Probable Causation

View Original

Episode 35: Erich Muehlegger

Jennifer Doleac

Erich Muehlegger

Erich Muehlegger is an Associate Professor of Economics at the University of California - Davis.

Date: September 1, 2020

Bonus segment on Professor Muehlegger’s career path and life as a researcher.

A transcript of this episode is available here.

See this content in the original post

Episode Details:

In this episode, we discuss Prof. Muehlegger's work on air pollution and crime:

"Air Pollution and Criminal Activity: Microgeographic Evidence from Chicago" by Evan Herrnstadt, Anthony Heyes, Erich Muehlegger, and Soodeh Saberian.

OTHER RESEARCH WE DISCUSS IN THIS EPISODE:

Transcript of this episode:

 

Jennifer [00:00:07] Hello and welcome to Probable Causation, a show about law, economics and crime. I'm your host, Jennifer Doleac of Texas A&M University, where I'm an Economics Professor and the Director of the Justice Tech Lab.

 

Jennifer [00:00:18] My guest this week is Erich Muehlegger. Erich is an Associate Professor of Economics at the University of California, Davis. Erich, welcome to the show.

 

Erich [00:00:26] Hi, thanks for having me.

 

Jennifer [00:00:28] Today, we're going to talk about your research on air pollution and crime. But before we get into that, could you tell us about your research expertise and how you became interested in this topic?

 

Erich [00:00:38] Sure. Happy to do so. So I'm an economist by training. Most of my research focuses on energy markets and energy economics, although it often overlaps with sort of other areas of economics. This - this paper relates to - you know sort of the overlap between energy economics and environmental economics. I also do a lot of work at the overlap between energy economics and public economics, thinking about tax policy, things like gasoline taxes and electric vehicle credits, things like that, and also have some research that focuses on energy markets and industrial organization, competition policy, things along those lines.

 

Erich [00:01:19] My interest in this topic really came about because few years ago I started to get really interested in this literature that thinks about the adverse impacts of air pollution. Not surprisingly, there are a bunch of papers that focus on the effect that pollution has on health outcomes. And I started to get, you know, interested in thinking about whether there might be impacts from air pollution that extended more broadly beyond simply health outcomes. And this is sort of, you know, an interesting - I pulled actually some knowledge from way back in my undergraduate days when I got interested in this topic, because back when I was an undergraduate, I took a social psychology class that had talked about a bunch of experiments that they've done in social psychology where they were interested in thinking about cooperation games where they would put people in a game and potentially expose them to adverse conditions, whether it was - whether it was a noisy environment or a smelly environment or, you know, an environment with really poor lighting or things like that. And basically documented that in those kinds of situations, people behaved less cooperatively. And so in this case, I was wondering whether there might be a similar mechanism whereby pollution might impact other types of social outcomes - social behavioral outcomes. And, you know, crime was one of the things that sort of came to mind as a potential outcome that we might be interested in.

 

Jennifer [00:02:42] So your paper is titled "Air Pollution and Criminal Activity: Microgeographic Evidence from Chicago," and it's coauthored with Evan Herrnstadt, Anthony Heyes, and Soodeh Saberian. And I think it's conditionally accepted at AEJ  applied last I checked.

 

Erich [00:02:56] That's correct.

 

Jennifer [00:02:57] Excellent. So let's start with something very basic. What do you mean by air pollution? What kind of pollution should we have in mind here?

 

Erich [00:03:04] So generally what we're thinking about in this - in this particular paper is we think about air pollution coming from from one of two sources. So we're thinking about air pollution that's being created by industrial sources. So many industrial processes create air pollution that air pollution goes up into the atmosphere, and then potentially drifts to other locations where it might impact things like health or behavioral outcomes as we examine in this paper. We're also interested in thinking about air pollution that comes about from source - transportation sources. So cars driving on highways generate pollution as they - as they emit exhaust, and that pollution can also be blown in particular directions to impact particular individuals.

