Dr. Carey W. King from the University of Texas at Austin delves into the intricate relationship between energy systems, the economy, and the environment, exploring his innovative HARMONEY Model and discussing the critical role of energy in shaping macroeconomic policies and future sustainability.

Podcast Transcript

00:06
Jed Dorsheimer
All right. Welcome. Good morning. My name is Jed Dorsheimer, and I will be your host for the Plugged In Podcast, brought to you by William Blair.

So today it an honor to introduce a friend of mine that I've known for a few years with the work that we've done together in the Biophysical Economics Institute. Doctor Carey King and his expertise is, focused on interdisciplinary research related to how energy systems evolve and interact with the economy and the environment.

This is, a broad topic, but really, really interesting and I think carries a theme on what we've been talking about on this podcast, and Carey is a research scientist at University of Texas, Austin. He's also a director at the Energy Institute and a director at the Biophysical Economics Institute. He's got appointments with the McCombs School of Business and LBJ School of Public Affairs, and he is the author of a fantastic book, if you have not read it, titled The Economic Superorganism. It is a must have. It's on the shelf behind me here.

And, so, Carey, welcome. I'm, I'm thrilled to have you on. It's, it's been, it's been a little while that we've talked about it. And I'm glad to, to have you today.

01:22
Carey King
All right. Thank you. Jed. Thanks for having me.

01:24
Jed D
So I thought maybe, just in our discussion here a couple minutes ago, you're actually one of the few people that I know who has this knowledge, in kind of building macro models and has an understanding of energy's primary role, within that. And this is a topic that, you know, we kick around a lot. I know that we talk about it the BPI, of where neoclassical economics fails to adequately account for the role of, of energy. And, and that leads to a bunch of other problems in terms of models that don't really tell you what's happening from a macro perspective. And you've developed a model, I think it's called HARMONEY that, that is more complex and actually has, it's, you know, appropriately accounts for energy.

02:20
Jed D
So maybe just as a starting point, why is that important? Why is understanding energy's role from a macroeconomic perspective important in telling us what's going on in the, in the world?

2:35
Carey K

Right. So, I, I guess I'll try to see why I think it's important just from a, basic, academic or theoretical perspective. And then why it's important from what we're learning from some data and analysis. So, from a fundamental perspective, it's kind of clear that, you know, energy is this thing we, you know, this thing we came up with and we give these properties to things called energy carriers, like gasoline and electricity.

And these go into machines, and they make the machines do something, and they dissipate this energy and create heat. But they also produce work and do physical things in the, in the world around us. So the economy, a large part of it, maybe, you know, not 100% of it, but a large part of it is about extracting materials, turning materials into some other form that we, you know, have value, a cup we drink out of or the computer, that we're computer, communicating on right now.

And from a basic physics standpoint, we know that to, to collect and refine and shape these materials requires, energy and transforming energy into, ultimately into work and heat. And so, we know the economy is a lot about that. And so that seems really fundamental to put in but most, as you're saying, sort of the neoclassical economic approach describes growth as, you know, capital and labor and then some sort of, you know, quote unquote technology that is not well described. And when some scholars have kind of looked at, you know, how would you really incorporate the idea of energy into the concept of a growing economy, or modeling an economy, even from looking at a data perspective, if you calculate, so far, the work that's been done to calculate how much work is done by all machines in the world or in an economy, at least the best approximations we have at the moment. This amount of work done is the closest proxy for gross domestic product. It's the closest metric of a 1 to 1 relationship, which is to say it's much more predictive than just counting all primary energy and relating that to GDP.

The 1970s. You know, GDP grows faster than, primary energy consumption at the global level. And for most countries, but, the country levels that people have studied, the quantity of work done by all the machines is much more of a 1 to 1 relationship. So, it kind of implies that, a lot of what we'd call technology or what a normal economist would call technology is really, inventing things or that can produce more work.

