In this episode, host Kristin Hayes talks with Ariel Ortiz-Bobea, an associate professor of applied economics and policy at Cornell University, about the impacts of climate change on agriculture. While climate change often is perceived as a future issue, Ortiz-Bobea elaborates on a recent journal article he coauthored that assesses the extent to which climate change already has affected agricultural productivity. This research finds evidence of substantial productivity losses over the past 60 years, and Ortiz-Bobea predicts that hotter regions could suffer more pronounced losses in the future if temperatures continue to rise.
Listen to the Podcast
- Climate change is a challenge that impacts more than just the future: “We tend to think about [climate change] as a future problem, or maybe something in the present that is becoming more imminent. Generally, we don’t see it as something that has already happened—a thing of the past. It is critically important to answer questions [about the past], so that we know and we have estimates of how much climate change has already impacted us, and so that we build not only [an] understanding of what the impacts are, but also the support to tackle climate change.” (5:06)
- Historical impacts of climate change on agriculture: “We linked country-level agricultural [productivity] data with detailed weather data around the world … What we did by linking these data sets is try to characterize how weather fluctuations affect agriculture at a global scale. That’s the first point that we’re trying to answer: how agriculture evolves when weather conditions change. The second one was that we needed to know how agriculture would have responded since 1960, if farmers were facing a different trajectory of weather conditions … without human [interventions] in the climate system.” (9:57)
- Temperature increases are likely to hurt crops in hotter regions: “In terms of what drives the variability across regions, I would say it’s more about how sensitive agriculture is to extreme temperatures … So, areas that are warmer tend to be hit the hardest. That’s a big part of why you might see Africa and parts of Latin America and Asia being particularly hit hard in our results.” (20:43)
Top of the Stack
- “Anthropogenic climate change has slowed global agricultural productivity growth” by Ariel Ortiz-Bobea, Toby R. Ault, Carlos M. Carrillo, Robert G. Chambers, and David B. Lobell
- Creating Abundance by Alan L. Olmstead and Paul W. Rhode
The Full Transcript
Kristin Hayes: Hello and welcome to Resources Radio, a weekly podcast from Resources for the Future. I'm your host Kristin Hayes. My guest today is Dr. Ariel Ortiz-Bobea, associate professor of applied economics and policy at Cornell University. Ariel is also a faculty fellow at Cornell's Atkinson Center for Sustainable Future. Much of his research focuses on the links between climate change and agricultural productivity, which is the topic of our conversation today.
In particular, we'll be hearing from Ariel about a paper that he, and a number of co-authors recently released in Nature Climate Change, on the historical impact of anthropogenic climate change on global agricultural productivity. The keyword here is “historical.” There's a large body of research focusing on future impacts, but this study looks back to see how much climate change has already affected agriculture globally. Spoiler alert, the impacts today have been fairly large, and we're fortunate to have Ariel join us to provide more insight. We're actually recording this episode before the final publication date, but we're excited to bring you this preview. Stay with us.
Ariel, it's very nice to connect with you again and thanks for coming on the show.
Ariel Ortiz-Bobea: Thank you for having me, Kristin. I'm a frequent listener of the show and I'm delighted to be here. I also think fondly about my stint at RFF, so I've always been a big RFF fan.
Kristin Hayes: Awesome. Well, it's great to have you back. So let's start with our regular introductions. Can you tell our listeners just a little bit more about your background, and in particular, how you came to focus your research on the intersection of climate and agriculture?
Ariel Ortiz-Bobea: So it's a long story, but I'll try to make it very short, and I'll start with the latter parts about my interest in ag. It really started just being the son of two agronomists, spending weekends going into the family farm in the Dominican Republic where I'm from, so it really started there. I studied agronomy in France. I started with my studies, and over time, I became increasingly interested in social issues, food security development, and that's really what I thought I'd be only working on.
The interest in climate came later when I worked in government. I worked at the Ministry of the Environment and Natural Resources of the Dominican Republic. I was a special assistant to the minister there and saw many different issues related to the environment in a developing country, and that's where my very strong interest in climate and agriculture came in. So I did my PhD at Maryland and that's where I focused on climate change and the impacts of agriculture. I’m still working on those issues, although I'm starting to diversify the different sectors that I'm working in. That's the story.
Kristin Hayes: Fantastic. What was your favorite part of life on the farm? I have to ask.
Ariel Ortiz-Bobea: My favorite part was going fishing on my own. Fishing was a fun part, growing my own tomatoes too.
Kristin Hayes: Very good.
Ariel Ortiz-Bobea: So yeah. Growing tomatoes and fishing.