 

Jennifer [00:03:48] So I admit that when I first heard about this study, I was very skeptical that day-to-day changes in pollution levels could affect crime. But you make a compelling, and I would say quite scary case, that this is probable. So walk us through the potential mechanisms here. Why might we expect air pollution to affect criminal behavior?

 

Erich [00:04:07] So we might expect air pollution to affect behavior more broadly. I think there are a few reasons why we might expect this to be the case. So one is already the literature that I alluded to in the context of social psychology. Social psychologist for a long time have documented that if you put people in an environment where, you know, they are exposed to adverse conditions, that they behave less cooperatively in various types of games, they might compete a little bit more aggressively with one another, sort of things like that. There are a whole bunch of reasons why we might expect someone's environment to potentially affect their behavior in the way they respond to other people.

 

Erich [00:04:49] But there's also a literature that looks from sort of a more medical biological perspective at the impacts that air pollution can have on animal physiology. And so, in particular, there are a set of papers that examine, you know, different types of animals, expose them to air pollution or not expose them to air pollution, and then record both how that physically affects the animal, but then also record how that animal's responses change. And these papers, you know, basically document in, you know, the case of mice and in the case of other mammals, that exposing mammals to air pollution, like the kind that, you know, are commonly - we might be commonly exposed to in an urban environment, can have impacts on sort of how the endocrine system of those animals operates and potentially how brain chemistry operates that can influence those animals' behavior.

 

Jennifer [00:05:48] So before you wrote this paper, what did we know about the effect of pollution on crime?

 

Erich [00:05:53] So we didn't know a tremendous amount. You know, there had been some papers that had been written sort of far in the past, which had documented evidence that, you know, locations that were more polluted also tended to be locations where sometimes there might be elevated levels of crime or potentially, you know, there might be other sort of adverse behavioral outcomes. There was a literature that was, you know, looking at things like particularly polluted days in communities and thinking about other negative consequences. So things like emergency room visits, things like psychiatric visits, you know, sort of things along those lines. But much of this literature was basically focused on really just comparing, you know, polluted locations versus  unpolluted locations.

 

Jennifer [00:06:46] And so why don't we know more than we do? So it - you know, you kind of got the basic correlations there. But as you started to think through this research question, what were the main hurdles you had to overcome in order to measure the causal effects? Is it mostly a data challenge or mostly an identification challenge or both?

 

Erich [00:07:05] So, you know, I would say it's mostly an identification challenge in the sense that, you know, from a data perspective, we have lots of data on criminal activity. Criminal activity is reported by many, you know, many police departments, you know, many law enforcement organizations. It's aggregated at national levels. It's sort of tracked by a variety of different databases. You know, similarly, there's a large set of pollution monitors that have been recording pollution readings for decades, scattered across the country and many in many urban areas. And that data is also publicly available. So from a data perspective, you know, there was a lot of data that was available both on air pollution and on criminal activity.

 

Erich [00:07:51] The challenge here really gets back to this idea that it's not going to be sufficient to simply compare locations that are polluted with locations that are unpolluted. Locations that are polluted, might also be locations that have lots of other characteristics which might be related to the amount of criminal activity that occurs there. In particular, more polluted locations might be more urban locations. They might be locations where there's more economic activity going on or more - or more travel going on. They might be locations that are less affluent in a variety of ways. And to the extent that these other things are correlated with the amount of criminal activity that goes on, it's going to be insufficient for us to simply compare the polluted locations with the unpolluted locations because we might be picking up the effect of those other variables.

 

Erich [00:08:46] So really, the, you know, the identification strategy here was to borrow some ideas from the literature that looks to document the effects of pollution on health. In particular, there are a set of papers that leverage this idea that wind direction can be a very sort of powerful way to separate out the effects that are common with the local community from the effects that are specific to air pollution itself. The idea in this context is to look at days which are very comparable on many dimensions with respect to the amount of pollution that's going - I'm sorry - with respect to the weather, with respect to the time of year, with respect to other conditions. But compare a day where pollution levels are high because of the wind direction, with the day where pollution levels are low because of the wind direction.