We either learn how to extract more resources so that we have more resources to go into machines and do more work, or you make the machines more efficient because work is essentially fuel input times efficiency to do work equals work. So two ways you can make more work output, increase efficiency, or increase the flow rate of fuel into machines, which means you need more machines.

So all these things sort of translate to more GDP. And that's why that's particularly insightful. Maybe it doesn't help us as much. It's a challenge to try to understand more information, communication technology type things and information processing. That's a bit, harder of a story, but, my macroeconomic modeling sort of, originated from these ideas.

And it was just to, you know, my HARMONEY Model that you mentioned, which stands for Human And Resources with Money Model. My, my background is physical system modeling, kind of dynamic systems and controls and what we call system dynamics modeling of energy systems and so this modeling of sort of energy flowing from one place to another is what I've been trained on.

But I had to learn the money flowing from one place to another and how to model money. And to do that, I kind of tacked on to what people would call the post-Keynesian kind of, way of thinking, which essentially treats banks as being able to create credit or money in that sense, when they make a loan to a company.

And so this is a way to track money coming into the economy and leaving the economy. When you pay off loans. So, my model essentially tracks energy, materials and money as stocks and flows, all consistently and has feedback on depletion effects. It's the simplest model I could make that had a set of features that would enable me to track debt, in a private debt sense, not in a government debt sense.

Track debt track energy flows, track material flows. So, we can have relate metrics like, you know, GDP, if you will, to net energy metrics, energy return on investment. So is essentially the, the, the model has this kind of fundamental factors that enable you to compare these kind of metrics and money that people talk about, but often can't have a hard time comparing to energy or have metrics and energy, but often have a hard time comparing to metrics of money.

The model is made to be consistent, to be a consistent way to be able to compare, these types of metrics and see how they evolve together.

07:45
Jed D
Did you have a particular moment in time where you know that you had this epiphany in kind of an understanding, or was it more of a gradual process? I mean, for me, buying a farm was probably where what really was the seed that I started to think of these things when I recognized that none of the farming could be done without my diesel tractor.

And, and so the mechanization and the, the workloads of, the, the scaling of getting five labor years and in that barrel of oil, for example, you know, that that I either need to have a lot more children, to help me on the farm or, or I need more, machinery. And, and that led me to Charlie Hall and kind of this idea of energy return on investment.

Did you have a similar type of experience that led you or was it just over time that, you know, you kind of got to this place where you're at now?

08:46
Carey K
Maybe I had, a sort of, I don't know, epiphany or light bulb in the head, or at least I kind of saw the vision of what I thought I could contribute or what, a new type of academic contribution that was needed. So, yeah, I mostly was hanging around this kind of net energy community and people thinking about energy return on investment and how to compare energy technologies.

And it's all on, most of it's on a micro level. What about a solar panel or a wind turbine or a gas powered, power plant, gas fired power plant. And then there's, a sort of leap, if you will, to take those kinds of metrics and talk about the entire economy. And the leap was too big, and there was no way to go from the micro to the macro.

And then we had the financial crisis, 2008. I hadn't really been studying, you know, the economy per se, or thinking about economic modeling until the early 2010s, but, the financial crisis kind of, you know, brought to the fore a lot of, inability of the normal, economic methods to be able to anticipate something like that.

And so I, you know, did reading and started learning and came to the realization that, yes, there was the fundamental problem was the inappropriate treatment of money, which is, say, almost, almost a lack of treatment of money in, economic growth models. Which means you can't track who's accumulating debt and who was owed money, from, from loans and these kinds of things.

And I thought, okay, there are people saying, yeah, the financial crisis was clearly this kind of speculative bubble and had to do with credit creation and money and mortgages. And that sounds really relevant. And the more I read you're like, okay, well, that's how the economy works. Banks make loans to people. Or companies and, you don't have to ask necessarily permission beforehand. We just witnessed that they were doing it, their own choice and, you know, risking their own, solvency, if you will. But and then we have this kind of net energy community, and we look at commodity prices and oil prices and the copper prices and natural gas prices were all going up, leading up to the financial crisis as well.

