Seeing the Invisible: Responses to Indoor Air Pollution, with Robert D. Metcalfe

Notable Quotes

• Household activities emit pollutants into the home: “These behaviors and technologies give off a pollutant called particulate matter 2.5 (PM2.5). That is particulate matter with a size of 2.5 micrometers, which is about one-twentieth of the width of a human hair. So, these are really, really small particulate matters that, once they get into the human system, can cause some damage to human health.” (4:52)
• Households aware of the levels of air pollution in their home make behavioral changes that significantly reduce indoor air pollution: “When we reveal that information to the households, across the whole of the day for two weeks, we find that the indoor PM2.5 levels fell by just under two micrograms per cubic meter … a big change.” (14:27)
• Most people reduce harmful indoor air by ventilating: “It’s not like people stop frying their steak or stop cooking their eggs. They still want to eat that food, or they still want to have warmth in their home from their fireplace. But they figure out ways to ventilate better … That’s probably the best thing that they could do.” (16:45)

Top of the Stack

• “Making the Invisible Visible: The Impact of Revealing Indoor Air Pollution on Behavior and Welfare” by Robert D. Metcalfe and Sefi Roth

The Full Transcript

Daniel Raimi: Hello, and welcome to Resources Radio, a weekly podcast from Resources for the Future. I'm your host, Daniel Raimi. Today, we talk with Robert Metcalfe, a professor of international and public affairs at Columbia University.

Rob recently published a working paper that looks at a fascinating and understudied issue: indoor air pollution. In the paper, he describes the results of an experiment where people were given information about the levels of air pollution in their homes. The experiment revealed that people often experience unhealthy levels of air pollution in the home, and that sharing this information leads them to open windows and doors, turn on fans, and take other actions to reduce their exposure. We'll talk more about what the study found and what it could mean for public policy in today's episode. Stay with us.

Robert Metcalfe from Columbia University, welcome to Resources Radio.

Robert D. Metcalfe: Thank you so much for having me.

Daniel Raimi: It's our pleasure. I'm really excited to talk about today's topic, which is indoor air pollution and this really interesting paper that you've written on the topic. Before we get into the substance, we always ask our guests how they got interested in working on environmental issues—whether they had some childhood inspiration or came to this topic later in life. What drew you into this field, Rob?

Robert D. Metcalfe: I remember when I was six or seven years old, I was really interested in maps. I was at a local fair with my mum and we went past a bookstore, where there was a book called Geography. It just piqued my interest, and I remember buying the book and taking it home. It had just a whole range of interesting human and physical geography aspects that just blew my mind at six or seven years old. So, since then, I've always been interested in taking geography at various levels in school and learning more about it.

When I went to college, I realized that you could take this subject called economics, and how you could actually incorporate studies of the environment with economics. That was a game changer for me in thinking about my career path. But I think it was that book that I bought when I was six years old that really affected my trajectory.

Daniel Raimi: That's really cool. Where did you grow up?

Robert D. Metcalfe: I grew up in a small city called Swansea in Wales in the United Kingdom. All of my family is still there, and I still go back. I've lost a bit of my Welsh accent, which my family complains about once in a while. But they're all still there, and it's a beautiful place. That's how I grew up.

Daniel Raimi: I can attest to the fact that Swansea and the area around are beautiful, because my mum was actually born in Swansea.

Robert D. Metcalfe: Oh, wow!

Daniel Raimi: Her parents lived in Clydach, which is right outside of Swansea. I used to go there in the summers.

Robert D. Metcalfe: That's a great pronunciation of it. It could be a lot worse.

Daniel Raimi: All right, gosh. Well, we'll have to talk about that more after we're done with the podcast.

Let's get into it now, Rob, and talk about this recent paper that you've coauthored with Sefi Roth. The title of the paper is “Making the Invisible Visible: The Impact of Revealing Indoor Air Pollution on Behavior and Welfare.” We'll have a link to it in the show notes, so people can read along if they'd like.

To get us started … When we think about air pollution, I think most folks primarily think about outdoor air pollution, like emissions from wildfires, coal-fired power plants, or other sources that we might see in the environment. But indoor pollution is also a significant issue. Can you give us a sense of what things in homes or buildings are sources of air pollution? What is the actual air pollution that might be affecting people's health indoors?

Robert D. Metcalfe: It's a great question. One of the main producers of indoor air pollution is cooking. Every time that you grill or fry any food, that emits some pollution. We also find that indoor fireplaces are contributing towards indoor air pollution. Some behavioral factors, like smoking, are also contributors to indoor air pollution. Those are the big factors, but then there is some evidence that there are some other sources, such as candles, cleaning products, and hair and cosmetic products, as well, that seem to increase indoor air pollution.

