Fitting Behavior and Politics into Climate Models, with Fran Moore

Notable Quotes

• Climate change must be addressed collectively: “What motivates people to make changes in their personal life in ways that reduce their environmental footprint? … I think maybe some economists, in particular, are a little skeptical of this framing of the problem or this emphasis on individual action, recognizing that what we’re looking at here is really a collective problem, that a lot about how and why we use energy and fossil fuels is really outside the control of any one individual, and that this needs to be solved at the collective level.” (9:43)
• Persuading people to change behavior can mobilize collective action: “If you can trigger [a] shift in public opinion, then that mobilizes action at the collective level. And that’s what’s really important in driving down emissions. This resolves this tension a little bit, where you see people trying to say, ‘Well, behavior change is important.’ And some people say, ‘No, that’s just greenwashing. This is not how we are going to go about reducing emissions.’ I think both of those can be true. It can be important, but it’s not necessarily important just because of the actual reduction in emissions—it’s important more because of how it persuades other people.” (12:11)
• Minimizing the effects of climate change is still possible: “It seems like we’re starting to bound these 2100 temperature ranges that are starting to make these higher outcomes look quite unlikely, and ours is not the only paper to do that … We’re on [a] trajectory that’s making something in the realm of two to three degrees look pretty likely. 1.5 degrees is increasingly out of reach—if it wasn’t already out of reach several years ago—but two degrees is certainly still possible if ambition ramps up more aggressively than is currently stated, or if mitigation technologies and the advancements are easier and faster than people anticipate.” (19:51)

Top of the Stack

• “Determinants of emissions pathways in the coupled climate-social system” by Frances C. Moore, Katherine Lacasse, Katharine J. Mach, Yoon Ah Shin, Louis J. Gross, and Brian Beckage
• Why Trust Science? by Naomi Oreskes
• The Knowledge Machine: How Irrationality Created Modern Science by Michael Strevens

The Full Transcript

Daniel Raimi: Hello, and welcome to Resources Radio, a weekly podcast from Resources for the Future. I'm your host, Daniel Raimi. Today, we talk with Dr. Fran Moore, assistant professor in the Department of Environmental Science & Policy at UC Davis. Along with several coauthors, Fran is out with a fascinating new analysis that expands the way we model the future of the climate system. I'll ask Fran and her team about how they built in aspects of human behavior, political decisionmaking, and technological progress into their climate modeling tools, and how those complex systems can interact with each other to speed up or slow down our efforts to limit climate change. Stay with us.

Okay. Fran Moore from UC Davis, welcome back to Resources Radio.

Fran Moore: Thank you so much for having me back on the podcast.

Daniel Raimi: Yeah, it is absolutely our pleasure. And of course, you've been on the show before, but it was more than two years ago, which is kind of amazing that we've been doing the show for that long. But can you remind us how you got interested in energy or environmental issues as a kid or as an adult, or how did you end up working in this field?

Fran Moore: Yeah, I think, mostly for me, it really came kind of through college. My undergrad was in Earth system science—kind of Earth science and geology—and I studied the paleoclimate system. From there, I had this trajectory of … I moved to Washington, DC, and if you have a degree in geology, in paleoclimate in DC, your comparative advantage was definitely working on modern climate change and climate policy. That's where I started getting more into that area, and it's just kind of snowballed from there.

And working on climate change is such a fascinating issue to work on, because it really bridges across the whole of the Earth system as well as pretty much everything about our social, political, and economic world, as well. I find it a kind of unendingly fascinating topic to work on, but also obviously with real societal importance at the same time.

Daniel Raimi: We were just talking about this before we started taping about your new paper that we're going to talk about today on the show. It's such a fascinating example of the ways in which Earth sciences and economics and other social sciences can start to be rolled into a single framework, which is exactly what you and your coauthors start to do with this paper. It's out in Nature; we'll have a link to it in the show notes. And the paper title is “Determinants of Emissions Pathways in the Coupled Climate Social System.”

So, let's start with some high-level background. Can you help us understand how most existing models estimate future greenhouse gas emissions pathways and how this new analysis adds new dimensions to those preexisting modeling efforts?

Fran Moore: Yeah. So the motivation for this work was this observation that if you look out to, say, 2100 (which is now actually not that far away), and you look at what the variance is in possible climate futures that we might be in, in that 2100 world, then the question of what happens to emissions between now and then is really key, right? And that, in turn, is a question of, well, what do nations and what does the world as a whole choose to do? How ambitious are they going to be in efforts to combat climate change?