And so there was this question, got implied, or getting thrown around that was in my head is like, okay, is this really like the 1970s? Is or is this, was this an energy or commodity, physical commodity constraint that tried to build too much stuff and didn't have enough stuff coming out of the ground? Or was it a debt constraint, and that we accumulated too much debt and then people didn't have money to pay for other things, and then therefore, they have to start going insolvent and other parts of their life because they're owing too much money.

And I thought, well, how would you know the difference? If you were to look at the data in this economic crash? Is that because of physical, aspects of energy materials? Is it because of debt? Do we actually have a way to know the difference? And the more I looked, I thought, well, I'm not sure anybody actually has a model that puts these ideas together, at a fundamental enough level to know if you can ever tell the difference.

So, they’re always going to kind of come hand in hand. You get too much debt, you're going to get too much materials and energy flow. And then are you going to be able to tell the difference. So, I thought, well okay, that's it. That’s the thing to do is to try to make this model. And since the 1970s, people have been thinking about and saying it's important to incorporate energy and materials and basic concepts of thermodynamics into economic models.

And, you know, obviously, you know, reference the world three model of the limits to growth, which did this in a in a pretty abstract and interesting way. It got lots of notoriety. From today's perspective, the downside of that model is it does not have explicit tracking of money. And so, in some sense, the HARMONEY Model is, sort of like a simplified world three model, but with the explicit tracking of money. In a way an economist would, would want to do it. You talk about things like wages and debt levels and labor and these kinds of things.

12:52
Jed D
So I'm curious, you know, I have a, probably a more simplified view, using, you know, what the metabolic breakeven is of, say, the US economy, which, I calculate, to basically be like, roughly, six and a half times or 6.22 to, to be more precise. And so the concept would be that anything that we're putting in to, from a primary energy that is below, in order for that to be profitable would require debt to, clear, and anything that is above would get real capital formation from that.

So you have this sort of mixing pool. I'm curious. Your HARMONEY Model sounds more complex and, you know, more precise. You know how I guess, would you what are the differences in terms of do you look at it in a, a metabolic, type of perspective in terms of what it tells you or what is that what are the signals telling you right now with your model as you apply that either to the US or to other regions of the world?

14:06
Carey K
To start with, the first answer is what I've published as kind of a toy model. So it's a model of nowhere. It has a structure and a framework that I'm now, I've been spending the last couple of years looking up data to calibrate to the United States. So, so I don't have a model of, you know, the United States yet, but that's what I'm working on is my first, country-based model.

So you can imagine sort of the hassles, tracking all money and energy and materials. It turns out materials is something I haven't worked on as much. And that's been the bottleneck, if you will. Energy and money data sets are a little bit more, common and straightforward. But with that said, what the model structure does is it sort of defines the vines, the natural resource, let's just say the energy or materials resource.

You know, it's the question on a lot of these models are when you're doing macroeconomics is, you know, what is the quote unquote exogenous thing you're assuming. There's some kind of just assumptions you have to make that are the premise of starting the model. And a lot of these effectively, some of these are dictate essentially how the model grows. Why if you will, if there's an answer to why the model grows. In my model is effectively a definition of the natural resource that translates to capital productivity, which is to say, how much capital do you need to function to extract this resource.

And so I define this in a similar way to what was done in world three, is just kind of a simple method, as, you know, as this resource depletes, you need more capital, to get a unit of resource as it depletes. So this is a feedback that prevents you from extracting all of the resource.

And so from my standpoint, I needed that basic concept in there that there's a there is a limit to where you cannot extract 100% of a resource, or at least that's the that would be the premise and that this feedback ultimately is what prevents the economy from growing forever.

16:05
Jed D
Can you give an example? Maybe using copper, you know, copper was 40% concentration, 100 years ago. And today, where does it come to?

16:14
Carey K
Yeah, I don't know where. Copper. I think I hear people talk about this and the point the point something percent or grade.

16:19
Jed D
Point 4% or something.

