Search the Site

Episode Transcript

My guest today, David Keith, is among the world’s experts on the subject of solar radiation management, also known as solar geoengineering. He’s a professor of applied physics at Harvard and co-founder of Harvard’s Solar Geoengineering Research Program.

David KEITH: Solar geoengineering is a very tempting thing because if you want quick reduction in real harms to some of the world’s poorest ecosystems, it might achieve it. But on the other hand, it’s absolutely true that if the world just chose to embrace solar geoengineering and forgot to cut emissions, we are screwed.

Welcome to People I (Mostly) Admire, with Steve Levitt.

There are few areas of science that are as controversial as geoengineering. Supporters believe it could be a valuable tool for fighting climate change, but opponents say it’s dangerous, unproven, and will weaken the will of society to take the other hard steps required to save the planet. In January of this year, over 60 scientists signed an open letter calling for a complete prohibition on government funding for solar geoengineering research, and a total ban on outdoor experiments in this area.

I discovered firsthand just how sensitive the topic of geoengineering is when Stephen Dubner and I wrote about it in our book SuperFreakonomics. Now I’ve taken many controversial and unpopular stances over my career, but nothing I’ve ever written has generated as much of a negative firestorm as the suggestion that geoengineering might play an important role in fighting climate change. Now, that was over a decade ago. Clearly, views haven’t completely changed since then, but that hasn’t stopped David Keith from continuing to research the topic. I’m interested to hear what advances have been made in solar geoengineering and what the future might hold for the technology.

LEVITT: So, I got to attend an academic lecture that you gave recently at the University of Chicago. You started your lecture by noting there are four broad strategies available to deal with climate change. The first strategy is that we can put fewer greenhouse gasses into the atmosphere or what you call decarbonization. The second strategy is that we can remove carbon from the air and store it somewhere. And then the third strategy is solar geoengineering — or solar radiation management. Essentially, it’s to do things which lower the temperature of the earth that don’t involve lowering greenhouse gas concentrations in the atmosphere. And then the fourth strategy is what you call adaptation. So essentially, to take steps to limit the damage caused, if the first three strategies don’t get the job done. And what I liked about your framework is that it’s incredibly simple, but it helps organize my thinking in a useful way. It helps me focus on the trade-offs that are apparent in this area.

KEITH: Yeah, thank you. But I certainly wouldn’t claim it’s mine. While people have argued about the names, the idea that there’s those four independent buckets, in fact goes back, oh, at least to the ’70s. People use a bunch of buzzwords, “mitigation,” or “adaptation” as one division and “solar geoengineering” is always contested, so people have lots of different names for it. But I often emphasize the four physical responses. Of course, there’s a complicated network of political pathways that one would take to implement any of them.

LEVITT: So Stephen Dubner and I, we wrote about solar geoengineering in our book SuperFreakonomics back in 2009. And we were strong supporters. And yet, when you and I spoke a few years ago, you said that you thought that SuperFreakonomics had set back the field of geoengineering by five years, that it slowed down events and research. You know, with friends like us, who needs enemies? What happened after our book that had such negative repercussions for you and other scientists?

KEITH: Your book had this fun way of pointing out lots of counter-intuitive economic examples. And that was the spirit of the book. But I think in this topic, just saying this kind of counterintuitive fact, “Oh my God, it’s much cheaper to cool the world this way than cutting emissions,” is a kind of oversell that really makes it harder to get people to take it seriously. And I think your book did come across that way a little too much. Not that what it said was wrong. And the thing that’s helped the big environmental groups inch towards taking this seriously and governments inch towards taking this seriously has been really putting it in context and being very careful not to overclaim. Several of us in the field, it wasn’t just me, were not very happy with SuperFreakonomics, to be honest.

LEVITT: It’s funny because I hadn’t looked back at that section of SuperFreakonomics in five or 10 years, but in preparation for our conversation, I re-read it. And admittedly, I think what you’ve just criticized was completely true. It was a little bit starry-eyed and optimistic about the prospects for what we could do. I feel bad that we set back research because I really do think this is an important area for research. I would just say you, obviously, who were into this 20 years before Stephen Dubner and I wrote about it, but even we, 20 years later, were ahead of our time.

KEITH: The world on this topic seems to be changing very quickly, even in the last few months. There will be a formal announcement of a global commission on this topic that has really got an amazing collection of people. It’s brought out of the Paris Peace Forum. And it’s going to be the most serious discussion about some of the high-level governance questions of how solar geoengineering and carbon removal might fit together. There are major university programs starting. I have this little program that I’ve got going at Harvard, but there’s now several others that look really serious in a way that wasn’t true before. The U.S. federal government is for the first time, seems like actively working on the design of a research program. And some of the big environmental groups seem to have a very different attitude towards this technology than they did even a few years ago. So I think it has changed a lot. And I think the way to advance it is to really think about it always in context of other things we do, and to be careful about overclaiming what it might achieve.

LEVITT: So virtually everyone agrees that reducing greenhouse gas emissions is critically important. But in practice, it’s been hard so far to induce that sort of behavior change. Do you think that’s a fair statement?

KEITH: No. Or rather, I think your statement presupposes that this is about behavior change and I think it mostly isn’t. This is about rebuilding the industrial infrastructure that underpins, really, human civilization, that is not an overstatement, to make it not emit carbon. Which I think is absolutely physically possible. What humans want are energy services, like transportation, communication, illumination. Most of us don’t care at all how those services are made. And the thing that we need to do is to reshape this heavy industrial infrastructure, to make it in a way where the primary energy going into it is not releasing carbon. Now we have multiple independent ways to do it. Solar and nuclear power being the most obvious. You can see that behavioral change can’t be central, because you basically have to eliminate carbon emissions. And you’re not going to eliminate consumption. So there’s some way in which behavioral change can help around the edges. But it’s about changing what energizes our civilization.

