants. So you’re looking at potential alternative policy interventions that you could take. There is a particular kind of emissions control technology that is going to get mandated. Alternatively, you’re saying that you’re going to tighten the standards by X percent, and everybody has to comply with it in whatever way they want. Alternatively, you’re saying these power plants have to be shut down. Now, trying to figure out what impact each of these alternative scenarios would have on air quality levels actually becomes pretty complicated from a scientific standpoint, and that has meant that the folks who do policy analysis are folks who also have the ability to do the scenario modelling and so forth. Now, I’d submit that the vast number of people who have the atmospheric chemistry expertise don’t really have any kind of expertise on understanding policy. And then the folks who can do the policy stuff and understand the nuances of how these things cost have no ability to do the atmospheric chemistry stuff. Which creates a gap in the policy analysis that gets produced, creates this particular gap that we have in not having nearly enough cost-benefit analysis of different policy actions. What the reduced complexity models will do — coming back to the grant — is basically simplifying that second step of the process. So essentially they make some approximations on the atmospheric chemistry side, but basically provide this very nifty computationally light tool that you can learn. So folks who don’t have the atmospheric chemistry training and so on can just use the tool, generate the scenarios they want, be able to figure out what impact it will have at a fraction of the cost and a tiny fraction of the time involved. And that potentially helps create the foundation for new quantitative policy analysis on alternative interventions. Which makes me super excited about the grant, but I can talk about the case against it. Rob Wiblin: Sure. So in a nutshell, this would allow someone who’s, say, an expert on transportation or on house fuel burning, who’s not a climate scientist, to actually, at a reasonably accurate level, a tolerably accurate level, model what implications this would have across the airshed if policy and pollution emissions were changed in this way or that. And this would allow them to say a whole lot more, and potentially to communicate to politicians and the public what benefits they might expect to see if they change policy in some direction. Seems pretty good. What’s the case against it? Santosh Harish: The case against it is that this is field building from an even earlier stage than with some of the other work on the governance support type stuff. In some sense we are creating or trying to create the supply of policy analysis, assuming that there is either a latent demand or that a demand will get created once this stuff is out there and is seen as credible and interesting or insightful — which is a nontrivial assumption. I think you could justifiably say that we are starting at a fairly low base. There are probably a bunch of actions that you could do without any of this very particular kind of policy analysis. And given the way politics works, and given the way policies get decided and implemented and so forth, there is enough messiness in there that this kind of thing might not actually add a tremendous amount of value, and you might end up doing the same stuff. I think that’s a perfectly legitimate point. Rob Wiblin: Yeah. Why do you support it anyway? Santosh Harish: Why I support it is that if you look at air quality governance in other countries — so the way the US EPA works; the way something called the CLRTAP, which is the Convention on Long-Range Transboundary Air Pollution, which was basically this surprisingly successful international convention which involved, when it was signed, the Soviet Union, the US, Canada, and most of the European countries in trying to solve acid rain initially, and then actually go ahead and try and tackle other types of air pollution — the way these things functioned had analysis of this kind almost at its core. So there was diplomatic engagement and compromises and reconciliation and all of that happening with the CLRTAP, but in many ways the policy analysis and scientific assessments provided the facts, if you will, that informed the contours of those negotiations. Likewise with the US EPA: I’m sure there are a tonne of political negotiations, but the cost-benefit analysis is almost by default. When states drop their implementation plans, running some of these models that look at alternative policy scenarios and so on, that happens as process. So there is good reason to believe that this adds value, becomes the foundation of evidence-based governance and policymaking of a particular kind that has basically been standard and has worked in other parts of the world. And therefore there is going to be some point — probably very soon if you have not already reached there — where the common-sense-based prioritisation of different actions and so on will hit its limit, and where you actually ought to be taking the support of more careful, thoughtful analysis. Which would be impossible right now — and which, potentially, the reduced complexity models can help solve. Partly it’s also addressing the constraint in terms of talent in India for running these more complicated atmospheric chemistry models across the country. Rob Wiblin: You need to make it easier. Santosh Harish: Yeah, a few groups can do it, but that’s not enough, given the scale of the problem. Rob Wiblin: Yeah, OK. There’s a historical example where it seems like something like this was extremely instrumental in getting