 

Jennifer [00:09:39] So in this paper, you're considering the effects of air pollution in Chicago specifically. So tell us more about the layout of that city. What are the big sources of air pollution there?

 

Erich [00:09:49] Sure. So there are sort of two sources of air pollution that we leverage in this particular paper. So just to give you a layout of the city of Chicago - the city of Chicago is located on the western shore of Lake Michigan, and in particular in the city of Chicago - Chicago has some fairly significant industrial facilities that are located to sort of roughly the southeast and the southwest of the city. So one source of air pollution that we leverage in this paper is we leverage variation in air pollution that's being driven by the direction from which the wind is blowing towards the city of Chicago. So on a day where the wind is blowing from the southwest or the southeast, pollution from those industrial point sources is being blown toward the city of Chicago. And if you look at the monitors in the city - the pollution monitors in the city of Chicago - air pollution readings tend to be significantly higher on those days. Air quality tends to be worse on those days.

 

Erich [00:10:46] In contrast, if the wind is blowing off of the lake - basically from the east, then air quality tends to be pretty good in Chicago because there aren't any - there isn't any pollution which is coming off of the water. And so the - the air that's being blown into the city from Lake Michigan tends to be sort of cleaner than the air, which is potentially coming from these other sources. So one source of variation that we use in this paper relates to the location of those industrial point sources relative to wind coming from other directions, like off of Lake Michigan, where air quality tends to be a little bit better.

 

Erich [00:11:22] The other source of variation that we utilize in the city of Chicago is - for any of you have ever been in the - in Chicago, there's an area in Chicago called the Loop, which is downtown. It's right up against Lake Michigan. And there are large interstates that basically radiate out from the Loop. So I'll spend probably a fair amount of time today talking about I-290. I-290 is basically an interstate which runs directly west from the - from downtown Chicago toward suburbs, you know, to the west. But there are also interstates that radiate out to the north. I-90 basically runs through the city from roughly south to northwest. And so a second source of variation we're going to use in this paper is the pollution that's created by those highways. So traffic traveling on those highways generates pollution. And when the wind is blowing particular directions, it blows the pollution from those roads to one side of the highway relative to the other side of the highway.

 

Jennifer [00:12:25] Okay, so you're going to use those daily changes in wind direction as a natural experiment in both of the ways that you just referred to - say, using the industrial facilities out in the city and the interstate highways that run through the city. So let's talk about the pollution from the industrial facilities first.

 

Erich [00:12:41] Sure.

 

Jennifer [00:12:42] So, you've got the wind direction as giving you something akin to random variation in pollution levels in the city and just as you just described. So how do you use that variation to measure the causal effects of air pollution on criminal behavior?

 

Erich [00:12:55] Yeah. So conceptually, the idea is going to be the following. We're going to - we're going to look at - we can sort of think about the following thought experiment, which is essentially what we're going to do empirically. We're going to run a regression where we're going to be trying to predict the amount of criminal activity that happens in the city of Chicago on a given day. What we're going to do is use a whole bunch of controls, including things like the time of year, which we know there are seasonal patterns with respect to criminal activity and with respect to pollution. We'll use things like temperature and other meteorological conditions. There's a long literature in economics that documents an increase in crime on hot days relative to cool days. And we also know that temperature is related to pollution - related to pollution formation. We're going to include other explanatory variables. And essentially what we're going to be doing is we're going to be comparing two days that are happening in a similar time of the year and happening under very similar meteorological conditions. But the difference between those days is on one of those days, the wind is blowing from, say, the southeast or the southwest. And air quality in the city of Chicago is relatively poor. On another one of those days, very similar temperature, very similar meteorological conditions. But instead, the wind is blowing from the east. So it's blowing off of Lake Michigan.