LEVITT: So often you do see this pitched as behavioral change. So if I think about Al Gore’s movie An Inconvenient Truth, a lot of that was pointing at consumers and how consumers needed to do something different. You make an interesting statement and a true statement, which is that it could all be in the background, right? If we had nuclear plants and we had endless solar and wind power, then we could do a lot of this without any individual changing what they’re doing. How are we doing as a civilization at changing the infrastructure towards decarbonization?

KEITH: We’re doing a crappy job, but when we talk about doing a hard thing in the future, it’s really helpful to think of what we’ve done in the past. The world has made enormous progress in reducing air pollution and water pollution, and some key toxics, metals like lead or some very long-lived organic chlorines, D.D.T., or damage to the ozone layer. We’ve made real progress in all of these. In some cases, really dramatic progress. The U.S. Clean Air Act has added a year and a half to life of an average American. That’s a stunning number. And that was done almost completely without consumer preferences. This didn’t happen because people just prefer to buy products that had lower air pollution footprints. It happened because we passed rules that said, “You’re not allowed to operate a factory that emits this kind of pollutant,” or “You’re not allowed to sell a car that doesn’t have a certain kind of catalytic converter.” And I think in some of the frustration of low progress about climate, the climate activist community has lost sight about a central fact that a bunch of this is about changing the rules, which drive the change of infrastructure. And there’s just a limit to what a consumer can do, because the consumer is just buying the final product that comes from this huge supply chain that they don’t really see or control.

LEVITT: So you’re saying if we want to fight climate change, we should change laws, not try to convince people to be vegans, even though obviously if everyone was vegan, it would have a great benefit to the climate.

KEITH: Well, interestingly, that one example is the most plausible place for consumer choice to do something useful. But I think it is fair to say that if people ate significantly less meat, that is a pretty easy-to-accomplish, at least in principle, without any obvious, other big impact on people’s lives, way to take a, you know, significant bite out of the greenhouse gas footprint of each person. And that’s something I actually could imagine a pretty rapid social change driven by concerns about climate, because that’s a choice you can just make at the grocery whereas you can’t just choose to make your house or your business or your airplane travel carbon neutral.

LEVITT: So, basically, through decarbonization alone, there’s likely no way that we’ll hit the sorts of benchmarks that the scientific community thinks are required to manage the risks of climate change. So being pragmatic, we’re not on track. And it seems like there’s going to be a need either for massive amounts of carbon capture or solar geoengineering. And that’s where people like you come into the picture.

KEITH: Well, I do want to emphasize how much progress is being made on decarbonization. Now, it’s much less progress than I would like. As a voter, I would vote for much higher, effective carbon prices, much more action. But the world is now spending well over half a trillion a year on low carbon energy. That number’s significantly higher than it was a decade ago. And in North America and Europe, carbon emissions have peaked in or are headed down, not anywhere near as quick as many of us would like, but I used to give climate talks 15 years ago saying it’s a phony war and people talk about cutting emissions, but they don’t do anything at all. And I can’t say that anymore. We are doing things. Not as much as we should, but there’s real action happening.

LEVITT: So we’re making real progress, but realistically, it’s not nearly so fast as the scientists think that we need.

KEITH: It’s important to say scientists don’t set the target. What science can do is tell us something about the consequences of climate change, how much the climate will change for a given set of emissions, and quantitative social sciences, such as yourselves, can say what the human consequences and ecological consequences are, given an amount of climate change. How many people die, or how much productivity declines with heat. But scientists can’t tell you the right target to aim for. That’s inherently a political choice because it involves trade-offs between cost of cutting emissions and cost of climate change that are distributed unequally over time and space. So it is actually very common to say, “Scientists say we should hit 1.5 or two degrees centigrade above pre-industrial,” but that’s in fact, false. Those are political targets. They may be good or bad political targets. We can argue about it. But they are most definitely not the kind of cold result of science.

LEVITT: Okay, let me ask you about that, because those numbers, my impression come out of the I.P.C.C. climate reports that are written by scientists. But those are really political activities you’re saying?

KEITH: The main body of the I.P.C.C. reports at least the volume one, which is mostly about climate science, is a very good summary of what we know about climate science, which is, how the climate will change with given amounts of emissions. But it doesn’t tell you that we should hit 1.5. There’s a very famous I.P.C.C. report, which is the report on hitting 1.5. And that’s the report that generated all the press about eight years left to save the climate. And this idea that there’s a very fixed time window to keep 1.5. But crucially, 1.5 was given to them by the politicians. It was a politically agreed target and I.P.C.C. was instructed to write a report about the consequences and difficulty in meeting 1.5. But IPCC wasn’t instructed to say, “What is the right temperature target?” And produced 1.5 is an answer. That is not what happened.

LEVITT: Oh, interesting. I had misinterpreted that.

KEITH: So you’re not the only one. And from my perspective, my confidence in the I.P.C.C. process went down. I am very keen on activism. I believe we won’t get the kind of changes we need to cut emissions without activists, including, civil disobedience. People locking themselves to coal trains, or tying themselves up in front of the White House. I think we need people like Greta Thunberg. But I.P.C.C. job is not activism. And I think I.P.C.C. crossed a line in not the actual text of the 1.5 report, which is good, but the way it was reported out. So much so that some authors of that report had to then go back separately and write articles, saying, “No, we didn’t mean that.” Because the I.P.C.C., I think deliberately, let the reporting have this kind of “scientists say” framing, even though if you pick up the intro of the report, that’s clearly not what it says.