16:20
Carey K
Yeah. That we're going for when you have one resource. I have unpublished work, which will slightly maybe annoy you with the answer, the question. And you have one resource. It's kind of, clear that you can't extract it all. So my resources. Well, let's just say you're extracting one big oil field or something. And yeah, at some level, you know, you go deeper and deeper and you need the oil to run a pump to pump the oil out.

And at some level, you're spending all of the whatever oil's coming out of the ground, 100% of that oil has to go to run the pump to get the oil out of the ground. And so that would be the physical limit at which you can no longer extract, any oil. So my model is kind of simple.

It acts like a force. So it's as if you're saying, like you have to go deeper and deeper into the forest to chop down the next tree. And at some level, you have gone so far into the forest to chop down the next tree that you need all the energy from the tree and all the resources from the tree to go to the next step.

And you don't have anything else left to run the economy or to give to anybody is households to stay alive or to build the next machine or anything like this. So you're just kind of 100% of the natural resources used to extract the resource. So that's kind of the ultimate limit.

When you have more than one resource this may not be particularly obvious in the sense that, you might be able to completely deplete a particular resource, it helps you deplete a second resource.

17:46
Jed D
Just an example of the resource depletion. I think you use oil as an example. So eventually, it's, you know, or you could use, that, that eventually the energy required to get that barrel out is going to subsume the amount of oil that you're getting.

18:05
Carey K
So say that an unpublished work or playing around with my model where I stuck in two resources, something akin to, like, a fossil resource, like oil and something.

18:14
Jed D
So as you think about this model in terms of the limitations, you know, in this, this wasn't that important, I guess, when we had a physical back currency, because you would inevitably be limited to the energy expended to mine gold, for example, in terms of the money creation, when the transition to a fiat currency, where there is no limitation in terms of the creation of that money, I would think this becomes a bigger issue now, in terms of, you know, data, etc..

18:55
Carey K
Yeah, it gets a little bit beyond what I have insight into at the moment, but at least I can appreciate there's at least two main kind of ways money is created. Some people would call it credit instead of money, but commercial banks making loans as credit. But effectively this gets translated for me, for example, in U.S. dollars to buy my home and pay off the person who built it, and then I pay off those dollars back to the bank.

So that's one thing private, sort of banks. And the other was, yeah, the federal government, creating money, which is what you're hinting at, which had a different relationship or end of the relationship with, a conversion to physical gold. And so it's not clear to me, I, I'd like to learn more about that in terms of how that was really different from the before the 1970s until ‘71, when the gold standard was ended, because commercial banks probably, you know, operated roughly the same, rules have been relaxed over time.

Yeah. It's a good question, but part of part of the modeling, frameworks are you should be able to express both of these ways that money is in some sense created and destroyed one by commercial banks lending. And then me paying back the principal money gets destroyed, and the other by the federal government spending and its currency, and then taxing to, take away some of that money.

So tracking that money is just as important as tracking the energy flows as well. And so if you can do the both of these, or at least you have rules or some rationale for doing both of these, then you have a basis for understanding the relationship between energy and money or energy intensity or any of these other kind of metrics.

So that's kind of the starting point is if you can't do that, so the neoclassical approaches don't track money that way. So in that sense, they would be fundamentally incapable of explaining, fundamental money and energy relationships.

20:41
Jed D
Well, let's take an example, maybe and make it, local. You gave the example. So, let's say that you're a really or let's say I'm an incredibly poor capital allocator, that I love to spend money on things that might bring me some joy, but I failed to provide adequate food for my, family. And I take on more and more debt in order to keep this lifestyle, up.

Now, I'm going to reach a point, and maybe it's like a Minsky moment, in my household, where I have so much debt that maybe I have to even take on debt to pay for that debt, because I've been a poor capital allocator. Now, if I, had sort of a windfall, maybe that's, you know, I find that there's oil under the, you know, under my ground that I didn't know about or some other type of, you know, I would be able to take care of that.