At the start of this project, I bought my own indoor air monitor. You'd be surprised by all the behaviors that you do in your home that contribute to indoor air pollution. Outdoor air pollution is also a contributor. Some of it gets inside people's homes, so it is a contributor to indoor air pollution. But it seems like, based on the studies that are currently out there and our study, there are indoor sources that are the main reason as to why you might see variation in the levels of indoor air pollution.

These behaviors and technologies give off a pollutant called particulate matter 2.5 (PM2.5). That is particulate matter with a size of 2.5 micrometers, which is about one-twentieth of the width of a human hair. So, these are really, really small particulate matters that, once they get into the human system, can cause some damage to human health.

What we find in some of our work, that we can talk about more, is that there are some levels of this indoor air pollution that are way above outdoor air pollution levels. That was really fascinating to us.

The science out there … Some people think that indoor air pollution, especially PM2.5, might be less dangerous for human health than outdoor air pollution. But there isn’t really a strong scientific claim for that. Some people think it might be the opposite case. And so, the World Health Organization, in formulating its indoor air-quality guidelines, determined there's no particularly convincing evidence that pollution from indoor sources is less dangerous than pollution from outdoor sources.

I think, given that, as humans, we spend most of our time indoors—there are varied estimates of whether 80 to 90 percent of our time is spent indoors—we need to know about this exposure. This is an area of study that Sefi introduced to me, and I find it really fascinating that we just don't know enough about what the levels and sources of PM2.5 are indoors.

Daniel Raimi: That’s really interesting.

Tell us a little bit about this paper; specifically, the experiment that you carried out in London to try to, first of all, measure the extent of pollution, and then how people responded to certain types of interventions.

Robert D. Metcalfe: We went to get really great data on the levels of PM2.5 in people's homes. That data is not publicly available. So, Sefi Roth, a professor at the London School of Economics, and myself partnered with Camden Council. They are one of the local government boroughs in London. Actually, the London School of Economics is within Camden Council. They were also interested in this air-pollution topic, especially indoor air pollution.

We partnered with them and were able to randomly sample residents of Camden to take part in this study, where we would place an indoor air monitor in people's homes. We had great take-up—a response rate of about 20 percent—from sending out cold letters to people, which is really, really great. So, people were very interested in this topic.

Sefi and I bought these indoor air-pollution monitors from a company called Kaiterra. They have this technology that allows real-time measurement of PM2.5. They can measure the air-quality index, as well as indoor humidity and temperature. We had these monitors and were able to convince just over 250 households to place them in their homes for four weeks.

For the first two weeks, no one could actually see the information on their monitor, even though the monitor was working. What we were doing was just gathering the data on what the levels of PM2.5 in people's homes were when they weren't aware of what the levels were. After two weeks, we randomly assigned some of the households to actually receive that information on their monitor. The control group, also assigned randomly, did not receive that information. So, what we had was the gold standard test to see whether revealing information about how the levels of PM2.5 in your home fluctuate every 15 seconds will affect, ultimately, your behavior, and then your PM2.5 levels.

You can think of this as a randomized control trial, like any pharmaceutical or drug company would do to test the efficacy of a drug or a therapy. We did the same thing, but for information. We wanted to test what the causal effect of information on PM2.5 is, ultimately, on PM2.5. So, this trial is really about exposure.

We know when people are predominantly in their home, and we know the air pollution is bad, so we get a good sense of what the exposure is. When we think about outdoor air pollution, we have really great data on outdoor pollution, but we don't know exactly how people are exposed to it. In this study, by placing these monitors in people's homes, we can see exactly when they're sitting in their home, when they're cooking, when they're cleaning—essentially, what the exposure is. The randomization of that information allows us to understand whether it makes the invisible visible, in this case.

Daniel Raimi: That's great. I'm going to ask you about how people responded in just a second, but first, I want to ask you about the levels of air pollution that you observed in the household. What did you find, and how did those levels vary across groups of people, maybe by income, education, or things like that?

Robert D. Metcalfe: Around 60 percent of the observations that we have in people's homes of indoor air pollution are at safe levels. That's below the World Health Organization threshold, which is 5 micrograms per cubic meter of PM2.5. But we do find that when people are home, and when they're engaging in household activities after work, indoor air pollution increases to about 15 to 20 micrograms per cubic meter. That’s 3 to 4 times the World Health Organization’s threshold of what the safe levels are.

