William Hogan, the Raymond Plank Research Professor of Global Energy Policy at Harvard's Kennedy School of Government, along with his colleague Scott Harvey, are the architects of the market structure employed in every competitive regional wholesale power market in the United States, known as the bid-based, security constrained economic dispatch model with locational marginal prices.
This market model has been the subject of criticisms from its inception some three decades ago, but in recent years it has come under renewed attack from interests who claim it is inadequate to support a transition to a clean-energy grid comprised of many zero marginal cost intermittent renewable generation resources.
Hogan and Harvey authored a paper in October 2022 defending and explaining the market structure incorporating locational marginal pricing, or LMP. In this episode, Hogan elaborates and explains that, not only are critics wrong to suggest LMP markets aren't up to the task, but they are in fact the only means available to facilitate the clean-energy transition. "If you want an efficient market in the electricity system, this is all the only way to do it – that we know of," Professor Hogan tells the podcast. "This model is ubiquitous and successful. I call it successful market design."
Hogan scoffs at criticism that the LMP market design isn't the correct vehicle to accommodate increasing amounts of zero marginal cost renewable resources. Such criticisms fail to recognize the role that battery storage and demand-side resources will play in decarbonized markets, he says. "Even if you had 100% renewables (on the grid), there will be periods of time when you’re capacity short because you just didn't have enough of them. And then the demand side is going to be much more important because we're going to have to adjust demand. And implicit in the theory, or explicit in the theory, is that the price will be set by the demand side, not by the supply side, but it's the same marginal cost principle. You get the same answer in terms of economic dispatch. And you get the same pricing outcome across the grid. So absent the fantasy world of an infinite supply of zero marginal cost energy, that theoretical problem doesn't arise."
The empirical proof that the criticism is misplaced is evident in the Western interconnect, where increasing numbers of utilities are joining the Western Energy Imbalance Market, which employs LMP pricing, primarily to accommodate the rapidly increasing amounts of renewable energy resources in the region.
"This is a very powerful empirical argument that says, no, you’ve got it exactly backwards. It's not that we need to get rid of this. We need to strengthen it and expand it," Professor Hogan says. "Because the basic design works in theory and it works in practice, and all the other things that we have tried – I don't know how many of your listeners know but PJM tried something else first. And it didn't work. And New England tried something else first, and it didn't work. And California tried something else first, and it didn't work. And Texas tried something else first. And it didn't work. And then they all had to revise and reform and go through the agony of that process to get to bid-based security-constrained economic dispatch with locational marginal prices, which is now done across all of these. And New York started this from the beginning, the Midwest started this from the beginning, the Southwest Power Pool when it converted to an RTO started with this, so that's why it's now true in all of these markets in the United States."
William Hogan, the Raymond Plank Research Professor of Global Energy Policy at Harvard's Kennedy School of Government, along with his colleague Scott Harvey, are the architects of the market structure employed in every competitive regional wholesale power market in the United States, known as the bid-based, security constrained economic dispatch model with locational marginal prices.
This market model has been the subject of criticisms from its inception some three decades ago, but in recent years it has come under renewed attack from interests who claim it is inadequate to support a transition to a clean-energy grid comprised of many zero marginal cost intermittent renewable generation resources.
Hogan and Harvey authored a paper in October 2022 defending and explaining the market structure incorporating locational marginal pricing, or LMP. In this episode, Hogan elaborates and explains that, not only are critics wrong to suggest LMP markets aren't up to the task, but they are in fact the only means available to facilitate the clean-energy transition. "If you want an efficient market in the electricity system, this is all the only way to do it – that we know of," Professor Hogan tells the podcast. "This model is ubiquitous and successful. I call it successful market design."
Hogan scoffs at criticism that the LMP market design isn't the correct vehicle to accommodate increasing amounts of zero marginal cost renewable resources. Such criticisms fail to recognize the role that battery storage and demand-side resources will play in decarbonized markets, he says. "Even if you had 100% renewables (on the grid), there will be periods of time when you’re capacity short because you just didn't have enough of them. And then the demand side is going to be much more important because we're going to have to adjust demand. And implicit in the theory, or explicit in the theory, is that the price will be set by the demand side, not by the supply side, but it's the same marginal cost principle. You get the same answer in terms of economic dispatch. And you get the same pricing outcome across the grid. So absent the fantasy world of an infinite supply of zero marginal cost energy, that theoretical problem doesn't arise."
The empirical proof that the criticism is misplaced is evident in the Western interconnect, where increasing numbers of utilities are joining the Western Energy Imbalance Market, which employs LMP pricing, primarily to accommodate the rapidly increasing amounts of renewable energy resources in the region.