And I give this example because if I look at, you know, the U.S. today and with our $37 trillion worth of debt, I see a limitation from a Minsky perspective in terms of more debt to solve the debt problem. The only other variable that that could possibly solve it would be a influx of net energy. Can you see any other variable?

It seems that energy is kind of the key issue in terms of solving that. Obviously, there's a reduction of spending, but if you have energy that you that end and depending upon what that primary energy actually is and the productivity from that, that would then turn into growth that would then be able to deal with the debt issue.

Is that too simplistic? What am I missing there?

22:42
Carey K
Yeah, I can't answer your question due to a little bit of lack of knowledge, but you have to I would say you have to separate the idea of, yeah, private debt or, the United States dollars owned by the United States government to itself. So, in some sense, the government does it borrow, the United States government doesn't borrow US dollars. It just spends and it taxes. And if there's a discrepancy, then this gets into sort of issuing bonds to kind of, equate, this. So, there's no borrowing from anyone, per se. And the United States…

23:15
Jed D
No, but there's a selling. Right? So if I, if I am going to auction off bonds from a Treasury perspective, then, and if, if those purchasers, whether it be governments or whether it be institutions, believe that my credit worthiness is lower or higher, that's going to affect the interest rates that I'm going to achieve on that auction. And that would affect the yield. Right? So that's going to affect the value.

23:45
Carey K
Yeah, we're getting beyond my knowledge. But my understanding is that the direct issuance of bonds by the Treasury goes in some sense to, you know, a certain subset of banks, right, that have, accounts at the Federal Reserve. And there's no choice there. They just say here, I'm going to exchange cash for bonds. And so, they're kind of an idiot if they don't take bonds that give interest, you know, instead of having cash that has no interest bearing, there's no money, they're going to take this bond, that has interest bearing.

And then there's this secondary market of them selling these to other people. Are these reserve banks selling these bonds to other people that effectively, is how the, the interest rate gets, translated and how much those bonds are willingly exchanged ends up affecting the interest rate, which makes the Federal Reserve, you know, do more or less issuance of, of bonds or buying them back to kind of maybe get to a basic, I know at least, at least an understanding, that I came maybe too theoretically through my model, which relates this kind of, you know, debt accumulation and, energy resource technology, energy capability, at least through my modeling.

Let's just say fundamentally, there's one behavior in the model, if you will, which is, is our agent in the model, there's these two sectors. One sectors job is to extract resources. The other sectors job is to make more capital. And both sectors need this capital to do their job and need capital to make capital. And the extraction sector needs capital, to extract resources.

And so each sector is just saying, well, what are looking at my let their profits and saying I'm going to invest a certain amount of my profits and in this case invest more than their profits, which is, when they invest more than their profits, than they take on a loan from the bank. And this is how the debt accumulates, in the model is by investing more than your profits, or at least trying to. Having the desire to do that.

And that's in the data that this has sort of historically happened. So, you have, you know, push go in the model. We'll start extracting this forest like resource. It'll grow into some other steady state. And you go, okay, I assume that all of a sudden there was some technology that was better at extracting the forest, and this lowers or increases effectively the quote unquote, capital productivity of the extraction sector. And by increasing that capital productivity, their profits go up, and by their profits going up, they invest more by investing more money, they induce the other sector, they need capital production from the other sector. The other sector starts producing capital. Their profits go up. So, the natural resources, or the capability to extract a natural resource and the fact that the resources exist still some there, you know, in some sense powers the, the economy growing, if you will, and you'll end up with a certain amount of total capital in the end, certain amount of population, because that's endogenous. This the population needs some resource consumption in order to stay alive or to have a, a tolerable death rate. So, so ultimately, the death rate becomes equal to the birth rate, and you end up with the steady state.

All right. So then the question is, okay, what other kinds of rules can you imagine that that change this? There's one rule in the model to if you have labor unions somehow talk about what their wages are and if the wages go up with inflation, you can kind of make an assumption to that, that wages go up with inflation within the model. Then the households or the workers will earn a certain fraction of GDP as their wages and the companies will have a certain fraction of GDP as the profits. And this will fluctuate a little bit, but it will kind of end up at the same place you started the steady state. It's a smaller economy and there's a certain distribution. Then there's a larger economy and the distribution is roughly the same. But if you take away this ability, if you change this parameter that says how the wages go up with inflation and you say, well, they don't go up with inflation anymore, then you can end up with a different distribution of capital and population, in the households.