What is really interesting is that it doesn't seem that outdoor air pollution is the main predictor of indoor levels. We find that outdoor air pollution can explain about 10 percent of the variance of indoor air pollution. So, 90 percent is coming from the sources that we mentioned at the top of the podcast, like cooking, cleaning, smoking, and heating. These are all the kinds of contributors that we see.

When we look at income levels, we find that levels of PM2.5 are particularly high in low-income households. When we look at individuals who are earning, say, in the bottom twenty-fifth percentile of income in the United Kingdom, we find that the average PM2.5 level in their homes, across the whole of the day, is about 20 micrograms per cubic meter. For those that are in the upper twenty-fifth percentile, their PM2.5 levels are just around 5 micrograms. So, we do see a fourfold difference between the very low- and the very high-income household indoor levels of PM2.5.

This income gradient with indoor levels of pollution seems quite steep. When we look at the ambient pollution levels across Camden Council households, the relationship between ambient pollution and income is not so steep. The gradient is, really, quite shallow. When you look at other studies that have been done on outdoor air pollution and income, we don't find the gradient to be as high as what we observe for indoor air pollution. So, we think this is the first study to show that this indoor pollution could be affecting the low-income populations dramatically more than higher-income ones. This is about exposure; this is when people are home. So, we got a good sense that this is the actual exposure that low-income households are being faced with—these levels of PM2.5—that high-income households are not.

Daniel Raimi: I know you didn't necessarily try to tease this out in the study, but do you have any guesses as to why low-income households might have that higher exposure? Does it have to do with the types of housing, access to mechanical ventilation, or things like that?

Robert D. Metcalfe: We definitely find that lower-income households are a bit smaller, so the indoor air pollution is more contained in the home for a longer period of time. That definitely seems to be correlated with what we find in the experiment. Secondly, we definitely do find that they have fewer ventilation options. Thirdly, there are also lifestyle factors. It might be more likely that people are smoking in these households. They might also have open-face fireplaces and not have the new technologies for indoor heating. So, there are a range of factors that might explain why the levels are much higher for low-income households compared to higher-income households. But, again, we don't really have strong evidence of what the big thing that's driving that gradient is. But we know that gradient exists.

Daniel Raimi: That’s really interesting.

Now, can you tell us a little bit about what happened when you performed your experiment and gave information to these randomly selected groups of households? How did people respond?

Robert D. Metcalfe: When we reveal that information to the households, across the whole of the day for two weeks, we find that the indoor PM2.5 levels fell by just under two micrograms per cubic meter. This is a big change in PM2.5. When people were home—we got time stamps from the surveys that we did on when people are home—that reduction in PM2.5 was about five micrograms per cubic meter.

These are really, really large treatment effects that we haven't seen, particularly in other studies, in terms of both the absolute levels of the PM2.5, but also the percentage change, as a result of the baseline. It really did blow us away, how large those effects were.

We've tried to understand where those largest effects were coming from. It did seem that it was coming predominantly through ventilation. We found three ways to uncover whether ventilation created those really large changes in PM2.5.

First, we just asked people: What did you do when you got this information? Over 70 percent said they started to open windows and doors and ventilate when they got the information that the PM2.5 levels were particularly high. So, we asked them … Some economists don't believe what people say, they trust what they do … So, we looked, then, at behavioral changes.

Two things really stood out. It didn't seem like the number of high levels of PM2.5, what we call the “peak levels,” were different between the control and treatment groups. But, it looked like the gradient of the reduction, or the decay, in the PM2.5, once you hit the peak, was much steeper for the treatment group. So, it seems like those that got the information, versus the control group that didn't get the information, were bringing the peak down quicker than the control group, even though the number of peaks didn't change. This is quite interesting, because it’s not like people stop frying their steak or stop cooking their eggs. They still want to eat that food, or they still want to have warmth in their home from their fireplace. But they figure out ways to ventilate better.

Thirdly, one of the great features of these air monitors is that they measure other things, like temperature and humidity. What we found is, in the treatment group, once these peak events started to occur, people's indoor temperature got a bit closer to the outdoor temperature during those peak events. We didn't find that in the control group. So, we definitely find a correlation between indoor and outdoor temperature, which we think is coming from people opening up their windows and doors to ventilate from the high levels of PM2.5. That, for us, was pretty exciting. We think we have enough evidence to say that once people learn about this information and how bad their PM2.5 is, the first thing they do is ventilate. That's probably the best thing that they could do.

Daniel Raimi: Right. It's easy, it's cheap, and it's right there, right? You just open up your windows.

Robert D. Metcalfe: Exactly.

Daniel Raimi: You and your coauthor talk a little bit about the potential effects of policy intervention here, where maybe the government subsidizes households to adopt these types of air-pollution monitors, or maybe air purifiers, or some other technology that can help them get this information that you've provided them in the experiment, or deal with the levels of pollution that preexist in the home.