This is a really key uncertainty that is really kind of driving the climate system. If you're interested in understanding climate change, you really have to understand policy, and this question of where does this ambition come from? As well as obviously also the evolution of, say, the energy system as well, and the energy technology that ultimately you're going to kind of translate that policy into emissions changes.

What current modeling strategies do is they kind of treat that uncertainty as entirely exogenous to the model. And by “exogenous,” I mean external to the modeling environment, so input by the researchers. In the climate modeling space, what that means is that, when you run a big climate model, what you do is you take a pathway—some pathway of possible emissions—and maybe you run your climate model with several different possible pathways of emissions. These are now standardized across climate models, and they're called the SSP, RCP kind of combinations.

It's an acronym that's not super important what it stands for, but essentially these are possible future worlds of socioeconomic development as well as emissions. But then in the energy modeling space, you see a similar thing where the alternative policies enter into these models of the energy system as constraints. What these types of models will often do is they will say, “Show me what the least-cost-energy-mix pathway looks like in order to limit global temperature change to, say, under two degrees, or to limit cumulative CO2 emissions to some constraint.”

What that means—if those emissions are external to the model, then you're not able to really say anything about probabilities over those different emissions scenarios. And that's what we've seen, is that what happens is that these alternatives get presented as purely alternatives without any assessment of the likelihood of these different outcomes. The question that we're trying to answer in this paper is a question of, Where does policy come from? If we understand where this policy comes from, then we can start to bound the space of possible emissions trajectories that are going to shape climate change over the twenty-first century.

Daniel Raimi: Yeah. That's such a nice way of putting it: Where does policy come from? I'm going to start to ask you about how you derive that estimate of where policy comes from.

I imagine our listeners are imagining all of the things that are difficult to quantify about where policy comes from, and they're curious about how you did quantify those things. Can you talk a little bit about how you incorporated … I know we don't have time to talk about all of them, but can you talk about maybe a couple of the social and political dynamics that you incorporated into the models, and what the challenges were with incorporating them?

Fran Moore: Yes. So this is highly ambitious (I would say is maybe the polite way of putting it) in terms of what we're trying to do here. The way we approached this was twofold. One was recognizing that feedback loops are … If you're looking at the long-term dynamics of a system, feedbacks can be really important in determining that—and particularly feedback loops that are connected together can produce highly, very, very, quick changes, or it can lead to sudden and unexpected changes in the system and multiple possible stable states of the system. We initially approached this by looking within relevant literatures for evidence of different types of feedback across the systems that are relevant here.

There's a lot of literature involved here, we've got a lot on psychology and social psychology. We're looking a lot at political science, too, because there's a question of: What do nation states do to combat climate change? That's really important. We're looking at—there's some legal feedback in there, and then also in the energy system, as well. We identify these feedbacks, and then we essentially put them together in what is, actually, even though it's rich in ideas, I think, it is very simple in terms of the actual equations that we're putting in. But because we have these connected feedback loops, it can give rise to quite complex behavior.

Daniel Raimi: That's great. So let's take some of those concepts that you've just described and maybe see if we can put a little bit more meat on them. Can you give us an example or two of how, within the modeling framework that you've developed with your coauthors, how a change in one system could lead to a rapid change or so-called tipping point in a different system in the model?

Fran Moore: Yeah, one of the interesting examples we look at in the paper, and we systematically explore using the model, is this question of the role of individual behavior change. This is kind of debated; there's an awful lot of … psychology literature is one of them, that has looked at what motivates people to take pro-environmental behavior—so, what motivates people to make changes in their personal life in ways that reduce their environmental footprint? But this has come under criticism, and I think maybe some economists, in particular, are a little skeptical of this framing of the problem or this emphasis on individual action, recognizing that what we're looking at here is really a collective problem, that a lot about how and why we use energy and fossil fuels is really outside of the control of any one individual, and that this needs to be solved at the collective level.

What we explore in this model, though, is potential feedbacks. In particular, there are two relevant feedbacks: There's a social-conformity feedback, where people tend to want to be similar to others in their social network. And what that means is that, depending on the strength of that effect, you can get very sudden and rapid changes in things like behavior or opinion as norms change and then spread through a population.

The other really important feedback we look at is what’s called a credibility-enhancing-display feedback (or, that's what we call it in the model). This is the idea that if you undertake costly personal behavior change in ways that are consistent with your underlying values, in ways in which are consistent with how you think, or how you are advocating that society as a collective should change, then that actually changes your credibility as an advocate for those collective changes.