 

Erich [00:14:10] And that sort of variation in wind direction is what's going to give us what we call exogenous variation in the amount of pollution that's experienced by people living in the city of Chicago. So effectively, what we're going to do is be comparing days where depending on the direction the wind is from, there might be more or less pollution in the city of Chicago and testing whether or not we see elevated levels of crime on the day where the wind is blowing from either the southwest or the southeast, relative to a day where the wind is blowing from the east and air quality tends to be pretty good.

 

Jennifer [00:14:46] And what data are you using for all of this?

 

Erich [00:14:49] So we're using a few different data sources. So as I sort of mentioned there's a lot of data available on pollution, both from particular point sources, but also from pollution monitors. And for some - for this part of the study, we're using data from pollution monitors that were installed by the EPA. The EPA tracks the pollution at these monitors. And so you can see in a very sort of high frequency basis what the pollution readings are in the city of Chicago. We're also bringing in this data that was collected by the city of Chicago, where the city of Chicago collects data on the number of crimes that are reported to the Chicago Police Department. In fact, Chicago is a city that provides a lot of data - publicly available. And so in the city of Chicago, they provide data on every crime that was reported to the Chicago Police Department. They provide information about where the crime occurred, the latitude and longitude of where the crime was reported. They include information about the time and the date of the crime and also information about the details of the crimes, so what type of crime was it, you know, and sort of other specific details.

 

Erich [00:16:07] And so, you know, for this part of the study where we're looking at Chicago as a whole, we're basically aggregating the data on crime together to a daily level value for the city of Chicago. We are taking daily information about the level of pollution in the city of Chicago, and we're taking information from weather stations that are located at the Midway Airport and Chicago O'Hare Airport that track things like what the temperature is in the city of Chicago, what direction the wind is blowing, how hard is it blowing, you know, and sort of other meteorological information. And we're using those three pieces of data together to understand the relationship between pollution that's being blown from either outside the city or from some sort of other directions and the amount of crime that's happening in the city of Chicago on a given day.

 

Jennifer [00:16:59] Okay, so what do you find? What's the effect of air pollution from industrial facilities on city level crime rates?

 

Erich [00:17:05] So what we find is that on a day where pollution is one standard deviation higher than the mean, so relative to the mean, one standard deviation is about a 50% increase. So this would be a typical day where the wind is blowing from the southwest or the southeast, and some of that industrial pollution is being blown towards the city of Chicago. What we estimate is that the amount of violent crime increases by about two to three percentage points in the city of Chicago relative to a day where the wind is blowing from a direction where there's less pollution in the city. With respect to property crimes - we separate out the two kinds of crime - we separate out crime by whether or not it's a violent crime. So things like assault, battery, and other crimes like that from more property oriented crimes of theft, burglary, larceny, sort of things along those lines. When we look at property crimes, we don't find a relationship between wind driven pollution and the amount of property crime that happens in the city of Chicago.

 

Jennifer [00:18:14] And so I gather that's in line with what you would expect based on kind of the science here.

 

Erich [00:18:19] You know, I think it is generally what we would expect. I mean, we might think that, you know, the mechanism which is going to drive violent crime - a violent altercation relative to the mechanism which is going to drive property crime might be a little bit different. Now, the literature from both psychology and from the medical biological literature with respect to animals and air pollution, suggests that air pollution exposure might lead to sort of less cooperative - less cooperative behavior, more aggressive behavior, sort of things like that, which we might think, you know, map more closely to a violent altercation as opposed to, you know, theft or burglary or sort of something like that in a property crime.

 

Jennifer [00:19:04] Yeah. Okay, so then you approach this question from another angle. Using pollution from interstate highways as the natural experiment. So remind us how changes in wind direction give you quasi random variation in pollution levels in this case.

 

Erich [00:19:19] Yeah, so that's a great question. So the idea here is to essentially say, you know, we have these interstates that run outside of the city, basically run from the Loop in downtown Chicago and sort of radiate out from the city like spokes. And I'll probably spend a little bit of time talking about I-290. Today, I-290 is, you know, an interstate which basically runs from the Loop basically straight west out to suburbs that are located in the west - in the west of the city. And so what we're going to do here is we're going to lean on the fact that those interstates are going to be generating pollution.