LEVITT: I see. Now you were part of the I.P.C.C. reports and then you’re no longer. Is that because you didn’t like the way things were happening?

KEITH: Yeah, I think it’s fair. I was part of early ones and I was part of the one before this, and I just found that it was less and less interesting and less and less a place where you could really have productive discussions about what would happen next.

LEVITT: So if we want to hit any of these benchmarks, we are looking at either massive carbon capture or the use of solar geoengineering. What are the menu of options we have when it comes to solar geoengineering?

KEITH: Going from the top down, if you like, the highest altitude method is to put some reflective shield in between the Earth and the sun. And this sounds really like ludicrous science fiction and I think in the near term, it is. But it’s actually been in the literature for 30 years. And the cost of access to space is changing pretty quickly. So I think it’s laughable in the next decade or few decades, but it’s not laughable over the kind of hundred-year timescale that we’re talking about. Next is to put reflective aerosols — and aerosol is just a fancy word for a small droplet or a particle — that’s just so small that it doesn’t fall quickly. To put aerosols into the upper atmosphere, the stratosphere, where they would reflect some sunlight back to space. That’s the best studied thing, a stratospheric aerosol injection, it’s also often called.

LEVITT: Now, we understand this pretty well because nature has given us some natural experiments in this regard with Mount Pinatubo and other large volcanic eruptions.

KEITH: That’s right. So volcanic eruptions teach us a lot about aerosols in the stratosphere, but that’s just one piece of all sorts of reasons why we know about the effect of aerosols on climate. So humans emit an immense quantity of aerosol pollution to the lower atmosphere. And that actually has huge human consequences. Kills on an order of 5 million people a year. And it also cools the climate and changes the way clouds look, and lots of enormous body of research on aerosols and climate dating back for half a century.

LEVITT: And one of the good things about this approach, this aerosols in the stratosphere is we know exactly how to do the engineering to make it happen. One way to do it is a medium-sized fleet of airplanes that fly high and deliver sulfur dioxide. What kind of scale would it take to make a real dent in what we’ve done to heat up the planet?

KEITH: So to get a degree of cooling, very roughly, would take a fleet of something of order of a hundred aircraft, depends how big they are. Make them bigger or you could do it with more like 30 aircraft. They’d need to be specialized aircraft. As a technologist, I’ve got a whole list of things we don’t know, but I think in the big picture, there’s no fundamental doubt that that could be done.

LEVITT: And cost-wise, we’re not talking very much in the grand scheme of things — $5 billion, $10 billion per year? To achieve this kind of cooling through this mechanism?

KEITH: That’s the right kind of number, which on the scale of the climate numbers is really tiny. I mean, if we want to decarbonize as rapidly as many of us think we should, we need to spend globally something like a trillion dollars a year in decarbonization and the econometric estimates of climate impacts mid-century are also on the order of a trillion dollars.

LEVITT: And it’s good to remind people because when numbers get big, people have a hard time thinking about them. But the difference between a trillion, which is a thousand billion and 5 billion is a lot. That’s 200 times bigger.

KEITH: The right way to think about the money is the money small enough that the money will not be central to the decision. This is really about the risks of doing it, which are very real, against the risks of not doing it, which are very real. And it’s about this risk balance, and questions of governance and trust. In this complicated way that humans will make decisions about solar geoengineering, at least if you’re talking aerosols in the stratosphere, it’s not likely that the ability to pay for it will play a big role in the decision.

LEVITT: And the other both good and bad thing about this approach is that if you stop that fleet of airplanes and you ground them, then this cooling effect will disappear very quickly, which is good in the sense that if you realize you’ve made some kind of mistake and you want to undo it, you can undo do it, but it’s also bad in the sense that you really need to keep on doing it in perpetuity, or at least with a sensible plan about how to stop it. And if something went wrong, you would have a very quick heating of the earth, if you had become dependent on this kind of technology.

KEITH: Yeah, I think that’s right. It’s certainly true that if one was doing a lot of solar geoengineering and suddenly stopped, then you’d get a sudden warming with bad consequences. But I think one needs to really consider what would be the features that would make that happen. And I think it’s actually, politically, very hard to stop. And it’s very hard to see a kind of system failure stopping them. Given that it’s cheap, the fact is it would make sense for multiple countries to maintain the ability to do it, even if they oppose doing it. If there was a lot of solar geoengineering happening and there was a legitimate fear about the existing players somehow losing the ability to do it, then it would be in say, ultra-green German’s rational self-interest to build the equipment to do it, even though they oppose doing it because there’s a danger of turning it off. And the global positioning system provides a clever analogy for why this is true. So the U.S. built the first constellation of global positioning satellites. I can get out my cell phone in a city I’ve never been before and know exactly where I am because satellites are whizzing by at many times the speed of sound over my head, and give me my precise location to a meter. And the Russians built one of these too, originally, but now we have a Chinese system and the Europeans are building separate system for global positioning. So there’s now four of these independent systems. So why is Europe spending money to build a system? Because they can use the existing systems for free. The answer is they want the power to decide whether it’s turned off. So the U.S. could in principle, they have a system for doing it, deny access. And so could the Russians. So the Europeans want to have a vote — if you take G.P.S. to mean having at least one of these systems on, effectively, all players must vote to turn it off for it to be turned off. If any player votes to keep it, then the service stays. And I think that is exactly the same as what happened with solar geoengineering. Once you’ve started, given that it’s fairly cheap, essentially, all the major entities need to agree to stop for it to stop. And I think that’s unlikely to happen by accident.