Kristin Hayes: Fantastic. Sounds very relaxing. Well, let's turn to the paper, and I want to start by honing in on what's new about this study. As you mentioned to me as we were planning this episode, there are quite a few studies that look at climate impacts on agriculture, but most are forward-looking rather than retrospective. So I want to start by asking, why did you and the team that worked with you want to undertake a backward-looking study, and why does that matter?
Ariel Ortiz-Bobea: That's a great question. The beginning of this project, the idea, really started with a bad interview that I had with a journalist a couple of years ago. A journalist was asking me questions related to a study that I published in 2018. The journalist was very frustrated that I couldn't give him answers to his questions, and the questions were good and they really planted the seed of this paper in my head. The questions were about the historical impacts. Where are places in the world where climate change is already affecting agriculture in a way that makes agriculture not viable in certain parts of the world? I just simply didn't have the answers. Could have answered qualitatively but not something quantitative, and so that really got me started. So I guess the moral of the story is that even in bad interactions, there's always a good side to them. So that's a silver lining of conversations like that.
When you look at opinion surveys about climate change in the general population, and I would even dare to say among people working on climate issues, we tend to think about it as a thing of the future—a future problem, or maybe something in the present, something that is becoming more imminent. But generally we don't see it as something that has already happened, a thing of the past. It is critically important to answer questions so that we know and we have estimates of how much climate change has already impacted us so that we build critical support also, not only the understanding of what the impacts are, but also the support to tackle climate change.
Kristin Hayes: We'll get to topline results in just a second, but I wanted to ask you one more contextual question about another innovation in the paper. This gets into one of the technical details here. You specifically note that you and your colleagues measure something called total factor productivity, or TFP, not a phrase that I'm particularly familiar with at all. So I wanted to dive a little bit into that, and you look at TFP as opposed to looking at narrower measures, such as crop yield or agricultural output, which my sense is that those are more standard or widely used measures. So tell us a little bit more about TFP and why you see that as an improved metric.
Ariel Ortiz-Bobea: Great. The first one is that we actually don't measure it ourselves, so that's one clarification. We actually use official statistics from the US government on agriculture TFB, that's the first one. It takes a lot of effort to put these data sets together, and that's why, generally, you see agencies putting these data sets together and not just individual researchers or smaller groups putting them together, but that's a important variable in the study, and so I'm happy to talk about this.
When people think about agriculture, generally, they tend to think about cereal crops or fuel crops, and people think crop yields, not only in the general population but also researchers. A lot of the research has been done on fuel crops, mostly cereal crops. The issue with that focus on just these crops is that globally, the value of that production represents about 20 percent of the global value in agriculture. So that's a relatively small share of the total packet.
In this study, we're trying to go broader and encompass all of the agricultural activities at the global scale, so that's where an aggregate measure is important. So that's the aggregate part; that's the first reason why TFP can be useful is because it boils everything down into one.
So what is TFP, total factor productivity? As the name indicates, it's productivity for all factors of production, instead of thinking about bushels per acre, which is a partial measure of productivity, because you're only measuring output relative to an input—in that case, land. Here, imagine that you could do that to all the outputs and all the inputs together in one metric. It condenses a lot of information, so you have to aggregate all the output. So we're talking about chickens, livestock, everything into a single output, and then all the inputs, land, fertilizer chemicals, put them in another aggregate input. Then all the output growth that you cannot explain by the growth in inputs has to come from somewhere, and we call that TFP. So that's the TFP metric.
There's a loophole in the data. The thing is, official statistics don't account for weather as an input. So when you look at the TFP data, it jumps around on years that we know are, say droughts or floods, and so you see all these changes in agriculture TFP, and it's not that people become dumber from one year to the next. It's that there are omitted factors in the data. And what we do is harness that volatility in the short term from year to year in that data to capture the response of agriculture to weather conditions. So we are capturing how farmers also respond to environmental shocks in this way implicitly. So, that's a great advantage of this metric.
Kristin Hayes: Alright, I've asked you enough preliminary questions and I feel like I should stop the suspense. Why don't you tell us a little bit more about how you conducted the study and then what the topline findings are?
Ariel Ortiz-Bobea: So what we did is that we linked country-level agricultural TFP data—that we got from the US Department of Agriculture’s Economic Research Service—with detailed weather data around the world. So that's a data set that goes back to 1960. What we did by linking these data sets is trying to characterize how weather change or fluctuations affect total factor productivity and agriculture at a global scale. So that's the first point, that we're trying to answer the question, how agriculture evolves when weather conditions change. That's the first step in the study. The second one was that we needed to know how agriculture would have responded since 1960, if farmers were facing a different trajectory of weather conditions consistent with the climate without anthropogenic forcing. What I mean by that is that we are getting from climate models is output that tells us how weather conditions would have been without human forcing in the climate system.