 

Erich [00:19:55] And so using I-290 as an example - on a day where the wind is blowing orthogonally to the interstate, so basically across the interstate in one direction or the other, the pollution from that interstate is going to be blown to one side of the interstate. So on a day, for example, when pollution is blowing from south to north, the pollution from I-290 is going to be blown onto the north side of the interstate, whereas the south side of the interstate is not going to be exposed to that pollution. On the other hand, if, let's say, the next day, the wind switches and starts blowing from north to south. Now it's going to be the south side of the interstate, which is going to be the downwind side. And the north side of the interstate is going to be the upwind side and potentially exposed to less pollution. So what we're going to be doing here is we're essentially going to be looking at days where the wind is blowing sort of across one of these interstates. In the case of I-290, it's going to be on a day where it's blowing from either north to south or from south to north. And we're going to be comparing the amount of violent and property crime that happens on the downwind side of the interstate relative to the upwind side of the interstate.

 

Erich [00:21:05] The thing that's really attractive about this from an identification perspective is that if we think about lots of other things that might be happening, so say they're sort of meteorological conditions that are happening on a particular day or that the wind, you know, that the wind tends to blow a particular direction at particular times of the year or it happens to be blowing a particular direction on a day where there's a lot of economic activity or relatively little economic activity. The nice thing about this identification strategy is that those types of factors are going to be common across both the upwind side and the downwind side of the interstate on a given day. And so we can essentially use the upwind side of the interstate - the side that's not exposed to the pollution as a control for the downwind side, as a control for sort of how much activity - how much criminal activity might happen on the downwind side of that interstate on a particular day. So it becomes sort of a much tighter comparison. Really, now, instead of thinking about all of Chicago, we're really just looking at a collar around these interstates on days where the wind is blowing from one direction to the other. And we're really focusing on this idea of whether or not we see an elevated amount of crime on the downwind side of the interstate relative to the upwind side of the interstate. And because the wind switches direction from day to day on one day, we might be looking, for example, for evidence that crime is elevated on the north side of I-290. And on a different day, we might be looking for evidence that crimes elevated on the south side of I-290.

 

Jennifer [00:22:48] And you couldn't do this with the industrial facilities just because there is no kind of south side. Is that basically the intuition there?

 

Erich [00:22:56] Yeah, I mean, the intuition is that, you know, even in sort of a world where wind direction is generating variation in pollution in downtown Chicago, there might be other factors which are correlated with wind direction that we haven't taken into account. Right. So, in principal, I've - I've never lived in the city of Chicago, but it - it's plausible to imagine sort of a story where even on days that look very comparable from sort of meteorological conditions - so same temperature, same level of humidity, sort of things like that, when the wind is blowing off the lake, maybe people go down to the lake. It's just a particularly nice day to go down to the lake. You go down to the lake, you're in a good mood, and you're more willing to, you know, look the other way if you end up in a confrontation with someone - sort of de-escalate the situation. So that kind of thing, that's - those kinds of omitted variables are always going to be sort of problematic, even in a situation where you think that air pollution is providing some, you know, quasi random variation in the levels of pollution, you always worry that there's going to be something else there that might be correlated with wind direction that you just haven't taken into account. And it's really that other thing which is related to the amount of criminal activity that's occurring, not the effect of air pollution coming from wind direction.

 

Erich [00:24:21] You know, once we sort of zoom into comparing pollution on the north versus the south side of let's say I-290 on a particular day, then we're really at a point where, you know, in principle, meteorological conditions are going to be identical across those two regions on a given day. Similarly, we wouldn't expect - you know, even if there is seasonality with respect to which way the wind blows - within a particular season, there are going to be days where it's blowing in one direction, they are also going to be days where it's blowing in the other direction. So, again, we can essentially use sort of what's happening on one side of the interstate as a control for the other side of the interstate.

 

Jennifer [00:25:05] All right, so what do you find in this case. What's the effect of air pollution from the highways on crime rates?