LEVITT: Yeah. What may be the greatest risk many people see of solar geoengineering is that we lose our will to do the harder thing, the decarbonization. I think that’s probably a legitimate point.

KEITH: That is absolutely the central concern. I think that goes by many names, “moral hazard,” I think in many ways, “addiction” is the right word, but that is the central concern and the biggest reason that I, personally have doubts about whether I should have spent my career talking about this. Because solar geoengineering has this, in a sense, terrible property and that in the short run, it’s always much more cost-effective and even just more effective than anything you do in cutting emissions. The reason is the single most important fact that all your listeners should know about climate change, which is climate change is proportional to the cumulative emissions of carbon dioxide, basically over its industrial history. The cumulative amount. That means that if you stop emitting in a given year, it makes almost no difference to the climate in the next few years. And if you stop emissions completely, you don’t stop climate change. You just stop it getting worse, because it depends on the cumulative amount emitted. And that makes it very hard to justify spending money in the near term to cut emissions, even though it’s absolutely the right thing to do if you care about the long term. And so, solar geoengineering is a very tempting thing because if you want quick reduction in real harms to some of the world’s poorest ecosystems, it might achieve it. But on the other hand, it’s absolutely true that if the world just chose to embrace solar geoengineering and forgot to cut emissions, we are screwed. You just can’t say that too strongly. If you just keep emissions going, so carbon concentrations, that’s the amount of carbon in the atmosphere, keeps rising, and then you have to keep like blocking out more of the sun, you’re walking off into a more and more risky future. And that is just insanity.

LEVITT: Yeah. And so really the right way to think about solar geoengineering, in your mind, is it’s some combination of an immediate fix. Something that we can use right away to buy us some time. Or alternatively, as a form of insurance, right? If things get worse in the future to the point where we feel a kind of desperation for immediate cooling, if we’ve done the research now, we’ll be well positioned to act thoughtfully and quickly in the future.

KEITH: I don’t actually love either the buy-time framing or the insurance framing. For me, this is really about simply an additional way to reduce real risks. So the buy-time framing, it seems to me, exactly exacerbates the moral hazard. Buy time implies that we then get to reduce admissions more slowly. Whereas to me, the central idea is that we could buy, if you want to keep that word, less risk. The central way I think about this is that a world with emissions cuts and solar geoengineering could have significantly less risks, especially to vulnerable ecosystems, some of the most vulnerable people, in a world with just emissions cuts. So it’s really about the additional opportunity to reduce harm. And I think that’s a little different from the buy-time framing or the future-insurance-against-horror framing.

LEVITT: Can you talk in particular about the SCoPEx experiment? What was the basic idea of that experiment?

KEITH: Stratospheric Controlled Perturbation Experiment. This is really about improving the processes that are embedded in the big climate models that in turn we use to predict the effects of say solar geoengineering or anything else. So these climate models are, you know, a bunch of computer code, but the computer code in the end is built on observations of the natural world. And so one of the key processes that we don’t know enough about is how aerosols would actually be formed in a plume in the stratosphere, how much they stick together, and some key ways in which they interact with the ozone chemistry we don’t know enough about. And there are things that you can’t find out just by modeling more, because we just don’t have enough observations. And this experiment was going to make a little plume in the stratosphere and allow us to observe that plume on very short timescales to observe the plume evolution. And there aren’t very many high-altitude balloon operators in the world. And there is a company called the Swedish Space Sciences that wanted to do the balloon flight and was well prepared to do it. And it has been extraordinarily politicizing, including for mistakes we made. So there’s been headlines about how Bill Gates is going to block out the sun. The actual experiment, even the strongest critics, admit the experiment, itself, doesn’t pose a risk the only risk is that the balloon gondola falls on somebody, which actually is something we worry about. But the atmospheric risk is really negligible. It would release a kilogram of aerosols, which is less than a minute of flight of a conventional aircraft in the stratosphere.

LEVITT: So obviously, you expected some resistance because you went to extraordinary lengths to build oversight and consensus around this experiment. What did you do on those fronts?

KEITH: Well, the obvious question is what are the regulations about such experiments? And the answer is what we’re doing is legal, or there’s a regulatory process that would authorize it. But we recognize that there were legitimate disagreements about solar geoengineering research. So we set up an kind of independent advisory committee, which, advises not us, but the Harvard senior administration about the reputational risk to Harvard of us running the experiment and about the ways that we should manage the experiment to take public input as much as possible. So we set up this independent advisory committee and then in a complicated process, the thing fell apart in Sweden. So in the end really, there were some prominent voices who had opposition, most importantly, something called the Saami Council, indigenous group, and that Saami Council letter and others helped to make the Swedish government basically order the balloon operator not to fly us.

LEVITT: So let’s keep on going down. What happens below the stratosphere?

KEITH: So next on the journey down, would be an idea that you could reduce or thin some kind of cirrus clouds. Cirrus clouds are the high wispy clouds that you can see on the upper part of the atmosphere. So any cloud by looking white and puffy will reflect some sunlight and that cools the earth. And low clouds really just act in this cooling way. But a high cloud also acts as an infrared heat trapper and for thin, high clouds, these cirrus clouds, that infrared-heat-trapping effect can be stronger than the solar cooling effect. And so the net effect of those clouds is to warm the planet. So if you reduce those clouds, you’re cooling the planet and that’s the idea of cirrus thinning.