"This is a very powerful empirical argument that says, no, you’ve got it exactly backwards. It's not that we need to get rid of this. We need to strengthen it and expand it," Professor Hogan says. "Because the basic design works in theory and it works in practice, and all the other things that we have tried – I don't know how many of your listeners know but PJM tried something else first. And it didn't work. And New England tried something else first, and it didn't work. And California tried something else first, and it didn't work. And Texas tried something else first. And it didn't work. And then they all had to revise and reform and go through the agony of that process to get to bid-based security-constrained economic dispatch with locational marginal prices, which is now done across all of these. And New York started this from the beginning, the Midwest started this from the beginning, the Southwest Power Pool when it converted to an RTO started with this, so that's why it's now true in all of these markets in the United States."
EMP S3E7 William Hogan, Raymond Plank Research Professor of Global Energy Policy, John F. Kennedy School of Government, Harvard University, transcript
(edited for clarity)
EMP: Welcome to the Energy Markets Podcast. I’m Bryan Lee, and our guest today is William Hogan, Raymond Plank Research Professor of Global Energy Policy at Harvard’s John F. Kennedy School of Government. Professor Hogan is an economist and along with his colleague Scott Harvey is the father of the market design used in the regional wholesale power markets across the country. In recent times he has been defending this market design against critics who suggest it is inadequate to support the clean-energy transition we must undertake to respond to the threat of climate change. Professor Hogan, welcome to the podcast.
WH: Thank you for having me.
EMP: Okay. So let me see if I can get this straight. So the market design used in ISO and RTO markets employ a single clearing price at congestion points on the grid. So whoever clears the hourly or day-ahead market are paid the same price at these locations as a resource that clears that locational price. And this is what's called locational marginal pricing (LMP), which is intended to take into account system constraints or congestion on the grid. Could you flesh that out for our listeners to understand a little – perhaps a little better?
WH: Well, first, just a slight clarification, which is the locations are determined by geography, electrical geography, where you’re located on the grid. And congestion is what induces price differences at different locations – primarily congestion. There's a little bit of effect of losses too. But the major part is from congestion. So the market clears and we get a price at every location, which may be different at every location, but for everybody at the location the same price applies to them.
EMP: Well, I mean, this is a market design that has been in place for what 30 plus years and I think throughout that time, it's been subject to criticism. I remember when I was at FERC, I had to try and defend it even though I still don't quite thoroughly understand it. But has this market framework benefited electricity consumers?
WH: Oh, definitely. There's, there's of course, an important starting point of decisions. So if you’re a traditional utility in the United States in the 1980s and a vertically integrated monopoly, then that's a model which can work. It creates problems of its own, but we were doing that for a long time. We have many places in the world where you have a single company, which is a state-owned enterprise, running the electricity system, and so on. Or you could make a choice which says, in order to address the problems that come from that monopoly, you can turn to markets. And that's a complicated policy choice with many dimensions to it. You could have different views about whether monopoly is better in principle in practice or markets are better in principle, and in practice. I'm a markets-oriented guy, so I know where I would come down. But the important point is once you make that decision, and you get past that starting point, if you start going to markets then there are certain things that become necessary, that you have to do, if you want to get good outcomes. And that's where this design comes from, as I can explain.
EMP: Please do.
WH: So, there are many characteristics of markets, but let me emphasize four of them that I think are relevant here for some of this discussion. The first one is that in markets, the participants have discretion. So they can choose whether to generate, how much they're going to consume, how to schedule things and so on. It's not somebody else making the decision on their behalf. So they have discretion, and they will respond to the incentives that they face. Second, we know because of many technical characteristics of transmission grids that we need a system operator to coordinate the flow of power and the generation and the load on the system at any given moment. And that exists in the monopoly model and has existed in the market model. That's a fundamental technological requirement given the current technology. So there is a system operator who is coordinating all the real-time activities. The third point is that the transmission grid is today and probably for the likely future is not a “copper plate,” as the term of art goes. So it's not that everybody's at the same location. That's the image of the copper plate, you can connect anywhere and it's all the same. We have transmission lines with complicated flow patterns and grid constraints throughout the system. So you're going to get sometimes very different economic conditions at different locations on the grid at any one moment, and that's important and has to be accommodated. And then finally, which is important, but I won't elaborate on that too much today, is the final customers who are buying the power are interested in what happens to them at their location. So you have to take that into account when you're thinking about hedging your [unintelligible].
EMP: So you have a system operator. In a monopoly, it's the monopoly utility that's operating system, obviously. So in a competitive market, we have what's an independent system operator, which takes control this system away from the monopoly utility, and I guess reduces any incentive for self-dealing in the process.