 

Erich [00:25:11] So in this case, we find something quite similar to what we were finding at the city level. Now it's a little bit - you know, it's a little bit difficult to do an apples to apples comparison because in the case of the city level analysis, we can observe pollution at particular monitors. So we can think about things in terms of, you know, what is the impact of a one standard deviation in pollution levels in the city of Chicago. In the context of the interstate analysis, there are a lot of pollution monitors across the US, but there are so many pollution monitors across the US that we have pollution monitors on both sides of every interstate in the city of Chicago. And so what we're basically focusing on here is the impact of being downwind of a particular interstate on a particular day. And what we're finding is that, you know, including sort of controls that essentially allow the upwind side of the interstate to act as a control for the downwind side, we're essentially estimating roughly a 2% increase in the amount of violent crime that happens on the downwind side of the interstate relative to the upwind side of the interstate.

 

Erich [00:26:30] We also do the analysis for property crimes. And just as in the case of the city level analysis, we find no relationship between being downwind of an interstate and the amount of property crime that occurs on a particular day.

 

Jennifer [00:26:44] So you do a whole bunch of additional checks to convince yourself that these results are indeed the effect of air pollution and not some other potential confounding variables, as you mentioned, is always the concern here. So walk us through, you know, one or two or three of those checks - whatever your favorite ones are, and what you find there.

 

Erich [00:27:02] Sure. So we do a few different - a few different types of checks. I think when we initially performed this, we were a bit skeptical as well. We sort of said, you know, look, this is, you know - we think we're estimating something, an effect of pollution on violent crime. But we want to make sure it's the case. And so we did a couple of checks you know to sort of think about these - and you can sort of think about these checks as sort of asking the following question. If this really is being driven by - if this effect is really being driven by wind borne air pollution from these particular interstates, what else would we expect to be true? And so I'll just give you a sense of a few of these. So one thing we can do is if the interstates are the source of this air pollution, we can look to see whether or not the effects are more pronounced closer to the interstate than further away from the interstate. In our base analysis, we basically look at a collar around the interstate that extends one mile in both directions. But in principle, if this really is something which is being driven by pollution from the interstate on a day where one side of the interstate is downwind, we should expect a greater impact close to the interstate where there might be lots of pollution impacting that area, then potentially a little bit further from the interstate. So this is one of the - one of the checks we perform. We find something very consistent with that, that when you look at the relative effect close to the interstate, within that - within that sort of mile band and sort of further away from the interstate, within that mile band, you tend to see larger impacts close to the interstate than further away from the interstate.

 

Erich [00:28:46] Likewise, you could imagine sort of doing similar types of analysis on days where let's say the wind is blowing much more forcefully. On a day where the wind is blowing much more forcefully, you might expect that the wind is going to disperse the pollution from the interstate much more effectively. It's potentially going to blow it over a much larger distance than on a day where the wind is blowing lightly, in which case much of the pollution might end up drifting a little bit to one side of the interstate, but not very far. And again, in that situation, we're finding sort of the largest effects of being downwind on days where the wind is blowing moderately but not extremely - extremely strongly such that the pollution might be being dispersed.

 

Erich [00:29:28] Another exercise we can do is we can look to see - using sort of I-290 as an example whether it seems like this effect is greatest, if you look right at I-290 versus if you look at other - what we would call placebo interstates that are either parallel to I-290 to the north or parallel to I-290 to the south. So you can think of this in the following sort of thought exercise where we can basically draw different lines which are parallel to I-290 at intervals, extending from about 15 miles south of I-290 gradually through I-290, and then north about 15 miles from I-290. And in each one of those placebo interstate's, we're going to estimate what the downwind effect of pollution would look like if the interstate had been located there. Now, the reason I call these placebo interstates is if this really is an effect of the interstate, okay, then what should - we should expect to see is that these placebo interstates, we shouldn't expect to see results or at least we should expect to see much, much potentially smaller results than if we draw the line right at I-290. And that's actually exactly what we see, which we found sort of - we felt like very reassuring that this is something about these interstates and being downwind of these specific interstates. If you basically estimate what the effect would be of being downwind of all these - all these hypothetical interstate's relative to the - where I-290 actually goes sort of east west out of the city of Chicago, you find the effect of being downwind is greatest right at I-290 itself. And in fact, if you look sort of at a hypothetical interstate one mile north of I-290, you find on that particular day that the effect of being downwind of that hypothetical interstate actually goes in the opposite direction, which is exactly what we would expect if pollution from I-290 was being blown onto the upwind side of that hypothetical interstate.