LEVITT: So airplane contrails are an example of that kind of cloud. And there was an amazing study that came out a year or two ago that suggested that you could dramatically alter the climate change effects of airplanes simply by redirecting planes and keeping them from making contrails, right?

KEITH: An aircraft contrail is more or less the same as a cirrus cloud. An aircraft will make or not make contrails depending on the humidity of the air that it’s flying through. So in some places you’ll see contrails and some not. And an aircraft could in principle make less contrails by deliberately altering its flight path a little bit to fly in air where contrails get made less. So it turns out that as of today, if you look at the total warming impact of all of global aviation about half of the warming impact, a little less than half, is actually due to this instantaneous warming from the contrails rather than the accumulated CO2 from the aircraft operations.

LEVITT: Let me stop you there because that is so crazy. It’s what happens so often in climate science is that things that seem really unimportant end up being incredibly important. Like I’m sure my listeners have heard about and thought about the burning of fossil fuels by airplanes for decades and understand exactly why that would be bad for the environment. But I bet for many people, they have never heard about this. And to think that half of the total effect is coming from these seemingly harmless contrails, it’s amazing to me.

KEITH: To be clear, there’s big error bars on that half, I should emphasize. And it might be as low as 20 percent. But it’s significant and there are papers that put it close to half. So then there’s this really interesting fact that it might be that aircraft operators could substantially reduce their contrails at the expense of very small, additional fuel burn by just redirecting the aircraft to optimally reduce the contrails while burning just a tiny bit more fuel. And this is now being very actively studied. And the politics are very interesting. So the global aviation industry might more publicly take responsibility for that contrail portion of their warming, and then reduce it by this technical fix of readjusting. And that would allow the aircraft industry to look like heroes in reducing emissions.

You’re listening to People I (Mostly) Admire with Steve Levitt and his conversation with David Keith. After this short break, they’ll return to talk about why the carbon offset market is bogus.

*      *      *

Morgan LEVEY: Hey, Steve. So our listener Simon wrote in, he is a mid-career scientist in a major university. And Simon is worried about spending a lot of time padding his resume, this feeling that he’s pushed to publish all the time to get grants, to get prestige, probably to get tenure. And he feels like that sometimes comes at a cost, that he is pushed to produce a large number of publications, a quantity over quality. Now, this is really a follow-up to our conversation with Dan Gilbert, who gave the advice to work on one project at a time. Really working on things that were most valuable. What do you have to say to Simon?

LEVITT: So you know how much I love Dan Gilbert, but I think Dan’s advice in this particular case to work on one project at a time is terrible advice. Because sometimes you’re in the mood to work on data, and sometimes you’re in the mood to do theory. Sometimes you’re in the mood to talk to people and sometimes you’re in the mood to read. I find that for myself, I have much more fun with my work when I have a bunch of projects. And each of them is demanding something different for me so that I can cater my whims to the needs of the project as they come.

LEVEY: But I thought you were really a fan of Dan saying to do less, better.

LEVITT: Oh yeah. I’m not talking about doing more. So I want to draw the distinction. Instead of doing 15 really mediocre projects, finding the three or four that are really good, that’s a totally different dimension than the idea that you could only do one thing at a time. So over the course of three years, Dan might do six projects. Over the course of three years, I might do six projects. We’re both doing less, better, it’s just that I think you’ll enjoy it and get a lot more done if you do those six projects in parallel, rather than doing them in series one after another. I’ve learned that my productivity and my enjoyment of life are both much higher when I follow Dan’s advice to do less, better, and when I ignore Dan’s advice to work on one project at a time. And those two are not in conflict, they’re two different ideas.

LEVEY: But I don’t think we’ve really answered the heart of Simon’s question. Simon is a mid-career scientist and there is a lot of pressure to publish, publish, publish, and get your work out there. And so, as a mid-career scientist, not someone who is maybe as established as you, Steve Levitt or Dan Gilbert, how do you tell them to balance this working on things that are interesting without having to produce this quantity that is so praised in the academic field?

LEVITT: Right. The real advice I have for Simon is if he’s doing projects that are just padding his C.V., then I think maybe he’s caught in this trap of not spending enough time thinking about new ideas. And I think it’s so common among researchers that they spend too little time on ideas and way too much time on execution, because if you’re doing things right, you’ve got more good ideas than you have time to explore. And in that world, there’s never too many projects, right? If you have so many good ideas, your constraint is time and how much you want to put in, but you never feel like, oh, I’m working on something I shouldn’t be working on.

LEVEY: But the trick I think is knowing the difference and that might not become apparent on a project until you’ve spent a lot of time.

LEVITT: Yeah. And the other key piece is knowing when to stop. Knowing when to quit. People don’t quit enough. And that’s when you have to be ruthless about killing your own ideas because otherwise you get caught in Simon’s trap of spending all your time on bad projects.

LEVEY: Simon, hope that answers your question. If you have a question for us, our email is That’s Steve and I read every email that’s sent and we look forward to reading yours.

*      *      *

I learned a lot in the first half of this conversation, both about the I.P.C.C. and about airplane contrails. In the second half of our chat, I want to hear David’s thoughts on carbon offsets and carbon capture, but also what that open letter against solar geoengineering means for his research and what it’s like to work for decades on a topic that arouses so much anger.

