Long Story Short: Forests, Fortran, and Climate Adaptation in Your Portfolio

Why is Fortran, one of the oldest and least-used programming languages, so crucial to modelling the impact of climate change? How do forest growth, cloud formations and soil quality feed impact portfolio risk assessment? And what do ensemble climate models and professional football teams have in common?

Matt Goldklang, Climate Scientist at Man Numeric and recent co-author of the award-winning Climate Investment: Positioning Portfolios for a Warmer World, joins Long Story Short to discuss climate models, adaptation versus mitigation game theory, and how your portfolios can adapt to a more chaotic climate future.

Recording date: March 2022

Episode Transcript

Note: This transcription was generated using a combination of speech recognition software and human transcribers and may contain errors. As a part of this process, this transcript has also been edited for clarity.

Shanta Puchtler:

Welcome to Long Story Short. My name is Shanta Puchtler, I'm president, Man Group. And I'll be hosting a podcast series for investment professionals. Every year, we're asked by our clients, investors globally all kinds of questions from data science, to monetary policy, the effect of inflation, how to trade or better risk manage portfolios. It's the goal of this podcast to have wide ranging and deep conversations on some of the most difficult topics that we face today.

Welcome back to Long Story Short, I'm your host, Shanta Puchtler, president at Man Group. Today we're going to look at a very pressing and timely topic, which is that of climate change. For investors, climate is now both a short-term and long-term risk as we see the impact of the already 1.2 degree rise in global temperatures. Portfolios need to be adapted, but what does that look like in practice? To help answer this question, I'm joined by Matt Goldklang, climate scientist at Man Numeric. Welcome Matt.

Matt Goldklang:

Thank you. Super excited to be here.

Shanta Puchtler:

So it's a real treat to talk to an honest to goodness climate scientist. There's so much we could talk about here. The geek in me would like to dive into climate modelling itself first. Tell us a little bit about the state of the art of climate modelling. What's changed in the last five years. Is it all about computational power or have there been more fundamental breakthroughs in methodology?

Matt Goldklang:

Yeah. Climate modelling is a really complex, slow burning process. So the changes happen incrementally over time. I would say there are a few different areas in which there's a lot of innovation specific to climate modelling. First is just the representation of different physical and biological systems within the climate models themselves. You can think about clouds and how clouds form. For example, that's a really difficult thing for scientists to wrap their head around. And so the representation of clouds has been a huge topic, at least within the innovation space, within climate models.

But on a whole, it's really the spatial resolution of climate models and the computing power that we have behind them that's really the driving force of a lot of the innovation. You can think of them as a picture, so to speak, and the resolution of that picture really matters.

Shanta Puchtler:

So you said something interesting there that caught my attention, you said biological systems, but you gave a physical systems example. What are examples of biological systems that impact the climate?

Matt Goldklang:

Yeah. Forests, for example, there's a lot of talk lately about the Amazon rainforest being the main lung of the earth system that's breathing in the carbon dioxide and breathing out oxygen. And so the representation of those forests is really important in determining the global climate, because if these climate models are trying to understand the impact of carbon dioxide on the climate, these forests store a lot of carbon, so they are really important in dictating the directionality of our climate.

Shanta Puchtler:

I've also read that soil is an important source and sink of carbon. Can you talk a little bit about that?

Matt Goldklang:

Yeah. Soils also store a ton, well, more than just a single metric ton, but figuratively, a ton of carbon. And soils too like forests are really important to get right in terms of our carbon and climate models. So essentially, the soils will store all of the debris from the forests that gets buried through time, and the amount that specific soils within a specific area can store carbon really depends on a lot of different factors, so getting that right is also really important

Shanta Puchtler:

On a totally different note, something you wrote caught my attention that much of the climate modelling is done in the language Fortran. I learned Fortran back in the 1970s, and I just can't imagine that it's the tool of choice today. Can you illuminate that?

Matt Goldklang:

Yeah. Well, thank goodness climate models are not run with punch cards, but I think Fortran is a choice for a few different reasons, one is longevity. A lot of the climate models originally were developed in Fortran. So for sake of being able to trace back the changes in climate models and trying to understand how climate models have evolved, there's a certain dependence and allegiance to Fortran. Another reason is that I think the name Fortran comes from formula translation, and that just means that Fortran is exceptionally good at doing speedy map math. And climate models in a sense are just a bunch of speedy math. So I think Fortran lends itself to a lot of processing of different climate information.