Based on the study that you've done and these policy options, what do you see as some of the costs and benefits of government intervention here?

Robert D. Metcalfe: There are various government interventions you could have to try to improve the levels of indoor PM2.5. What we asked ourselves was, What would happen if the government would subsidize the purchase or the adoption of these technologies, whether it be an air purifier or an air-pollution monitor? What would the benefit-cost ratio be? And so, in the paper we take one class of benefit-cost ratio called the marginal value of public funds. That tells you what the social benefit or social impact is of a $1, or £1, government subsidy from adopting this technology.

What we find is that there are multiple channels here. First of all, we show that for every $1 or £1 spent from the government on this subsidy, it leads to about a 75 pence, so 75 cents, increase in health benefits. We find that it also increases people's productivity, because other studies have found that a reduction in PM2.5 is good for people's productivity. We use that study to ask, What would happen in our study if those results were true? We then find, using the other studies, that PM2.5 is correlated with or causes differences in crime rates. We use that in the welfare calculation and find that for every pound spent, there's a 50 pence reduction in crime costs. So, there are potentially large benefits for the individuals for this $1, £1 subsidy, but they are government benefits, as well.

In the United Kingdom, our healthcare is pretty fully subsidized. What we find is that for that £1 subsidy, 18 pence flows back to the government in avoided health costs caused by this PM2.5 in our study, if the subsidy remains, year after year. We also find that, because people are more productive, that would lead to higher wage rates, which would lead to higher tax revenue. This is the really big finding in our paper. If we believe the studies out there about how PM2.5 affects productivity, that £1 subsidy in our experiment actually leads to about 14 pounds back to the government for increasing revenue. So, that's really driving the welfare benefits here. So, what the government is spending is lower than what the government is getting back.

We call that a Pareto improvement, when these things pay for themselves. So, you can imagine, even if some of these assumptions of how PM2.5 affects crime or health or productivity were pretty high and we drew them back a little bit, we still find that this benefit-cost ratio is still infinite. It's still paying for itself. So, there might be good justification for subsidies.

This is one study; we would love to see more studies on this, obviously, to replicate our results in different settings. But this could mean that the government could be in the game of trying to subsidize technologies in people's homes to make their homes cleaner, and that would be beneficial to the government, because they're getting the money back from spending less on healthcare and receiving more revenue from increasing productivity from the workforce.

We even found this for very small changes in PM2.5 in our study. So, we are quite excited about this, as it is potentially a way that we can improve human health and improve productivity for individuals, while the government benefits.

Daniel Raimi: That's so interesting that there’s such a big effect on productivity gains.

I'm curious about that specific question of increased productivity and wages and then tax revenue. Did you assume that people were working from home and were exposed to their at-home levels of PM2.5, or did you assume they were at home for a certain part of the day and then their exposure changed where they went to an office? Can you talk through that a little bit?

Robert D. Metcalfe: We assumed that they're exposed to a 1.9-microgram-per-cubic-meter reduction in PM2.5 over the course of the day. That would lead to them having reductions in productivity either in their workplace or if they're working from home. The interesting thing with this literature is that it is very difficult to know people's exposure. The previous research that we used to look at how PM2.5 affects productivity looked at PM2.5 coming from ambient sources—from outdoors. So, we are using that variation in those studies, in terms of how that relationship works, and bringing that into our studies.

So, we are making the assumption that that exposure throughout the whole day affects their productivity, whether they are at work, outside for some part of that day, or whether they're working from home. That's something that further research should really try to understand—how much of the exposure that you get to PM2.5 at home will actually lead to causal changes in productivity in the workplace, which might not be at home. That's an area we have no idea about, which is fascinating.

Daniel Raimi: That’s really interesting.

Well, Rob, one more question before we go to our Top of the Stack segment. It's about another type of policy intervention that at least some governments have considered.

In the United States, there are some local governments in California, Washington State, and maybe elsewhere, that have either proposed or enacted bans on natural gas appliances that are used for things like cooking and that are sources of indoor air pollution. What would you think about a policy intervention that bans these types of appliances, versus ones that do what you did in your experiment—technologies that give people information and, maybe, reduce peak exposure?