LEVITT: Could you explain the basic principle behind carbon offsets?

KEITH: So the classic version of a carbon offset is roughly the following: If you were building a big office building development and I was somebody who wanted to buy carbon offsets, we could have a transaction, where I pay you some money, and in return, you build a better building with lower carbon emissions than you otherwise would have built. And that generates some carbon offsets. And there’s a third party of that transaction of course, a regulator effectively, who’s recognizing those offsets. That sounds a little bit like a normal economic transaction — it is not. And it is much harder to verify and fraud is much more likely. If you’re selling me an apple and I’m buying an apple, there’s this nice way in which you want the price to be high, and I want the price to be low, and we can, you know, close the deal. Here I am lecturing this econ professor, you should be laughing. But in this offset transaction, it’s a three-way transaction. And the key thing is the offsets are generated by the difference between the apartment building you actually built, which is observable and the apartment building you were going to build, if I hadn’t paid you the money, which is fundamentally unobservable and you and I have every reason to exaggerate it. You were going to build this absolutely crappy apartment with leaky windows until you got my money and you built a nice one. And this is not just theory. Early on in the climate policy world, there was a thing called the clean development mechanism that worked on offset credits. There was emissions from generating a certain kind of chlorofluorocarbon. And Chinese factories were deliberately engineered so that the base factory design emitted a lot of the bad chlorofluorocarbons, so they could get paid money to not do it. So this offset market is corrupt in a deep way, because it’s so deeply hard to verify, because it depends on a counterfactual that’s not observable.

LEVITT: So how big is the market for offsets these days? It’s huge, right?

KEITH: Yeah, it’s huge.

LEVITT: If offset markets were acting the way we expected, does huge mean that the impacts will be so big that we should actually be able to detect in the global temperatures the impact of these markets?

KEITH: I’m so glad you asked that. Not at global temperatures, because those are really a late measure — there all sorts of reasons you wouldn’t see that. But here’s where you should start to see it, is scientists now can make really good measurements of the total flux of carbon into the land biosphere and the oceans, separately. So we really know that number increasingly well. And the global market for these offset transactions, this idea of planting trees — indeed, I have just beside me here a little coffee napkin that I picked up on a Delta flight the other day that said “Carbon neutral since March 2020. Travel confidently knowing that we will offset the carbon emit on your Delta flight.” This is obviously complete and utter nonsense. If it was really that easy for Delta to zero out their carbon emissions, the whole climate problem would be easy, but it isn’t. There’s all sorts of reasons why tree planting doesn’t really end up offsetting the fundamental impact of burning fossil fuels. But I love your question because I think pretty soon we’re going to get to a point where the offset market that’s claimed is big enough that we should be able to see it in the change in the biospheric stocks, which are not happening. And then it will become more obvious that this is — maybe fraud is too strong — but this is just a set of, in many cases, well-intentioned people, some not so much, who have built a system that just is deeply unverifiable, where there are all these reasons to make the numbers bigger than they really are.

LEVITT: There’s really a buzz around carbon removal right now. Billions of dollars of venture capital going into it now. How optimistic are you towards the prospects for the carbon-removal approach more generally?

KEITH: So moderately, but here you need to flag for your listeners, conflict of interest. So I generally think of myself just as an academic, but kind of by accident I ended up starting this little company called Carbon Engineering, really just because I wanted to attack the problem of whether you could build a cost-effective way to capture carbon from the air in an industrial process and that company’s done very well. Publicly traded company has now announced that they look like they’ll build a first plant using our technology the biggest plant by far that would use technology of this kind. So obviously, I have self-interest in that though I find myself academically pretty far away from it. There are ways we could really remove carbon, in the sense that the climate problem is driven by taking carbon from the geosphere from deep underground — coal and gas and oil, burning it, and putting it in the active biosphere. It is possible to reverse that process. So if you take carbon that was in the atmosphere and put it back deep underground, and you could do that using the kind of technology that my company did or lots of its competitors, or you could do it by burning biomass that came from the atmosphere and putting that carbon underground. Or you could do it by other methods of putting what’s called alkalinity in the ocean, where you permanently hold the carbon in the ocean. Those things really are true negative emissions for carbon removal. And they are an important part of the way humanity gets itself out of this problem. But I think the carbon removal market actually is a little over-hyped. There’s a great deal of excitement that somehow in the near term, people are thinking of those as a substitute for cutting emissions. And my view, again, is that the vast majority of effort should go into cutting emissions, not into building up a big infrastructure for removal.

LEVITT: That’s a good point. And one thing I didn’t appreciate for a long time is the massive scale on which carbon removal would need to operate to have a big impact. So take Carbon Engineering — so you’re building a factory, which will take carbon out of the air. If you wanted to undo 10 percent of our current emissions, how many billion-dollar factories would you need to build to be operating on that kind of scale, say to offset, you know, what the transportation industry does?

KEITH: Global emissions of carbon dioxide are about 40 gigatons, 40 billion tons a year. That’s 40 cubic kilometers of liquid carbon dioxide every year. So I’ll go with your 10 percent number. That would be four gigatons. And the plant that our partners have announced they’re going to build with our technology is half-a-mega-ton-a-year. So you need a thousand of those mega-ton-a-year plants to be one gigaton. So you need 4,000 to be four gigatons. Which, it’s not totally ridiculous, but it’s not going to happen overnight. And I think the main point is we should focus on cost-effective ways to cut emissions now.