 

Jennifer [00:31:39] Okay, so that is your Chicago paper. That's been around a year or two now, I think. So what other studies have come out since you first released this paper that shed light on how pollution affects crime?

 

Erich [00:31:52] So there have been a number of other papers that have looked at the effect of pollution on crime using identification strategies, which I would say sort of in spirit are similar to our city level analysis. So leveraging this idea that we might aggregate crime over larger areas and then look at situations where either the wind or another factor is causing pollution to be a little bit higher in that city on a - on a given day or a little bit lower in that city on a given day.

 

Erich [00:32:24] There's a study that's been done that's looking at the city of London, looking regionally within the city of London. There's also another study which has looked a little bit more broadly, looking at multiple metropolitan regions in the US, but approaching the question from a very similar perspective to our city level analysis. And what those other papers are finding is something very consistent with what we're finding in this paper, that it seems to be on days where air quality is low, where the air seems to be polluted, that there's an uptick in the amount of violent crime that occurs and less of an effect on the amount of property crime that occurs.

 

Jennifer [00:33:03] So what are the policy implications here? What should policymakers and practitioners take away from both your research and the other work in this area?

 

Erich [00:33:12] So I think this is a great question. So I think there are sort of two takeaways from this research, from a policy perspective. So the first is that whenever we think about air pollution and the adverse impacts of air pollution, we think of those, you know, in economics, we call these externalities. These are impacts that are being imposed on other individuals that the person who's generating the pollution is not necessarily taking into account when they're making their decision about how much pollution to potentially generate. And so environmental externalities are the motivation for much of environmental regulation. In particular, things like pollution taxes, pollution - trade - tradable permit programs, the sort of things along those lines - in all of these cases, basically those other - those regulations or taxes or tradable permits are designed to get people to essentially take into account the impacts that they're having through some sort of financial means where they're either paying taxes on those emissions or there are regulations which are basically reducing the amount of pollution they're allowed to emit. Again, to sort of mitigate these adverse impacts that pollution might have. So one policy implication of this paper is that there's a long literature which documents the health impacts of pollution and a growing literature which documents a variety of different cognitive impacts, so impacts on test scores and sort of things like that. You know, what I think this paper adds is it adds sort of another way in which pollution might lead to adverse outcomes. And so when we think about pollution regulation, environmental regulation of pollution, these - this would suggest that we want to be a little bit more stringent with environmental regulation than we would have otherwise been if we were not taking this into account. That would be sort of one policy application.

 

Erich [00:35:15] I think the other policy implication here is that one of the things this paper does is document that pollution can impact a wider variety of outcomes than simply thinking about the physical health impacts of pollution or potentially the cognitive impacts of pollution. To the extent the pollution impacts people's behavior in adverse ways, then we might think much more broadly about - we might be concerned in much more general ways about the impact that pollution is having not just on our physical health, but potentially on things like the decisions that we're making, and potentially more broadly on our behavior.

 

Jennifer [00:36:00] In the crime context in particular, do you know if any of these papers have tried comparing, say, the cost of reducing crime through reducing pollution to other crime reduction strategies, like hiring more police or something like that?

 

Erich [00:36:14] Yeah, so I am - I'm not familiar with papers that have looked at that comparison specifically. Admittedly, I think I would say the impacts that we're finding in this particular paper are relatively modest. So we're finding sort of a potentially a relatively large impact - a relatively large change in pollution is translating into relatively modest impacts in the amount of violent crime that's happening, even though we still care very much about sort of a two to three percent change in violent crime on a given day. These changes in violent crime are at least modest relative to other estimates from the environmental literature with respect to how much, let's say, a particularly hot day has on violent crime relative to a cooler day on violent crime or things like that.