up one of the most successful international environmental treaties ever. So that makes you think that maybe it could be useful in this case as well, which makes a lot of sense. Speaking of historical analogues, are there examples of other cities or countries that have managed to greatly reduce their air pollution and improve health significantly, starting from something like the situation that India is in now? Are there other case studies that you can learn from? Santosh Harish: There are. In fact, one of the reasons that I remain optimistic that this is a solvable problem is simply that it has been solved elsewhere, and in many ways starting from a lower base of knowledge. So starting from the Great Smog in London in 1952 — where over the course of a few days the updated estimate suggests that something like 12,000 people died as a direct consequence of that smog — to legislation a few years later, and then to control measures in terms of cleaner coal for residential use, coal power plants having emission controls or being shut down entirely. So that has meant that London has made tremendous progress. Across the US, even using a more recent timeline, the US EPA claims that from 2000 to 2022 across the United States, there has been something like a 41% reduction in PM2.5 due to the various types of actions that have been taken. In China, which in some ways is most comparable, in terms of just the population and the scale of the pollution — although I think even at its peak, Beijing, for example, was never as polluted as Delhi is; sure, there might be episodic highs, but on average the pollution was always lower than Delhi is — but across China, there seems to have been a 30% reduction from about 2010 to 2019, 2020, primarily because of better industrial emissions control, transitions of the fuel used in many of these industries from coal to gas. Something that, for example, is happening in parts of India right now: basically almost internalising some of the state capacity constraints, and saying there are probably more economically efficient ways of getting these industries to emit less, but it’s unlikely that given the state capacity constraints, that the pollution control boards can actually sort of pull that off. So there could be a more command-and-control type mandate on fuel conversion from solid fuels to a much cleaner gas. So that’s one thing that China did, improving the quality of residential fuels for heating from coal to gas. I think that’s something that really worked well, which I think has a quite direct analogue in India as well. So there are things to learn from other countries. Arguably, the state capacity thing is one big constraint. The variety of sources makes it more challenging in India, but I think there is plenty to learn from elsewhere. Rob Wiblin: I guess the difference with China is maybe in 2010, China was several times richer in terms of GDP per capita than India is now. So maybe there are options that were financially viable for them that India might struggle to pay for now. What does “state capacity” mean exactly? I mean, there’s lots of governance issues in China, but do we just think that the bureaucracy in China is just more able to do things? It’s more able to deliver services; it’s more able to tell businesses what to do. Could we be more concrete about what is the difference between the government in India and the government in China? Santosh Harish: I can’t claim to be an expert on governance in China, so some of these are maybe high-level, superficial takes. Partly, I think, China has been able to resolve some of the political economy-based challenges for industries. So I think there are ways in which, if the pollution control boards were to get the political backing they need to tell industries, “You’ve got to clean up your act and be able to comply with the standards you’ve got” — or potentially even more stringent standards — it need not necessarily be economically prohibitively expensive for the industries to do so. But because they don’t have that political backing, because even from within government the pollution control boards are seen as sort of a hindrance to industrial development, in many ways they have been trying to signal success by how much they can cut red tape and make it easier for industries to function than how much they’re able to clean up the air or water. So I think changing those incentives, changing the success metrics, could play a huge role. I think once there was enough of a public demand that seemed to have built up for cleaner air, China was able to take some of these tougher actions, which has been harder to do in India. On the state capacity thing, in terms of the routine service delivery, my prior would be that some of these issues, the bureaucracy is able to implement stuff better, like implement a better waste management infrastructure better or something like that. I have no basis to believe that though, so I won’t believe it. Rob Wiblin: Yeah, makes sense. OK, we’ve been recording for quite a while and I should let you go, let you get to your kid there, and eventually get to sleep as well, because it’s getting somewhat late over there in India. As a final question: We have a nontrivial number of listeners in India, actually. What sorts of things could someone potentially do if they were a smart undergraduate in India? What kinds of things could they study, or what career might they pursue if they wanted to help with the air pollution issue? Santosh Harish: I think one way of trying to solve for the absence of successful, scalable interventions is there just being a lot more pilots out there, and a lot more organisations that are able to focus on