What this then does … So for me, I'm trying to eat less meat in order to reduce my climate footprint. The effect of that on emissions is small. The effect—even if everyone did —would still be relatively small. But if I am able to articulate to others about the values that lead me to do that, and why I think it's important, and how I think we should act collectively to reflect those values, the fact that I'm making these changes can make me a more effective advocate for those and can make me more persuasive. So, that is a feedback from that behavior change back to public opinion. And if you can trigger that shift in public opinion, then that mobilizes action at the collective level.

And that's what's really important in driving down emissions. This kind of resolves this tension, a little bit where you see people trying to say, "Well, behavior change is important." And some people say, "No, that's just greenwashing. This is not how we are going to go about reducing emissions." I think both of those can be true. It can be important, but it's not necessarily important just because of the actual reduction in emissions—it's important more because of how it persuades other people.

Daniel Raimi: That's so interesting. We actually did an episode on the show, maybe a year or a year and a half ago with Shahzeen Attari, who's a professor who studies this, like the way that climate scientists are perceived when they are viewed as, let's say, flying a lot versus riding their bike a lot, or something like that. It's such a fascinating topic.

Is another way to think about this, your idea going viral? So, if you stop eating meat, Fran, and all the students in your class and all the people around you in Davis, they see you doing that, and they're inspired to not only stop their meat eating, but take other actions as well that are pro-environmental—is that a suitable analogy?

Fran Moore: Well, the important thing is it's not just the spread of the behavior change, but it's actually the change in public opinion that leads then to collective emissions reduction. Because even if people were to change a lot of their behavior, the decisions about how we produce electricity, how we design our cities—these are not things that individuals can change by themselves. You need to have some process of aggregating through the political system, but because you have this feedback, those two things are not unconnected to each other.

We show that it's possible to have states of the model where this willingness of climate policy supporters to undertake costly personal behavior change can be really decisive in triggering a cascade of feedbacks that lead to a tipping point. But it's not necessarily common, so I don't want to overemphasize that, but that's just one example of the feedbacks in the model.

Another interesting feedback that listeners might be quite familiar with, because it shows up a lot in economics literature, is this learning-by-doing feedback in the energy sector.

We have this representation where you have this feedback where early energy technologies are very expensive, but if you can push that early installation, people learn how to reduce costs, how to get more efficient at doing the installation, supply chains improve, and that cost comes down over time. So, that leads to more installations, which leads to lower costs, which leads to more installations, and so on. That's another example of a tipping point in a really different system—but that's clearly connected in this overall system.

Daniel Raimi: That's a great example.Solar PV and wind technologies come to mind there, and maybe batteries, too, in the years ahead.

And now, a word from our sponsor.

Hi, it's me—Daniel Raimi, again. That's right: Resources Radio doesn't have any sponsors. We depend on listeners like you to make our show possible. If you're enjoying today's episode, please consider making a donation to RFF. Visit rff.org/support to donate online and find out other ways to contribute. Thanks.

Let's move now from talking about the model itself and the ways that you construct and parameterize the model to some of the main results. As always on the show, we're skipping over lots of great detail that I would encourage readers to go check out in the paper. But once you've started to incorporate all these dynamics into the model, can you talk about the range of outcomes that you find and whether there was a most likely outcome that emerges from the model?

Fran Moore: Obviously, there's a lot of uncertainties here, so we don't run the model just once and then look at the outcome; we run it 100,000 times. What we do is sample over the uncertainties in these parameters that we put into the model. Then we cluster (or group) those 100,000 outcomes into clusters of similar trajectories of policy and emissions. What that's trying to do is generate a tractable number of scenarios that emerge from the model that represents different possibilities of what this social, political, technical system looks like. If we look at what those scenarios look like, we see a range of temperatures in 2100 of somewhere between about two degrees to about three degrees—that’s where an awful lot of those emissions trajectories come out.

Certainly we have a central cluster of trajectories, which we call the “modal path.” About half of our scenarios fall into that bucket. And that sees relatively ambitious scaling-up of climate policy over the next two decades, which in turn leads to peaking of emissions by about sometime in the 2030s, and net zero by about 2080. That is not a Paris Agreement–consistent target, but it is something that is very different from, say, the baseline business-as-usual-type emissions. And it gets us to about 2.3 degrees by 2100.

Daniel Raimi: That's really interesting. I've been talking about this with some other people. It's so interesting to think about how these trajectories have changed over time, because technologies have changed, because policies have changed, because public opinion has changed. I know you didn't study this in the paper, but I'm curious: if you did the same analysis, let's say 15 years ago, do you have a guess as to whether you would end up with a higher or lower trajectory in terms of the end state temperature rise that you observe in 2100?