Once we link those two things, how agriculture responds to weather conditions and how weather conditions would have been in a different world, then we can figure out how much anthropogenic climate change is contributing to changes in agricultural TFP. So that's the essence of what the study is. What we find in terms of the bottom line, is that when we look at agriculture since 1961, we find that anthropogenic climate change has reduced agricultural TFP by about 20 percent since 1961. So there's a 90 percent confidence interval between minus 10 and minus 35. So there's some uncertainty around that result.
If you think about that in levels, instead of just percentage terms, what that means is that the level of TFP that we globally are projected to reach this year, 2020, we would have reached that level in 2013, in a counterfactual world without anthropogenic climate change. So in a way, anthropogenic climate change has wiped out seven years of TFP growth since 1961. With some uncertainty, between 4 and 13 years of TFP growth, years of TFP growth lost to that. We also get into some regional impacts and country-level impacts, but those are more uncertain. Results are more crisp at the global scale, once you start going down into the weeds in terms of regional effects, there's more uncertainty there. But yeah, the results are fairly substantial.
Kristin Hayes: So I want to dive a little bit deeper into this question of the counterfactual too, because that's always something that I find particularly challenging to wrap my head around. In this case, as you mentioned, you used climate models to tell you what the world would have been like if human contributions to climate change had not occurred. Can you say just a little bit more about the uncertainty that you embedded into that baseline assumption and just how you establish that important piece of the study?
Ariel Ortiz-Bobea: Yes, and that's why we have climate modelers in the team, so that important part of the team is not only economists. We have our agriculture ecologists, and also climate scientists in the team to get all these pieces together. One innovation, although it's maybe been done in another study as well, is the use of climate models—general circulation models—to figure out what the weather conditions would have been in a world without anthropogenic climate change. If we want to know the cumulative impact of anthropogenic climate change, we need to know how the world would have looked with anthropogenic climate change, which is the world we're living in, and a world without it. The problem with that is that we don't see that world. The current world is the only one we have, and so we need to rely on climate models to play God and go back in time and only remove the human influence on the climate system and see how it would have evolved.
So that's something that is used a lot in climate change literature. Folks are trying to figure out to what extent, say, a heat wave is consistent with, say, internal variability of the climate system, or is it due to anthropogenic warming? So that's what we do in terms of getting these counterfactuals out from the climate models.
Kristin Hayes: Thank you very much for that explanation, and certainly models are something that we at RFF are quite familiar with, and there's a fair amount of comfort with models like that, both in the economic community and obviously in the climate science community. I did want to ask—these models are complicated, they're mysterious. I guess I just wanted to ask you, what would you say to someone who was skeptical about relying on something as similar to a black box and as complicated as a model to underpin an analysis like this? I know you have to. You articulated very well why it's really a necessary component of the study, but how did you build your own faith in the models? Maybe that's my ultimate question for you.
Ariel Ortiz-Bobea: Great question. The first point is that climate models are not the only type of evidence that climate scientists use to determine that humans are the cause of climate change. That's the first one. There's empirical evidence that that's the case. So when you look at the way that the atmosphere is warming at different layers, it is consistent with warmer lower levels than at higher levels. There's a change in the signature at different depths of the atmosphere that are consistent with something happening in the lower levels at the surface of the world. So that's the first point, so if the forces were, say, solar, then you wouldn't see this type of signature and how different parts of the atmosphere are warming. That's the first one. This is only one way that scientists establish this effect of humans in the climate system. That's the first one.
The second one is that these models have been fairly good. If you go back in time and use earlier models, they're fairly good at predicting the growth in global temperatures that you would have seen in decades to come. So you could go back to earlier versions of this. Now we're up to CMIP6, which is the latest intercomparison of climate models that are going to fit into the six assessment report. If you go back to CMIP3, those are much older models, and even before that, they do fairly well, because they were forward-looking, but now we're in the future and you can compare the global temperatures to what those models predicted, given the emissions, the paths that we followed and they're they're spot on. I know that a climate scientist will come up with a whole encyclopedia of things to say, but those are two of the things that I find that are very important that have built credibility to my eyes in how reliable these models are.
Kristin Hayes: That's great. Thank you for talking me through that. I want to turn to the regional variation, something that you mentioned when you were talking through the topline findings. I wanted to ask if you could say a little bit more about that piece; it seemed like that showed up fairly clearly in your results. My understanding is that there is some uncertainty there, but is there anything that you can tell us about which parts of the world seem to have been affected more or less? In particular, I also want to ask if the study sheds any light on the causality behind those regional differences. So can you speak to why some places would have been more affected over the past decades than others?