 

Erich [00:37:03] One of the exercises we do do in this paper, those we do sort of ask sort of how comparable are the costs associated with the impacts on violent crime to other types of impacts that have been estimated with respect to pollution on, let's say, health outcomes or things like that. And what we find there is that the health impacts of pollution are considerable, but at least relative to some subsets of those. So Impacts on fetal health outcomes, on infant mortality, sort of things like that, at least our back of the envelope calculations suggest that the cost of crime created by pollution might be of at least similar magnitude to some of the impacts on infant health.

 

Jennifer [00:37:54] Yeah. I mean, violent crime is really expensive.

 

Erich [00:37:57] Yeah, no absolutely. Yeah.

 

Jennifer [00:38:00] And especially when - I mean, we were talking earlier about how correlated, you know, elevated pollution levels are with other problems that a community might have, like poverty and other negative, you know, health shocks and so on. And so thinking about all of these different costs really compounding for the same communities over and over is striking.

 

Erich [00:38:22] There's a - there's a tremendous amount of policy interest in the state of California and thinking specifically about the impacts of pollution on communities that have historically borne the brunt of pollution, might be slightly less affluent communities, might be communities that have a higher fraction of minorities living there. So there's a lot of concern within the state of California about making sure that to the extent we are trying to improve the environment through environmental regulations, that those environmental benefits are spread throughout citizens of the state and not simply captured by more affluent communities.

 

Jennifer [00:39:13] Right. So what's the research frontier here? What are the next big questions that you and others will be thinking about going forward?

 

Erich [00:39:21] Well, I can sort of think of two from related to this study. I mean, one is I think there's still significant work to be done in trying to understand the mechanism by which we see pollution impacting criminal activity. So I sort of alluded to a few different mechanisms. There's the psychology literature which suggests that if I - if I am not feeling good because it's polluted out, I might be less cooperative, more prone to act aggressively in a confrontation or things like that. There's also the literature from biology and medicine that documents the physiological impacts on the endocrine system and the central nervous system in animals. So I think there is - there's still some work to be done at trying to understand how does pollution actually impact our behavior and how does it actually impact the decisions that we make. So I would say sort of that's one area of exploration.

 

Jennifer [00:40:24] I mean, I think the other sort of area of exploration is what I - what I feel like, one of this - one of the things that this paper does is demonstrate that pollution can impact our behavior in a lot of ways outside the sets of impacts that we've thought of historically as being sort of adverse impacts from pollution, so health impacts or things like that. And so what I would expect and what we've started to see are more and more papers that have started to think about other types of impacts on the decisions that we make. So to what extent does pollution impact cognition? To what extent does pollution impact other types of decisions that we make? Because to the extent that even though in our - in our particular situation, we're finding what I would characterize as relatively modest impacts on violent crime, if we think that pollution might impact behavior very broadly, there are lots of decisions we make every day. And if pollution is impacting all of those, even just a little bit, in aggregate, that could add up to quite a bit.

 

Jennifer [00:41:32] My guest today has been Erich Muehlegger from UC Davis. Erich, thanks so much for doing this.

 

Erich [00:41:37] Hey, thank you for having me.

 

Jennifer [00:41:43] You can find links to all the research we discussed today on our website, probablecausation.com. You can also subscribe to the show there or wherever you get your podcasts to make sure you don't miss a single episode. Big thanks to Emergent Ventures for supporting the show. And thanks also to our Patreon subscribers. This show is listener supported, so if you enjoy the podcast, then please consider contributing via Patreon. You can find a link on our website. Our sound engineer is Caroline Hockenbury with production assistance from Elizabeth Pancotti. Our music is by Werner and our logo is designed by Carrie Throckmorton. Thanks for listening and I'll talk to you in two weeks.