Fran Moore: Yeah. I'm almost certain that if we were to have done this 10, 15 years ago, our probability mass around those 2100 temperatures would be substantially higher. A change in public opinion is part of it, as well as a scaling-up of carbon pricing, just even in the last few years about the number of countries that are doing it, the rate at which those carbon prices are increasing—and those are what are going into our model to inform this analysis. And I think that has really changed.

I would also say I kind of agree with your observation, that it seems like we're starting to bound these 2100 temperature ranges that are starting to make these higher outcomes look quite unlikely, and ours is not the only paper to do that.

Actually, one thing we're able to do is use a recent paper that came out that looked at these pledges countries have made. So, countries have now stated what they're going to do under the Paris Agreement for emissions in 2030, and a lot of countries have also said things about their 2050 goals. If you look at, and then have to make some estimates about how likely those are or what exactly some of those pledges actually translate into in terms of emissions, if you look at what the 2030 and 2050 emissions look like, given those stated policies, they're very similar to what we find in our modal pathway. That's quite surprising, given that none of that information enters the model at all, but we're still recovering something similar that seems to match these Paris Agreement commitments.

It's not just this paper—there are several other ways at which you could get at this problem that seem to suggest that we're on this kind of trajectory that's making something in the realm of two to three degrees look pretty likely. 1.5 degrees is increasingly out of reach if it wasn't already out of reach several years ago, but two degrees is certainly still possible if that ambition ramps up more aggressively than is currently stated, or if mitigation technologies, if the advancements are easier and faster than maybe people anticipate.

Daniel Raimi: All great comments, and they absolutely dovetail with work that I'm doing right now as part of our annual Global Energy Outlook report, where we look at all sorts of different long-term energy projections, and we're certainly finding something similar to what you've just described there.

Stepping back a moment and going back to this question about modeling: As listeners are probably intuiting, developing models like this is hard. It's complex. And even when it's informed by the best available research, there are still lots of uncertainties, which you've already talked about. But I'm hoping you can talk a little bit more about those uncertainties, and where you think some of the biggest ones are, and what the implications of those larger uncertainties are for the results that we've just been talking about.

Fran Moore: This is definitely a key question here, right? You can come up with the greatest model in the world, but then if you don't have any data to inform it, to inform what the parameters are to constrain it, then it's not necessarily particularly reliable. And we address this in the paper. There is a mix of approaches there. Some of them have fairly well-established estimates in the literature; there's been a lot of work on this learning-by-doing effect and just looking at historical evidence of how quickly costs come down over time with installed capacity. So, for some parameters, we're able to rely on that. In other cases, particularly for these parameters related to the social system, we have parameters related to people's observations of climate change.

There really is no data there. And so what we try to do is we run the model historically, and then we try to match model output to observations of what changes in public opinion have looked like over time and what changes in policy have looked like. We are doing this probabilistic joint constraint on some of the parameters in order to try and get the model to match those historical observations as best we can. And that is definitely limited by the fact that those historical observations are partial—they're only for a subset of countries. It would be really nice to have much more data from many more countries about changes in public opinion, about changes in policy, about changes in behavior, about changes in observation of climate change and recognition of it. But that just doesn't exist. And so, we do it with this kind of partial calibration of the model.

The other way we deal with this is basically trying to recognize that uncertainty and incorporate it into our results. This is why we do these 100,000 runs of the model—to try and fully sample that parameter space of uncertainty. When we do that, we can identify—given the variants in our emissions and our policy trajectory—we can pull out the parameters that seem really important in driving that variation. These are parameters that are maybe not particularly surprising. They tend to be parameters related to these feedbacks, particularly either in the social system or the political system.

This political response is really important, and if you have a lot of … we call it “status quo bias,” which is an idea from political science, that institutions tend to have certain inertia associated with them; they might respond only slowly to changes in public opinion that could wrap into it a lot of ideas around political economy and things like that. So, that tends to be quite important. There tends to be the question of, How quickly do opinions diffuse through the population? It’s quite important, as well as parameters related to improvements in the energy sector. And just how effective is this mitigation technology? How quickly will it improve over time? That's going to be really important in determining emissions.

Finally, there's another set of important parameters that comes through what we call the “cognition component,” which is related to people's direct perception of climate change through their experience of weather. This is an idea that's definitely out there, and there's a lot of papers asking this question of, Do people's beliefs about climate change or opinions on climate policy relate to the weather they're experiencing? It seems like papers tend to find some association there, but there's also a lot of evidence that people's perception of climate change is imperfect because of certain cognitive biases and because of limited memory. If you allow for the question of just how large those cognitive biases are, it is another important set of uncertain parameters.