Ariel Ortiz-Bobea: Great question. The first one about uncertainty: so there's a lot of uncertainty in studies like this. There's uncertainty related to the econometric model that we first estimate, so basically the linkage between changes in weather and changes in TFP, that's an uncertain relationship; there's uncertainty there. There's also uncertainty in how much warming or changes in weather or trajectories or patterns are caused by humans. That uncertainty is captured by looking at different climate models. Different climate models will tell you that the warming might be more pronounced or less pronounced, in different parts of the world and started earlier or later. So there's a lot of uncertainty there, a lot. And so we factored in those two sources of uncertainty in the results that we show in the paper. There's a third source of uncertainty related to the econometric model itself because the plots that we show in the paper are based on a baseline model that we happen to pick first. So it's the main model and we show a lot of the results there.
Another researcher might have picked a slightly different model, with slightly different weather variables and things like that, small variations. In the paper, we also explore that source of uncertainty. So that's more specification uncertainty, and how do we characterize the weather and the agriculture-TFP relationship. So even when we consider all of that, the results are—at the global scale—fairly robust and negative. So we find that we didn't cherry pick the model that we showed. It just falls well in the range of all these, about more than 200 models that we explore.
That uncertainty, when you get deeper into the different parts of the world, the uncertainty is larger, so we get real uncertainty. There's more variation depending on the model that you pick. If you allow, say, weather conditions to affect agriculture in different ways depending on which part of the world, you get slightly different results at the regional scale. So that builds up the uncertainty at a fine scale, although you don't see it as much at the global scale, because you wash out a lot through aggregation.
In terms of what drives different response functions, the variability across regions, I would say it's more about how sensitive agriculture is to extreme temperatures. That seems to be the main driver of the results so far. So areas also that are warmer, are already a higher level of temperature, tend to be hit the hardest. So that's a big part of why you might see, say, Africa and parts of Latin America and Asia being particularly hit hard in our results.
Kristin Hayes: Interesting. That feels somewhat counterintuitive to me, to be honest, where you would think that places that were already adjusting for high temperatures might have some protective benefit from that, but it sounds like that's just the opposite. In fact, when temperatures are already high, those increments actually make a bigger difference. Am I interpreting that correctly?
Ariel Ortiz-Bobea: Yeah, that's what we would think, right? You might think that if you're more exposed to higher temperatures in your baseline climate, then you should be more adapted, but that's an assumption. It's an assumption that we think that people are adapted to their local climate. One thing that we find in our study is that over time, the response function to temperature is becoming steeper. What that means is that higher temperatures are found to be increasingly detrimental to TFP growth at a global scale. That's not something that you would think of as something that you would see in an agricultural sector that is adapting to a changing climate. It's actually appearing to be doing the opposite. So we don't know exactly what that is, what we find is agriculture becoming more sensitive, but we do find that it's consistent with some of our previous research, finding similar signatures in US agriculture.
In US agriculture, we find, particularly in the Midwest, that agriculture is becoming increasingly sensitive to high temperatures, and by a large factor, actually. In that study, we found that it's really two compounding forces. One is a change in the composition of agricultural output in the region. So we found that agriculture is becoming more specialized in crop production in the Midwest. So that specialization into something that is inherently more sensitive to weather fluctuations is making the region more sensitive overall. That's the first point.
Then when we look specifically at the crop output, we find that crop production is also becoming increasingly sensitive in the Midwest. So it's like a compounding factor of increasing specialization and a technological part that is making agriculture more sensitive to high temperatures. So it's not unheard of that we find that in the study. There's also precedent here in US agriculture, but we don't know exactly what's driving that at the global scale. So more research is needed to get at the bottom of that.
Kristin Hayes: You've jumped ahead to one of my questions. I definitely do want to talk to you about where this makes you want to look next, but one more question before that. So like all good researchers, you and your co-authors make sure to include caveats and important context for your findings, some of which we've talked about today. But some of those caveats relate to the benefits of fossil fuels and rising CO2 concentrations in agriculture. And I think this is something that occasionally comes up in the narrative, which is, there are actually some benefits from having a greater concentration of CO₂ in the atmosphere on agricultural output. So can you just say just a little bit more about those and how that relates to the findings of the study that we're talking about now?