LEVITT: So let’s stick with that 4,000 plants. Because thinking like an economist, which I know you’re good at now because I’ve been talking to you it takes real resources to build those plants. And in the short run, those plants are actually net negative on greenhouse gasses, and then you earn it back over time. So these are no bargain in the short run, these carbon removal approaches.

KEITH: I do think carbon removal does something that none of the other methods can do. In a sense that cutting emissions just stops us making the problem worse. But even if we remove emissions, we have the carbon that’s in the atmosphere. Solar geoengineering, adaptation do not deal with the long-term risks of the carbon in the atmosphere. The only way to really reduce that risk on a policy relevant timescale of a few hundred years is carbon removal. So I actually am very bullish on carbon removal in the long run as a way for humanity to undo some of the damage that the fossil fuel economy did. It’s simply that if you think about it as a short-term way to reduce acute damages mid-century, then I get a lot more skeptical. A very interesting thing might happen in near term, and I credit this idea to a friend and colleague Jane Flegal, who just went from the White House to go work for Stripe, a company that’s pushing carbon removal technologies. And she said something I thought deeply insightful, which was that all these high-tech real removal methods may help to kill the B.S. offset market in a way that’s very healthy. So just speaking of my company, because it happens to be public recently — was announced that AirBus has purchased 0.4 million tons worth of this pure carbon removal. That’s by far the biggest announced purchase in this world and the price is not publicly disclosed, I’m not letting any secrets to say that it might be north of a hundred dollars a ton. And the point is, it’s just not going to be stable for AirBus to pay more than a hundred bucks a ton and at the same time, have, I don’t know what Delta paid to make my flight supposedly carbon neutral, but it was a lot more like $2 a ton. And I do think that having the real technological carbon removal may help to shake the nonsense out of the offset market.

 LEVITT: In January of 2022, a group of 60 scientists and scholars signed an open letter that was demanding an international non-use agreement on solar engineering. It called for a prohibition on national funding agencies from supporting the research and it wanted a ban on outdoor experiments. And that seems like a very extreme and an unusual stance for scientists to take about other scientists who are trying to do basic research to understand a problem.  

KEITH: They even called for blocking the I.P.C.C. from assessing the topic, which is really an extraordinary call for ignorance. To be clear, the actual call for a non-use agreement, at least in near-term moratorium, is one that I would support. I don’t think we should use these technologies right away. And I think we do need a way to restrain a potential early action that’s not well-governed. But effectively what they called for was a ban on research. Now, obviously, I don’t buy it. I don’t think this is simply that scientists should get to do what they want. I think scientists need oversight. But solar geoengineering, where the fact is major bodies, like the I.P.C.C. and many high-quality published studies, show that it could substantially reduce climate risks. The idea that you’re going to ban that strikes me as a high bar.

LEVITT: So it’s been my experience that in the domain of climate research, many people wear dual hats as scientists and as activists. Do you think this letter is an example of science or activism? It feels more like activism to me dressed up as science.

KEITH: I’d say yes, but I don’t dismiss it for that. I think activism is vitally important, but I do question the ethical basis of their claim that research should be banned. And it’s worth saying that the people who signed that mostly live in north Europe, in societies that are pretty well protected from climate change, and even some of the signatories are in societies that are far enough away from the equator they might actually benefit in the short term from climate change. You have the people who will be most harmed typically are among the world’s poorest people living in the hottest places because an added a little bit of temperature is much worse in places that are already hot. So you need to think very hard about the moral case for banning research. Even if you have this legitimate concern about moral hazard, because you’re potentially taking away the possibility of reducing risks for the most vulnerable and I think it’s a pretty high bar to argue that’s the right thing to do.

LEVITT: My impression was that it’s already been incredibly difficult to get any public funding for this kind of work. If you were to guess, what is the total dollar amount that the U.S. government has spent on research in this area over the last 30 years?

KEITH: Tens of millions at the very most, which is just tiny, when you consider the atmospheric science, including the sort of satellite world, is like 10 billion a year. And that’s tens cumulatively.

LEVITT: It’s nothing.

KEITH: It’s nothing. I mean, rounding error to zero.

LEVITT: One of the things that’s so interesting as there’s been this boom in venture capital money going into carbon removal, that’s all facilitated by the idea that there’s a market for carbon, you’re going to buy and sell carbon. But because solar geoengineering isn’t dealing in that currency, there’s no market. There’s no economic drivers behind the scenes pushing the investment. I think that’s a subtle, but really important point when you think about the economics and the science and the politics of these choices.

KEITH: Yeah, I think it’s utterly different. If I think about a carbon removal machine, like the one that the company I found that is building, people shouldn’t trust everything we say. In the end they should trust what our actual plants do, because when our plants are built, it will be easy to verify what goes in and what goes out. It’s easily measurable. And so it’s a verifiable thing. And our company and our companies should just like compete to do the best job. That’s where I think capitalism can work very well. I think solar geoengineering is deeply different. It’s already so cheap. We don’t need to make it cheaper. The whole issue with solar geoengineering is about trust. This is a dangerous thing. Manipulating the earth’s system, where humans are going to disagree about it, sometimes very strongly. It’s healthy and natural that that’s true. And some coalition of countries may eventually push to do it. And other countries will have doubts about what’s happening. You won’t be able to directly observe the effects. You won’t know for sure whether it’s been working. So this is really all about trust. And however imperfect it is, trust like that is gained by distributed systems like academic research internationally. And open science, not closed science, such as in companies. So my view is there’s almost no room for for-profit activity for solar geoengineering. To be clear, if it’s actually implemented, a for-profit company might build a satellite that a government pays for that goes and measures the aerosols — fine. But I don’t want a for-profit company in charge of the key mechanisms of solar geoengineering, because there’d just be no reason to trust them. And there’s also a way which it just doesn’t make sense as a market. Because if I cooled the whole world, how do I make anybody pay for the service of cooling the world?