Daniel Raimi: Is that kind of similar to the status quo bias in which we come to perceive that whatever we're experiencing currently is normal, as opposed to maybe abnormal?

Fran Moore: We have two distinct biases that we look at in the model. One is—it's actually something that I've done work on before using Twitter data—this question of shifting baselines. This is kind of like the status quo bias, where the idea is, well, maybe you perceive anomalies, but you perceive anomalies relative to some shifting idea of “normal.” And that idea of normal changes over time. We found evidence using the Twitter data (and this is incorporated into this model) that that happens on about a five-year time frame. And that's relatively fast compared to the speed of climate change. It suggests that people are not going to perceive the true degree of climate change.

The other really interesting cognitive bias we have in there is what we call a “bias assimilation effect.” This is the idea that your political opinions affect how you perceive the weather. So, you're kind of primed, because—say you support climate policy, and that then primed you to notice warm anomalies more than cold anomalies. And the opposite: so, if you oppose climate policy, maybe you notice those cold anomalies more than those warm anomalies. The combination of those two cognitive limitations—those two cognitive biases—is actually very powerful. Even if you think that perception of climate change affects opinion on climate change (which, even that itself is not necessarily well-established—but even if you think that's true), the combination of those two cognitive biases can really diminish that effect and lead to sustained opposition to climate policy.

Daniel Raimi: Yeah, absolutely. I'm reminded of the time that Jim Inhofe, the Oklahoma senator, brought a snowball into the Capitol on a day it was snowy outside to talk about climate change.

Well, Fran Moore, again from UC Davis, this has been so interesting. I really encourage people to go check out the paper, which again, we have a link to in the show notes. Let's close out our discussion with the same question that we ask all of our guests, which is to recommend something that you've watched or read or heard. It can be related to the environment, or even just loosely related to the environment, that you'd recommend to our listeners. So, what's at the top of your literal or metaphorical reading stack?

Fran Moore: I went on a mini-binge on reading some philosophy of science–type books. This began maybe during the pandemic, and watching this interaction of science and policy and society in ways that are very—I think, for people that work on climate change, this was not unfamiliar—what you saw happen with epidemiology as we went through the COVID pandemic. I got interested in this set of books—maybe you could think of them as this post-postmodern approach to science, where we recognize that the way science is done, the way knowledge is created: there is a social setting to the way in which that happens that is embedded in a kind of particular political, social, economic environment.

And yet, we still think there's something real there, right? There's something valuable to science. What is it that we can point to about the scientific method, the scientific process, such that, even though it's done by people, it's done in institutions, so in that sense, it's kind of contingent on a particular social setting, and yet, it's still probably saying something valuable.

There are two books that I would definitely recommend for anyone interested in those questions. One is this book, Why Trust Science? by Naomi Oreskes. She writes fascinating books. Many of your listeners might know her from her work on Merchants of Doubt, showing this climate-denial industry. This is a book based on a series of lectures she gave, really teasing out what is it about science that leads us to trust it? And when should we not? What are some of these blind spots or weaknesses where the scientific process is understating uncertainty or overstating confidence? I think, for scientists, it's really valuable to read that. And Naomi Oreskes is really interesting, because she has a science PhD as well as a history PhD, so it's written by someone that is very familiar with the scientific process.

The second one is this book, The Knowledge Machine, by Michael Strevens, which is a longer-term history of science, but is also pulling out these questions of, What is it about science and the scientific process that leads us to have confidence in it under certain conditions?

And so both of those I would definitely recommend. They’re highly readable. I think it's also—I know many of your listeners are economists, and thinking about economics is becoming more scientific in many ways—in the way questions are asked, the importance of empirical research, the way funding is applied … economics is in this kind of transition phase at the moment between … It's starting to look a lot more like the natural sciences, I think—from my observation as someone kind of sitting at the intersection of the two. So, I think a lot of these would be recognizable to people working in economics today, too.

Daniel Raimi: That's really interesting. Great recommendations. We'll have links to those in the show notes for people to check out. One more time, Fran Moore, thank you so much for joining us today on Resources Radio.

Fran Moore: Thank you for having me.

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

Resources Radio is a podcast from Resources for the Future (RFF). RFF is an independent nonprofit research institution in Washington, DC. Our mission is to improve environmental, energy and natural resource decisions through impartial economic research and policy engagement.

The views expressed on this podcast are solely those of the podcast guests and may differ from those of RFF experts, its officers or its directors. RFF does not take positions on specific legislative proposals. Resources Radio is produced by Elizabeth Wason, with music by me, Daniel Raimi. Join us next week for another episode.