a particular source of pollution and test out new technology, test out new methods of monitoring and enforcement or something. I think there is a significant demand out there, even from within government. From my conversations with bureaucrats, they’re like, “Everybody who wants to work with us, wants to set up monitors, wants to help define the problem better, we think that’s helpful, no doubt. But in many ways we get it. We know that there is a significant air pollution problem. We know that, give or take, this is what the most important sources are.” So if you had to go beyond the problem of defining stuff, and actually figure out what you can do about it and test out new methods, test out new tech. The governance support grants at the moment are only trying to make the existing processes more efficient; they’re not necessarily bringing new ideas to the table. Having more folks who could help in the actual problem solving, as opposed to the problem defining, I think there’s a big gap. I think it’d be great if there were a larger number of young folks trying out new things — not assuming constraints that may not exist, actually, in practice. Just trying out stuff, doing things, I think that’d be great. There’s clearly an appetite for it. Rob Wiblin: And is Open Phil potentially going to be hiring for more people to work on the South Asian air quality programme? It seems like it’s such a huge space that you’d need quite a large team in order to be able to fully understand all of the different options. Santosh Harish: We might in the future. I think everything turns on the experience of the first couple of years of grantmaking and what we learn from it. So there is a real possibility that we actually update negatively in terms of the tractability — that, sure, it’s an important problem; sure, it’s neglected; but actually being able to make progress requires organisations that don’t exist, or ideas that aren’t already out there. So that might be a negative update. I don’t know if that’s the case, but things do turn on that. If our grantmaking were to increase in the other countries — in Pakistan, Bangladesh, and Nepal, there could potentially be, depending on the scale of it, the need for folks who understand those countries better, who understand what works. Because for all intents and purposes, I’m quite a generalist; I have no idea how policymaking in Pakistan functions. I think if that becomes a larger part of the portfolio, and there are more opportunities out there… Again, I think tractability has been the big constraint in the other countries. There just don’t seem to be enough organisations that work on air pollution. Rob Wiblin: Yeah, it’s a bit of a difficult chicken-and-egg problem, where there hasn’t been funding available, so people haven’t gone into it — but that means that there’s currently not really people who you could fund. Maybe it makes it hard to scale things up super quickly in any area. It’s also been a kind of classic issue in AI, in artificial intelligence policy and AI technical stuff as well. Well, I remember I spoke with Alexander Berger three years ago, when they were considering going into South Asian air quality. And everyone was super excited about this possible new programme area, but they weren’t sure whether they would find someone who was able to actually take the bull by the horns and take forward the programme. So I’m super glad that they did manage to find you, and you’ve got so much to say and so much knowledge about this. There was always a risk that this programme wouldn’t even start because the field just wouldn’t have been built yet. So yeah, thanks for all of the work that you’re doing. It sounds super exciting as well as being kind of challenging. Santosh Harish: Thanks so much, Rob. I guess the jury is still out on how I’ll actually perform as a grantmaker. It’s still early days, and it’s been a fantastic learning experience. And Open Phil has been a really good organisation to be a grantmaker in and learn the ropes and stuff. It’s been wonderful, and I enjoyed our chat today and I hope it was coherent enough and focused enough. Rob Wiblin: Definitely more than coherent enough. Well, we’ll all be rooting for you. Maybe we can check in a couple of years’ time and see how some of these grants have gone, and which direction the research has ended up moving. Santosh Harish: Yeah, I’d love that. Rob Wiblin: My guest today has been Santosh Harish. Thanks so much for coming on The 80,000 Hours Podcast, Santosh. Santosh Harish: Thanks, Rob. Rob Wiblin: If you enjoyed that conversation, you’ll be glad to know I have an interview with the founder of the Lead Exposure Elimination Project coming up sometime soon, which covers many more issues related to dangerous pollution in developing and sometimes developed countries. And if you liked that episode, you can show your appreciation by telling a friend about this show, or forwarding it to someone out there who might really enjoy listening to this episode, or another of our episodes that you loved. You really are able to get the show into the hands of people who would value it and might be able to usefully act on what they learn in a way that’s hard for us to do, because you know, we only personally know so many people! All right, The 80,000 Hours Podcast is produced and edited by Keiran Harris. The audio engineering team is led by Ben Cordell, with mastering and technical editing by Milo McGuire and Simon Monsour. Additional content editing by Luisa Rodriguez and Katy Moore, who also puts together full transcripts and an extensive collection of links to learn more — those are available on our site. Thanks for joining, talk to you again soon.