Robert D. Metcalfe: It’s a great question. In our study, in our setting with just under 260 households, when we compare just the differences in the PM2.5 levels between those that have gas hobs versus electric hobs for their cooker, we actually find slightly higher PM2.5 for the electric hob sample, as opposed to the gas hob, which is kind of the opposite of what we might think. This could be due to selection—there could be other reasons as to why there might be differences here. So, we're not making any causal claims, but there is not a clear correlation between using an electric hob and reducing your PM2.5 levels. However, all of those gas hob conversations and debates are more so around nitrogen dioxide, which is not part of PM2.5. Ideally, we would love to have different measures of pollution from indoor sources beyond just PM2.5. So, that would be a nice follow-up study.

I do think that mandating reductions in PM2.5 from ambient sources—for example, provisions in the US Clean Air Act or congestion pricing experiments around the world—only reduce PM2.5 by between 3 and 15 percent. In our study, we observed changes between 17 and 35 percent, just by having information. So, I do feel that there could be benefits of subsidizing some of this technology, which might be better than mandating that people have this technology in their home. Or, there may be other types of policies they might be doing for ambient PM2.5 levels. I think this is an area where it could really be beneficial.

We've done some studies with Octopus Energy Group in the United Kingdom through the Centre for Net Zero. We have some good studies showing that subsidizing heat pumps does lead to a good benefit-cost ratio. These are good things and are good subsidies to have. But we have no idea, really, what the causal impacts of these appliances or these inputs to how to heat your home are on PM2.5 or nitrogen dioxide. I think, again, that's where a lot more research needs to be done. But I think in the United Kingdom example of heat pumps, we find these large benefits come from just a reduction in carbon dioxide, as opposed to looking at indoor air pollution. That's definitely an area for future research.

Daniel Raimi: Right. That’s very interesting, and there are clearly lots of additional threads to pull here to learn more about this topic.

Well, Rob, this has been fascinating. I've learned a lot from this conversation, and I'm sure our listeners have, too. Of course, we'd encourage people to check out the whole paper, which you can find in the link at the show notes.

I'd love to ask you, now, to recommend something that you think is great. It could be a book, magazine, movie, or anything, really—even another podcast. It could be related to the environment or not so closely related to the environment, but anything you think our listeners might enjoy.

Rob, what's at the top of your literal or your metaphorical reading stack?

Robert D. Metcalfe: At Columbia University, we have a center called the Center on Global Energy Policy, and last week they had their annual energy summit at Columbia. The keynote speaker was this guy named Jonathan Ross, who's the CEO and founder of a company called Groq. Groq is an artificial intelligence (AI) company that's producing language processing units to compute and accelerate workloads in AI and machine learning. What this company has done is use what's called “language processing units,” rather than graphic processing units, which is what Nvidia, OpenAI, and so forth, use. These language processing units are super fast, low-latency inferencing tools. When you're using inference, they can be so efficient, even 10 times more energy efficient than the graphic processing units.

So, I'm really quite interested in how this boom or this innovation that's taking place in processing power in AI can really be beneficial for the environment and not a harm, because, obviously, there's a lot of doom and gloom right now about AI and how much energy it uses to train and compute and make inferences. Once those models are trained, there's going to be really, really cheap ways of getting large amounts of human intelligence. I think energy is going to be at the core of where we are going to go for human intelligence, and whether you believe in Moore's Law or Wright's Law of the cost of these things coming down over time, I do think we are starting to see a boom in the productivity and the efficiency of using energy to create human intelligence. And I think that could be beneficial for all individuals in society, and, hopefully, it will get us to a faster and more sustainable future for all countries. I think reducing the cost by using energy more efficiently is something that's good for everyone.

Daniel Raimi: That's fascinating. I hadn't heard about that, and we've only done a couple episodes on this show about energy and AI, but I'm sure we'll be doing more in the future, so I'm glad you mentioned that.

One more time, Robert Metcalfe from Columbia University and the wonderful country of Wales, thanks so much for joining us on the podcast. It's been a pleasure.

Robert D. Metcalfe: Thank you, Daniel, for your time. It’s been great to chat with you.

Daniel Raimi: You've been listening to Resources Radio, a podcast from Resources for the Future, or RFF. If you have a minute, we'd really appreciate you leaving us a rating or a comment on your podcast platform of choice. Also, feel free to send us your suggestions for future episodes.

This podcast is made possible with the generous financial support of our listeners. You can help us continue producing these kinds of discussions on the topics that you care about by making a donation to Resources for the Future online at rff.org/donate.

RFF is an independent, nonprofit research institution in Washington, DC. Our mission is to improve environmental, energy, and natural resource decisions through impartial economic research and policy engagement. The views expressed on this podcast are solely those of the podcast guests and may differ from those of RFF experts, its officers, or its directors. RFF does not take positions on specific legislative proposals.

Resources Radio is produced by Elizabeth Wason with music by me, Daniel Raimi. Join us next week for another episode.