And so you literally start to see how with the same description of the physical world and the same exact description of how good technology is, there's this social parameter that you could play with, in this case, how the wages change with inflation that affects how many machines accumulate versus how many people, how many people accumulate in the model.

And in some sense, both machines and people, to accumulate more people or to accumulate more machines, you need to extract resources and turn it into a machine or effectively extract resources like food. And you turn it into people. So you're literally sort of able to play this little game or to understand some really fundamentals about how social factors like our, our, workers in unions and allowed to bargain for wages that go up with inflation, how this would translate to their, you know, you know, in some sense fundamentally growth rate, but some sort of measure of, you know, welfare if you were, how much of the proceeds of the economy are going to them, and if they don't have the wages, they'll be less of the proceeds of the economy that go to the households versus, the companies as profits. Included, as well as bank interest payments. So you can see little fluctuations in debt go up and, and down with the, with the gross boom, and play around with these kind of, these little games.

And depending on how you choose to model prices and wages, you can see different patterns. But one of the patterns that emerges is similar to the, you know, metabolic scaling relationships that you've talked about a little bit with other guests on your on your podcast, you can plot resource extraction versus accumulation of capital in my model. And it looks like similar patterns, as you see, for biological organisms in terms of their metabolic power, consumption and their mass.

And for me and my model, accumulation of capital is like accumulation of mass. So, so you can see these things and that gives me some confidence to, hey, this model is doing something that has a, and an accounting basis and is a fundamental principles of feedbacks that's probably on the right track or has the right tendencies, because when you look at the real world data in definitive modeling, you look at energy consumption of the whole world and mass accumulation of the whole world or of the US.

You see these similar kind of, scaling patterns which is to say, you know, sublinear scaling energy consumption goes up slower than the accumulation of mass.

30:25
Jed D
You know, keeping it to maybe US. Where this becomes really important is if you have a model, a macro model that only tells you, say, 50% of what's actually going on and has a hard code for sort of what can't be accounted for. You could get yourself into real trouble where you think you're actually managing things in a proper way, but you see all of these social effects where you could have reproductive rates that start dropping off.

30:56
Jed D
I mean, that's what I'm hearing you say, is that the value of having a model that's actually telling you what's going on and properly accounting for the role of energy, I'm interjecting that in that model is really important to understand what's going on for policy, etc. because the risk then is that you have all of the wealth accumulate to a small portion, and you actually have these what we think are exogenous, but they're actually endogenous to the system, right?

31:27
Carey K
Yeah. So, you know, my last paper with my toy model, I still tried to compare it to try to understand, to see if it has some help, see if it helps us understand patterns in the real world that you're kind of hinting at here. So the 1970s were obviously a big kind of transformation point. Most of energy and economic data sort of go through, inflection point in the 1970s.

And one of these relates to this, you know, bargaining power, if you will, of the wage share of, workers or the share of GDP going to workers started to decline, at least in the United States and some of the other developed countries, during or right after the 70s. And so I'd have, you know, someone at a conference from a union was like, oh Carey it wasn't energy your crazy. It was just because workers lost wage bargaining power. And I said, well, I'm trying to go to the question before, which is why did they lose wage bargaining power in the 70s? And they didn't lose in the 1960s and the 1950s. Rather, they even get it in the 30s. And we have the New Deal and all these kinds of things.

And so I was trying to impress upon him that, hey, there's a new constraint in town. It was called energy, and oil. And it was harder for the, for companies to maintain profitability. There is now this cost that they're not in control of it. The price is kind of coming external. The Texas Rail Commission, for example, is no longer in control of regulating the price of oil, because the oil wells are all at 100% capacity utilization and women are less than that.