WH: It solves many problems, the independent part of the story. But the critical requirement that flows from those four conditions I talked about earlier. One of them is you have to focus on real time, what's going to happen in real time with the actual power flows, which is going to be coordinated by the system operator. The way you design the real-time is going to have an effect on everything else. Because market participants who have discretion will be looking forward to the real-time when they're making preparations or contracts or whatever. They will respond to the incentives. So if, for example, you were to say as we did for a short while, imbalances could be adjusted essentially for free, then everybody would be out of imbalance pretty fast, because they would be relying on the fact that they get that service for free. But you can't do that. So you have to design the real-time market first. Most people don't want to think about the problem that way. They want to think about well, what are we going to do a day ahead a month ahead, long-term planning and all that sort of thing. All that is very important, but in the context of markets, the logic flows the other way, what are you going to do real time? Then what are you going to do day ahead? Then what are you going to do a month ahead? And so on. So the critical thing is to get the real time to be efficient. That brings us to, since the system operators coordinating the real time, what should the system operator do? What should the market participants do? And in the monopoly world, we learned long ago, the system operators know how to do this is the basic concept is economic dispatch – security-constrained economic dispatch. So you should take the combination of generating plants that is the cheapest in order to meet the collection of load. And it's the combination part that's important. So you have to have to look across all of them. System operators know how to do this. They have the capability to do this coordination. They were doing it for years before under the monopoly system. So that's not an innovation. It's just continuing something that was already there and that was important. The new thing that arrives with markets is, what do you charge people for the power that's taken in real time? Either directly because that's the only price, or indirectly through imbalance payments if they have a schedule in their little bit off schedule as everybody is. The answer to that problem falls out of the economic dispatch problem. And the answer was laying there in plain view, but hidden, because people hadn't thought about it this way. But the economic dispatch produces automatically what are called marginal costs, or locational marginal costs, or locational marginal prices. And the idea in markets is to use those prices which are also readily available. And now to use that to pay people for the power they provide or to charge people for the power that they take. And when the system is constrained, which it is in most systems a good part of the time, those prices will be different at different locations. And so that's where you get to locational marginal pricing. That model is internally coherent, it marries the economics and the engineering, so it's consistent between the two of them. And it is the only model that provides the efficient solution in real time. And it sets up a scheduling regime which people can anticipate in day-ahead schedules and planning forward and so on. There are certain things that we could go into further about financial transmission rights, but that's probably an appropriate topic for another conversation.
EMP: Well, I don't know. I think I was just going to tee that up, you know, so, what you have with an electricity system is contractual arrangements either within or without the market that don't follow the contract path, so to speak. Electrons follow the path of least resistance. So depending on what sorts of transactions are occurring, you can create a lot of congestion on the grid which limits the amount of electricity that it can transmit. So by having locational marginal pricing, you take these constraints into account, you price them, and then those can be taken into account in these financial transmission rights that you were talking about, which can in and of themselves be traded back and forth in a market-based system.
WH: Right. This is a hard problem. Electricity is different than most other things. So there are kind of two simple versions of the electricity system that are often in the back of people's minds when they're trying to develop policies for it. One version is the one I've already mentioned, which is the copper plate. Everything is at the same location and location doesn't matter. And we know that's wrong. And we know it's wrong in a big way. The other version, the simple version, is I can make an arrangement from my generator and follow the contract path through the network to get to you as my customer. And if I can get the appropriate transmission assigned to me to use that contract path, then I could trade and do all the things that we find in other markets, and you wouldn't need anything different or more complicated. That would be good if it were true. But it's not true in a way that's very different than the copper plate. So the power doesn't flow through along the contract path, it flows according to the laws of physics through every parallel path to minimize the – equate the marginal impedance or minimize the total losses if you want to think about it that way. And the power flows create enormous what economists would call externalities. Because if you and I are doing a transaction, we can have a serious effect on the capacity of a third party who's not involved in that transaction, and they may not be able to do their transaction. So you get this very complicated interaction problem. The only solution I know about is economic dispatch with locational prices. And then you create financial transmission rights, which are point-to-point. They don't specify the path through the network. They just say, we're going to give you the right to inject at one point and take it out at another. That's the point-to-point story. But not in a physical sense, but in a financial sense. So if you actually match your physical schedules against the financial point-to-point then it's like a perfect hedge. And if you deviate, you have to pay something for transmission, or you're paid because of these rights. You can, under certain reasonable conditions, you can assure the, what's called revenue adequacy of those financial transmission rights so that the system operator who is maintaining this system actually doesn't take any financial exposure from the system of financial transmission rights. So put all that package together and you have an efficient way to operate the market in real time. And then you have these financial rights to get from one location to another and then people can start doing contracts or schedules day ahead, month ahead, week ahead if they want to, or they can do a whole bunch of other things mixing and matching. We can have a trading environment that provides the efficiency that we need for both short-run and long-run. It's an integrated package and a critical part of this is the locational marginal pricing, which marries the engineering and the economics. If you want an efficient market in the electricity system, this is all the only way to do it – that we know of.