Ariel Ortiz-Bobea: Yeah, that's an important point to think about, exactly the nature of our study and how to think about the counterfactual. The way to think of our results: it's not really comparing a world with fossil fuels to a world without fossil fuels. I don't think that that's even a very useful counterfactual. It would be a completely different world. So many things would be different. Maybe we wouldn't be talking here in a counterfactual world, so who knows?
The way to think about our results: it's a world without anthropogenic climate change. So in a way we keep everything else in the world, meaning our living standards going up, our use of fossil fuels, everything is the same, the research and development to make agriculture more productive that is coming from those high living standards: it's also in that counterfactual world. The only thing that changes is the ability of humans to affect the climate system. It's like we had a magic wand, and we just change the rules of physics right when the emissions are coming out and just doesn't affect the climate. So really, it's a narrower counterfactual, but I think it's a useful one to think about how anthropogenic climate change is starting to become a headwind to what we're doing. So that's how we're thinking about it because we have CO₂ in the atmosphere that we don't remove, so that stays in there.
Kristin Hayes: Okay. Well, I love that you said before that more research is needed because that's definitely something that we joke about at RFF at times, that the answer to all of our problems is more research. But there really genuinely is always more research to be done, and every study reveals new questions. So I'm curious, yeah, what issues did this study make you want to take up next?
Ariel Ortiz-Bobea: Plenty, to be honest. So more questions and more ideas to keep going. The first one is about the rising sensitivity of global agriculture to higher temperatures. As I said, we found a similar signature in US agriculture, but we don't know why that is happening at the global scale. That's definitely something that is bugging me, and I want to understand more about, why did we find that? So that's one thing. The other one is that, and this relates more to the productivity literature in agriculture, and it's that it's been documented that there's an increasing slowdown of agriculture productivity at a global scale, so that there's certain regions where you see agriculture productivity growth start to slow down. It's unclear exactly why that is the case. Is it the changing nature of R&D between private and public? Are the returns going down?
I think that there's a lot going on there, and what really interests me is what's going to happen to the future. If we are projecting climate change into the future, how much R&D do we need to counterbalance future climate change impacts? That's something that I'm curious to know more about and dive deeper into, so that's definitely something that it's on our agenda. The other part is that, as I said, there's uncertainty here at the regional scale, so some of our efforts that we're already starting involve looking at regional studies. We have a study working now on trying to do an analogous study on US agriculture. We have one on Chinese agriculture, and we might soon start a project on European agriculture, so that some of the uncertainty that we have in this global study, we can hopefully address them in more regional work.
Kristin Hayes: Alright. Well, that sounds like a full plate. I'm sure you'll have plenty to keep you busy over the next few years. That's for sure. But yeah, but these are such important topics and they have such global import for all of us who eat, which is all of us. I'm grateful to have folks like you looking into them with such rigor and enthusiasm.
So Ariel, thank you again for taking the time to talk through this paper with us. Again, it's published in Nature Climate Change, and folks are welcome to check it out to the extent that they are able to. Let me just close the podcast with our regular feature, Top of the Stack. What would you want to recommend to our listeners? Good content either on this topic, or any topic really, that you might want to give a shout out to. What's on the top of your stack?
Ariel Ortiz-Bobea: Yeah. Sure. Well, literally on top of my stack, there's a bunch of overdue reports, but that's not what I'm going to talk about.
Kristin Hayes: Are you sure you don't want some of our listeners to maybe take those over for you? I'm sure they wouldn't mind.
Ariel Ortiz-Bobea: Yeah,I should probably start outsourcing them soon. But one book that I'm reading that relates to our previous discussion, actually, is a book by Alan Olmstead and Paul Rhode. Those are two economic historians at UC Davis in Michigan. And the book is called Creating Abundance: Biological Innovation and American Agricultural Development. The book chronicles the story of the role of innovation in the development of US agriculture and how innovation, even in a rudimentary way back in the 19th century and earlier, really conquered new regions. So you had wheat expanding into regions that were previously out of reach, and by trial and error, wonderful things can happen.
Ariel Ortiz-Bobea: So I find this book really encouraging. I find humans are brilliant creatures and capable of amazing things when we put our minds to it and looking back at agricultural development over several centuries here in the United States and the technologies that we have today available, I think that there's so much that we can do. That's why incentives are so important, to channel that energy to meet these emerging challenges like climate change. And that's where I think also that RFF plays a great role here thinking about how those incentives are put together to meet these challenges.
Kristin Hayes: Well, that is both a very optimistic and very kind note to end on. So thank you so much. It's been a pleasure talking with you.
Ariel Ortiz-Bobea: Great talking to you. Thanks.
Kristin Hayes: 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 a 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 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 Daniel Raimi. Join us next week for another episode.