And if you need the price of oil to stay the same with more oil, well, then just open up the valve a little bit more. So that physical control was gone, at least within the United States. And now I've resided, to some degree elsewhere. And so my modeling, I played around with this idea, which is, okay, what if you lost bargaining power at a same type of interval in my model, as was experienced in the 1970s, which is to say, that's when there was a peak in energy consumption per person.

So if I look at my model, when it reaches energy consumption per person is a peak. And starts to slow down and decline, I say, okay, I make the time. This is when workers start to lose bargaining power. I change that in the model, and then I see a decline in wage share. At that point. And so I think, okay, well, maybe that's exactly what happened is that the 70s came about. It was too hard to maintain profitability. You're no longer in control of energy supply. This is a new constraint. Got off the gold standard, as you said. And, what a company's roughly in control of how much they pay wages. We elected, you know, Ronald Reagan in the US and Margaret Thatcher in the party in the UK. And there was a push against unions and voila, the social contract, the change now, from what it was the previous, almost half a century. And so, it's maybe a question no one will ever agree on is that because there were energy constraints, that there was pressure to change the social contract? To me, it looks like the evidence, there's evidence that that was the case and that, was a real rationale.

34:36
Jed D
So this is particularly germane to what we're seeing today, where the with the advent of AI and as we know, AI consumes these large language models, consume, an enormous amount of energy. We've written some white papers on this, but you have orders of magnitude, more energy consumption per, per query, compared to, you know, traditional, search.

And obviously there's, there's the idea that there's going to be greater value to society from these models. But if I kind of here, you know, take what you've just framed out in terms of the Arab oil shock in the 70s and the consequences that kind of rippled through over the next two decades. From an energy perspective, I'm curious your thoughts is, you see, you know, energy at the tip of the spear debate right now. Combined with, you know, a new kind of mechanization, from an IT perspective, what are your thoughts? It would seem like there's, you know, what? Have we learned anything from the past? And how were you thinking about that in terms of the future?

35:59
Carey K
Yeah. To some degree, I would say there's maybe 1 or 2 consistencies. The particulars are different. So I'm not sure how my generalizations, how much they're going to help, but, I was at one conference where someone just kind of got, I don't know, sick of some discussion about, investments or this or that and why they're made. And he was just like, well, capital has been trying to replace labor for centuries. This is what happens. Right. And so, so that's one constancy, I see is the I just as another attempt to kind of replace allocation to people, let's call it labor, wages going to, to people. Just as in general, fossil fuel machines, if you will, machines with, with various fuels are there to replace physical labor, by humans and animals. So we've gone through that transformation. Not everywhere in the world, completely, but, you know, largely for the developed world, that, you know, most of the physical work is done by machines, not people and animals. So those are two things. So it's a question of whether AI is just an extension of this sort of trying to, replace labor with capital.

And now it's hitting, not going for the physical nature of humans and animals, the physical, our physical capabilities, but our mental capabilities, our, their, language and mental capabilities to some degree. So, so how much energy is it take for that? That's what. Yeah. As you say, you're trying to study and we don't know.

And there's, you know, who obviously there'll be incentives for companies working on a as to minimize the energy consumption. But yeah, it's a good question. How much energy consumption can they take up and still replace a lot of, if you were, intellectual labor, which in principle just tries to increase profits across the board for all companies. That would be my that's my first take on it.

37:58
Jed D
Yeah. I mean, the point that, where I was trying to get, I guess, the value of having a model that appropriately accounts for these things is what is going to coming in to these turning points. That's what's actually increasing. So, do you have something that understands what's actually happening or are we flying blind? And it seems to me that lessons of the past are we kind of went through things in a in a blind way.

Solow's residual comes to mind in terms of those plugs versus today. Where can we actually learn or are we destined to repeat, failures of the past? Because, you know, AI does present an opportunity for another mechanization and sort of scaling effect. And I hope this is my goal here with the podcast. And, and, in the work that, that I do at William Blair is, is to hopefully educate on that role of energy because it seems like that is the key ingredient for all of this.