LEVITT: Everything I know about you makes me think that you are an environmentalist, that you love the planet and nature, and you want good things to happen to it. Do you ever personally find your own activism tendencies conflict with your scientific outlook?

KEITH: They certainly shaped it. I’ve been lucky enough to spend a lot more time outdoors. And also I was the child of wildlife biologists, so I’ve been influenced by that world view. I was brought up on Aldo Leopold and Rachel Carson and so on, and I actually am motivated by trying to reduce the impacts of the natural world in ways that are certainly not just rational. They’re my values. And there’s no question that influences the research projects I pick. And I do think one of the reasons I’ve chosen to work on solar geoengineering is that humans can probably find ways to adapt and, we will eventually cut emissions, and we could muddle through the climate problem, but the natural world won’t. And if we want to leave more of the natural world less mucked up from human activity for future generations, solar geoengineering might be an important piece of that.

LEVITT: So that makes sense, but I was actually pushing even further. Let me give you an example. So I know a scientist who was working on H.I.V. research very early on, and he was meeting with U.S. Senators and he outright lied to them about how H.I.V. was spread in order to get a big funding bill passed. And much to my surprise, he didn’t relay that to me with any shame. He’s like, “It was something I had to do. I thought it was really important to get funding. So I said what I had to say.” That feels like it’s such a violation of my norms as a social scientist. And I just get the sense that in climate science, there’s tremendous social pressure to emphasize the doomsday, because you’re worried that people won’t do the right thing, unless you do that.

KEITH: I think it’s the role of academics to be as honest and comprehensive about all they know. If you’re working for a company or an advocacy group, I don’t think you should lie. But in a sense, it’s your job to push a specific agenda and to choose the facts that you present to support that agenda. But I feel actually very strongly that is not your job as an academic. I can’t remember the exact quote, but there’s an Einstein statue outside the National Academy building that says something like, “Those of us who get the privilege to seek for knowledge,” meaning, who get paid good salaries to just write papers and study stuff, “have the duty to say all we know.” Now, if you’re a startup company, you shouldn’t lie, but you don’t have to expose all your warts. But if you’re an academic, it is your job in my view to say the things that may go against what you personally would vote for if you’re presenting a set of facts, it’s your job to present the relevant facts as thoroughly as you can.

LEVITT: What really surprises me, listening to you talk is you’ve devoted so many years to studying these issues. So much opposition against it. So much demonization of you and your research agenda. And yet, you seem so easy going and relaxed about it. And I’m wondering, many people find themselves in your situation, believing deeply in something that others believe just as deeply against. Do you have advice to people who are stuck in this kind of situation about how to navigate it?

KEITH: Oh, man, I feel like I’m not even keeled in private and I’m more even keeled on interviews like this. So I think some of it is a front. Coming to this through applied public policy, I really do believe that while I may know more facts on some of this than some other people, that my vote ought not to count more than anybody else. And that there really are value questions here where nobody has the right answer. I deeply believe that scientists should not just say what happens and that there isn’t an objective answer about whether or not we should use solar geoengineering. So while my judgment is that it’s technically true that solar geoengineering could be used in a way that would really quite significantly reduce climate risks, especially to the most vulnerable. And for that reason, me as a citizen would likely vote to deploy it. I really believe in the importance of separating out the kind of technocratic scientists from the public policy process. And I don’t believe that I, on this topic, should somehow have extra weight in the legitimate policy debate. But I think the other feature is that I’ve just learned that if I do show much emotion about it that just digs a deeper hole so I try not to. So, I mean, if you ask me how I really feel when there’re ridiculously distorted headlines it makes me want to kick the wall. And it really hurts and it hurts emotionally. And when there’s people in the environmental community, that I feel myself absolutely to be a part of, who regard this as some kind of oil-company front and anti-environmental, it hurts a lot. But that hurt is not an effective way to move the conversation forward.

Thirteen years ago, when we wrote SuperFreakonomics, I believed geoengineering deserved much more attention and investment than it was receiving, and I continue to believe that today even more strongly. Geoengineering is not a panacea, but it’s one more arrow in the quiver. And maybe we’ll never choose to implement these strategies at scale, but to not take careful steps to better understand the potential costs and benefits, that just seems shortsighted.

*      *      *

People I (Mostly) Admire is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, Freakonomics M.D., and Off Leash. All our shows are produced by Stitcher and Renbud Radio. Morgan Levey is our producer and Jasmin Klinger is our engineer. We had help from Alina Kulman. Our staff also includes Neal Carruth, Gabriel Roth, Greg Rippin, Rebecca Lee Douglas, Zack Lapinski, Julie Kanfer, Eleanor Osborne, Ryan Kelley, Emma Tyrrell, Lyric Bowditch, Jacob Clemente, and Stephen Dubner. Our theme music was composed by Luis Guerra. To listen ad-free, subscribe to Stitcher Premium. We can be reached at, that’s Thanks for listening.

LEVITT: You talk like an economist, but you have zero training as an economist.

KEITH: I have collaborated on a lot of econ papers over the years. And had initially an extremely allergic reaction to economists as many physicists do, but I’ve somewhat tempered it.

Read full Transcript


  • David Keith, professor of applied physics and public policy at Harvard University.



Episode Video