Shanta Puchtler:

Interesting. I may have to brush off my Fortran manuals. I don't know where I would even find them at this point in my life. The other thing that I've read about is that there is no one climate model, there are actually many climate models. And that ensembling technique is something we do in quant land and finance, but I didn't realize that was being applied in the climate space. Typically, ensemble modeling is used when you have a huge amount of variance and low accuracy. I the implication that each individual climate model is actually fairly inaccurate? And what does that say about the state of climate modeling?

Matt Goldklang:

That's a great question. So I'll just start by saying each climate model that is used, and there are over 50 of them at this point that different global climate modeling centers deploy for global climate model studies. And I would say some are really good at capturing specific processes within the earth system and some are not as good at capturing other processes within the earth system. For example, we might have one model that is really great at representing soil, just as we spoke about, but we don't want to rely on that model to tell us much about ice sheets and perhaps what Antarctica is doing.

So we're using these different climate models against each other or with each other to understand the different components of the earth system. So I would say that this process is more like ensembling a sports team. You wouldn't have one soccer team full of goalies, you would have a soccer team that has a goalie and a forward and all the other positions that a soccer team needs, just because each of those players is really good at doing what they're good at. But there are definitely some climate models that I would say we may have to bench, that may not do as well, but it's really up to the coach or the climate scientist in this case who can choose which climate models to use for what and how to actually ensemble them to get a good picture of how the climate is changing.

Shanta Puchtler:

And these climate models, how good are they at answering what-if questions, because that feels like that's at the crux of everything, what if we produce less carbon, what if we pollute differently, what if we regrow the forests, et cetera.

Matt Goldklang:

Yeah. I think you are a 100% spot on there. These climate models are huge models, processing terabytes of data. So to just flip a switch and rerun them is actually really difficult. So from the get go, we have this international climate modeling community that says, "Hey, I, what scenarios are actually plausible?" And then we're going to run the climate models within the bounds of those scenarios. So we start from the get go with a descriptive set of ways that us as society will perhaps unfold. And then we try to understand the climate within those parameters.

There are some less complex climate models that don't have the granular resolution where you can't say, "Hey, what's the climate doing here in New York? What's the climate doing in London?" But have a more global scope where you can switch those different parameters relatively quickly, but that won't tell us much information at the actionable level.

Shanta Puchtler:

Okay. Interesting. The other thing that I assume climate modelers do is test their models over historical data, and that sounds like back testing to me, again, using a quantitative finance term. Is it basically the same idea, and how do you prevent overfitting or confounding variables influencing the outcome in a way where suddenly your predictions actually are not valid?

Matt Goldklang:

Yeah. So this is perhaps one of the biggest differences between climate modeling and financial modeling or back tests. When it comes to climate models, we're trying to measure over four billion years of climate history in a sense. And we only have recorded temperature data, for example, or precipitation data for certain parts of the globe for the past 100, 150 years. So observational data really doesn't serve us fully and are trying to understand and back test these climate models. So we rely much more on our physical understanding of the earth to back test these climate models or to hind-cast these climate models. And that's informed by fossil plankton in the ocean, ice cores. So it's much more piece-wise process.

So really, these models are more similar to fundamental physics than they are similar to financial back testing. And we can use the history to help us constrain some of the models, but we're not necessarily fitting the models to the past.

Shanta Puchtler:

So for example, you'd rely on the expertise of a physical oceanographer who would have models of currents and temperatures in the ocean, and you extrapolate that physics in the context of a climate model, that kind of exercise?

Matt Goldklang:

Exactly. And then we would like to see how well the climate model does relative to the observational data, but we can't really rely on the observational data to tell us much. If the climate is so many billions of years old, only a 100 years of data doesn't really give us much insight to the climate. You could just have variability that is as a defining factor. So it's really important that we rely on our physical understanding of the earth, even more so than data.

Shanta Puchtler:

So it sounds like a pretty tricky modeling environment to operate in. We could talk about modeling the climate all day long, but this is an investing podcast, so maybe we should switch gears a little bit. One of the common areas of misunderstanding seems to be the difference between climate adaptation and climate mitigation. Let's start with mitigation. What is it and how is it different from adaptation?