39:06
Carey K
Right. You know, I go back to this idea of, useful work or we get nerdy, call it useful exergy a little bit, but, a lot of the solo residual, something like 80% of it, if you look sort of over the last 100 years for the countries that have been studied, something like 80% of that residual is described by the increase in efficiency of converting primary energy into work.

So that's a, you know, in some sense a real physical thermodynamic type of description. And when you go into AI or interpreting data and whether it's a medical scan or something like that, or just writing some sort of article, faster than, than a human, otherwise, alone would have written it, that becomes a little that's sort of a, a question in this community of how in the world to try to think about the conversion of energy into work effectively, it's conversion of energy into electricity as work, which then runs a computer that does some processing and trying to translate that into, yeah, yeah, some sort of value or technological growth.

But from the, you still can't throw away the physical description of the world. So I would just say that AI, and this kind of, computer IT based description of the world or as, technologies is something you just have to you have to add on. It's not a replacement for, the physical work going on and the efficiency of machines doing physical work.

It's an add on to try to understand the next step. So a little bit beyond what I know how to think about right now to translate something like, yeah, AI and to ultimate economic value or GDP. But it's a good question.

40:49
Jed D
Well, it sounds like you're working on a few things. Maybe to wrap up here, how can listeners find out more about your work or, contribute to what you're doing, how anything that you would like to plug, in terms of what you're working on?

41:09
Carey K
Yeah, I'm semi active on Twitter, I guess Carey King. You can find me, University of Texas. If you somehow wanted to donate to my research program, you can find a donation link on my website, CareyKing.com. And in terms of things I'm working on one thing that's good right now that if anybody wanted to follow up on it would be good to follow up, is that I've arranged, I've been funded to organize a couple of workshops over the last year, and these workshops essentially gathered people who do what I view as the correct type of macroeconomic modeling.

People I've learned from and people using what we call is, yeah, for the most part, a post Keynesian framework or stock flow. Consistent monetary accounting just means you're kind of tracking where money comes from and where it goes and how it accumulates. So these are all really important. And I work on the biophysical aspects of integrating energy and materials into these types of models.

So, so we have a philanthropy here in Texas that has, sort of paid us to get together and start to talk about how we would work together and how we actually kind of ultimately make, a macroeconomic model that could, you know, be what we all might envision as the, as the, a model that can actually inform climate policy and decarbonization policy because it would have these feedbacks about how much you invest and how much energy and materials it takes to direct investments towards decarbonization.

It would have the aspects of how the efficiency changes of the economy, if you integrate large scale carbon capture and storage, which in some sense, by definition is a decrease in the efficiency of machines doing work because you're, you know, going to prevent CO2 from going in the atmosphere, you're going to suck it directly out of the atmosphere.

So accounting for that work done, accounting for the materials and energy, how does that translate to money creation or value creation for the companies involved, and the firms involved. So, so we're actually trying to figure out how to work together as a group, and write proposals and reach out to organizations to be a part of this because, so I'm writing a couple of papers out of these workshops, and one of them will just kind of be about how the current suite of integrated assessment models are insufficient, to actually answer questions about decarbonization for reasons you've touched on, and mostly because they're neoclassical based and, based on an equilibrium type, modeling instead of non-equilibrium type modeling, where you can actually understand the dynamics of debt and energy, during a transition. So, it sounds odd that these models are used to help us understand the energy transition, but sort of the concept of time and transition is actually not really, embedded fundamentally enough in how they work.

So obviously we would look for help on that for anybody who wanted to, to help with that. And you could. Yeah. And reach out to me through my website or.

43:59
Jed D
That sounds great. Or if you want to get in touch with Carey, reach out to me and I'll put you in touch with them and, yeah, I'm looking forward to having you back. And maybe that's the next topic to try to dig into, to some of the integrated models. So, Carey, thank you for, for joining me today. I really appreciate it.

44:16
Carey K
Thanks for having me, Jed.