Matt Goldklang:

That's a great question. So I think mitigation is the more classic when you think about climate policy and what policy makers are speaking about. It's the deployment of renewable energy, it's the penetration of electric vehicles and to our car markets, it's the eating of less meat, all the hot topics that we normally hear when we think about climate change, it's the actual reduction of greenhouse gas emissions at their source. And so mitigation is really about global collective policy making to reduce greenhouse gas emissions.

Shanta Puchtler:

Contrast that to adaptation, what are examples of adaptation and how do the two meet in the middle in solving for and addressing the climate problem?

Matt Goldklang:

Yeah. Adaptation is the flip side to mitigation in many sense. It's more about, what do we do now that it's warmer? For example, I'm going to have to buy more t-shirts because my sweaters aren't going to do me less good. It's about actually changing the way that we live and our fundamental economic, physical biological to be more resilient in the face of climate change. And really there are two ways that one can adapt. There's this incremental adaptation, so this is me buying more t-shirts, or there's transformational adaptation, where you're having countries that are on low lying islands having to move their entire populations somewhere else. That's a transformative change in terms of their response to climate change.

Shanta Puchtler:

So if we have $1 to invest in either adaptation or mitigation, where do we invest it?

Matt Goldklang:

Yeah, that is the golden question. It really depends on who you are and where you are and what your values are. I could be in a country, for example, that is now a frozen Tundra and a little bit of warming might do me some good. So in fact, I may be more likely to say, "Hey, I want a little bit of climate change so I can perhaps unfreeze this Tundra have access to more natural resources." And in that case, a separate country, perhaps this low lying island nation state may say, "Okay, we're going to put all of our money towards adaptation, given how other societies are acting."

So it really depends on the global cooperation, this game theory game that different countries are playing with each other.

Shanta Puchtler:

So what is that game theory and how do people think about it playing out?

Matt Goldklang:

I think in the most simple sense, we can think of greenhouse gas emissions, or permission to emit as this global common resource, as in CO2 and the atmosphere is a global bad, but the ability to emit perhaps, or those permission slips a global good maybe, and the more you emit, perhaps the richer you are. So you can say greenhouse gases are directly tied to an economy doing well. But if everyone decides to think just for themselves, then there's going to be catastrophe. But if everyone thinks about the other, there won't be catastrophe. So it's about weighing the costs and benefits as an individual player, how much do I want to sacrifice in terms of my economic growth in order for the entire world to be better off? Again, it really depends on who you are, what the climate change consequences are for you and how reliant your economy actually is on greenhouse gases.

Shanta Puchtler:

One of the things we do as investors is we will screen our portfolios of certain stocks, for example, in a stock portfolio or certain bonds in a credit portfolio, based on their ESG characteristics or the climate impact of the underlying positions and the industries they operate in. And I'm curious because we make an argument that there's a cost of capital issue that is being levered here, that we're increasing the cost of capital for the businesses that are harmful and we're reducing the cost of capital for the businesses that are either greener or making energy transition or doing some other good.

But it takes capital for, or an energy company to become cleaner, to extract more efficiently, to create alternative energy, to store and transport it in a better ways. By increasing their cost of capital, are we actually contributing to the problem in some subtle backwards way?

Matt Goldklang:

Yeah. So this is a really hot topic, I would say, in the climate world right now, trying to understand what it actually means to do the right thing with your money, essentially. And I think perhaps there's been too much attention on emissions, looking back and say, how heavily emitting is the specific company? And while we see energy companies, they emit a lot, let's just ditch them. In my mind, we really have to focus on the energy transition and look out for those who are actually putting the capital in place now to transition.

So if we have a big energy giant who has the intellectual and physical infrastructure to really change the energy game, if they're investing a lot of their own capital now in renewables, rather than fossils, then we should think of that as a good sign, that in fact their emissions will decline over the next 10, 20 years. And they may be really important players and actually reshaping our global energy systems. And so I think we need a more nuanced view when it comes to who we're lowering the cost of capital for. And we need to detach ourselves a little bit more from the standard emissions game, how heavily are these companies emitting specifically among energy?

Shanta Puchtler:

Interesting. Parallel line of inquiry relates to the adaptation side of the equation. A lot of the work that is done in finance is thinking about the mitigation side of the equation. Are there investment opportunities that fall out of the urgent need to adapt? And what do they look like?

Matt Goldklang:

Yeah. I think we have overlooked adaptation as a potential investment avenue. We're really focused on mitigation and ensuring a better climate future, but what happens if that doesn't happen? And we are already baked in about 1.2 degrees of warming into the earth system that will continue to impact us over the coming decades. We begin to feel that not now, but will feel that later. So we really need to think about what our future looks like and how to shield specific portfolios from a detrimental climate future. And that involves really creating a market-based incentive for adaptation. We need to ensure that the companies that we are invested in are ready to have a supply chain that is resilient.

I think one example that I like to think about that is relatively simple, is wine. There's a lot of people who love wine and specific wine growing regions like California, where I'm from, may not actually persist very much into the future. We have these rapid wildfires and Sonoma, devastating wine yields and the likes. So maybe as a wine investor, it's my duty to say, "Hey, where can I start investing in now? What land will actually potentially be productive for wine in the future?" And that's a method of adaptation, that is incremental adaptation where you're saying, "Okay, now, it's New York State's time to shine, where there's going to be beautiful wine from New York State." I don't know if that's the case, but it's about thinking ahead in that way.

Shanta Puchtler:

That connects directly to some work that I know you've done looking at the specific geographic impact of climate change forward 10, 15, 20 years and further out, and what that means for businesses and for local economies associated with increased water levels or increased temperatures or risk of fire or other natural disasters. Can you talk a little bit about that modeling? How do you do that and then how do you it to an economic outcome?

Matt Goldklang:

Yeah. That's again, some really great questioning. I think we need to go back to the coach analogy here. So this is where we say, how do we actually take the climate models that do exist and use them in the best possible way for this purpose. So we're taking those climate models, we're saying, "Hey, these are the ones that we believe in, and this is how we think we should create an ensemble." And from there we say, "All right, we actually need a finer resolution picture than these climate models can actually provide." So we increase their resolution with statistical tools. And we can get a better sense of how climate change will be distributed amongst the globe.

And from there, we can take a look at perhaps a 10-kilometer resolution and say, "All right, what's going to happen here?" And assign those locations to actually where companies operate today. So if I were this wine investor, for example, I could use this climate model ensemble that has a finer resolution to actually pick apart where I should begin growing wine. And then if we want to translate that into a dollar's terms, we can begin employing econometrics where we're tying temperature, at least in history. So we can look back for the past 50 years and say, "All right, how does temperature actually impact wine sales? And can we use that econometric relationship to project that going forward."

The same applies to when we think about labor supply or energy demand and all these bottom line costs that every company faces.

Shanta Puchtler:

Transportation, etc.

Matt Goldklang:

Exactly.

Shanta Puchtler:

Let's close on some forward looking and perhaps more optimistic questions. Given what you know about physical risk due to climate change, what are some exciting opportunities looking forward and solutions for what is such a daunting problem?

Matt Goldklang:

It's really important to shine some good light on this often injury picture. So I think there are a few things that we can look into. One is, in this adaptation space, we can begin to be a little bit more opportunistic and optimistic trying to think, all right, what do we need? And how do we get that done? That's perhaps really exciting. We say, "Okay, climate change is happening, we have to deal with it. What do we got to do?" And that frame of reference change is often useful for some positive thinking.

Another thing we can think about is actually having this data and having the disclosure of physical risk will be really important in actually pricing climate change in markets where we can actually begin to see some movement on climate policy that we would otherwise not have because we have more visibility into what potential bad climate futures may look like. So now we know what perhaps 50 years down the line may look like for us and it's our job to say, "Hey, we don't want that." So hopefully, there's a bout of new legislation that's been passed and that will be passed in many different countries where we're actually mandating that this physical risk be disclosed. And so hopefully, that will push the ball both in the mitigation and adaptation space.

Shanta Puchtler:

Well, let's close there. Thank you for joining us today on Long Story Short. Besides learning that I need to dust off my Fortran manual, we heard a lot about mitigation versus adaptation, and we delved into both geographic specific econometric impacts on companies and on industries, as well as a wide range of other topics related to climate modeling. Thank you so much for joining us today, Matt.

Matt Goldklang:

Thank you for having me. I'm glad we got to chat.

Shanta Puchtler:

And thank you for listening to Long Story Short. Until next time take care investing. For Man Group, I'm Shanta Puchtler.