EMP: And so by having an LMP-based real-time market, you create the foundation for day-head markets and these other contractual types of transactions, forward markets and forward transactions, and all of that, and that gives you market liquidity.
WH: Correct.
EMP: Okay, well, you said this is the only way to do it, or at least the only way we know how to do it, and apparently there's disagreement about that. At the top of the paper that you and Scott Harvey wrote last fall in which you were explaining the LMP, defending the LMP, and why it was the best vehicle to take us into a clean-energy economy. There's an active element that's questioning that. And you noted at the top of your paper that FERC Commissioner Mark Christie in his concurring opinion in the commission's Modernizing Wholesale Electricity Market Design proceeding (AD21-10), he said quote, “...it is time to put the all-important question of the continued use of locational marginal pricing (LMP) in these market constructs on the table for serious scrutiny and discussion.” So apparently, Commissioner Christie's opinion has been informed by a report that was put out a couple of years ago by former FERC Commissioner Tony Clark and Vince Duane, a former longtime PJM official. It was called Stretched to the Breaking Point. And they questioned what they called the orthodoxy of RTO and ISO markets as the best ways of accomplishing the clean-energy transition. They wrote that given policies promoting clean energy resources the markets need to be reevaluated from the ground up, as Commissioner Christie echoed, because they are “misaligned with public policies that seek to advance grid decarbonization.” I’ve read through it a couple of times, but apparently they feel that, or they're concerned, that since all of these renewable intermittent resources that we want to bring on to create a clean-energy grid are zero-marginal-cost resources. And so how do you have a locational marginal price construct when the majority of the resources on the grid have zero marginal cost and can't be dispatched?
WH: There's a theoretical answer to that question, and then there's an empirical answer to that question.
EMP: Well, let's do it one at a time.
WH: Let's do the theoretical answer first, which is the underlying model that I summarized earlier about economic dispatch. When I was describing it, I did not – because I didn't have to – I did not say anything about the cost structure of the generators on the system. I didn't say that the cheapest ones were perhaps new nuclear when we were designing this, operating costs, and then the next would probably be oil, and the next would be gas or coal and taking on all of the conversations about the cost of different generating plants. I didn't say a word about it because the design is neutral to this. Yes, it doesn't matter. The same theoretical story applies. Now, the prices that you'll get at different times will be, of course, different depending on what the composition of the fleet is, the generating fleet, and the demand participants and all the other things that are going on. So it'll have a major impact on the outcomes in terms of prices and quantities. But there's nothing in the theory that depends on any of that cost structure story. Now, let’s take the implied extreme case where the variable cost is zero and it’s the entire fleet and we have a lot of it. So the demand is always below the capacity of the fleet at any point in time. Well, in the short run, if you want an efficient solution in real time, what should you charge? The answer is zero. Because that's what it costs. And if you do anything else, you're going to deny some users the benefit of using things that are really cheap. And so just from a logical point of view, you end up with zero prices and that's the right answer. Then there's the question of how you get the investment and how you get the capacity there. But that's a separate question than this real-time efficiency question. In reality, that's not what's going to happen. So you're not going to have infinite capacity. That's the first thing. So even if you had 100% renewables, there will be periods of time when you’re capacity short because you just didn't have enough of them. And then the demand side is going to be much more important because we're going to have to adjust demand. And implicit in the theory, or explicit in the theory, is that the price will be set by the demand side, not by the supply side, but it's the same marginal cost principle. You get the same answer in terms of economic dispatch. And you get the same pricing outcome across the grid. So absent the fantasy world of an infinite supply of zero marginal cost energy, that theoretical problem doesn't arise. And then you'll have other technologies, which I didn't discuss in particular, but an important one would be storage. So if you have some periods where prices are high in some periods where prices are low, then you have an arbitrage opportunity, and you could build storage and you can buy low and sell high and then that brings the prices back together. And so they aren't so volatile. So one of the things that storage will do, does now and will do in the future is to remove some of that volatility in the prices. You'll still get more volatile prices if you have more volatile sources, but nonetheless, it won't be the extreme cases. So when you're talking about the reality here, in theory, of how this all works, nothing actually changes in the design, the principles. The numbers are going to change in how you invest and how you make hedging arrangements. And plans might be different, but the basic design stays the same. That's the theoretical answer. Now we have an empirical test of this, which people often don't recognize, but I know FERC recognizes, and I use this in talks all the time, which is the picture of what's happening with the energy imbalance market out west in the Western interconnection. So you mentioned that all of the organized markets in the United States use the model I've been describing, which is true, and I can explain a little bit about the history of that. But one very large area where they don't do this is in the Western interconnect. It didn't do this. And we have California out there but the rest of this is not so much or was not so much in these organized markets. And what you saw arise was two things. One was an opposition to becoming an independent system operator or regional transmission organization a la FERC. But a recognition that the growth of renewables was going to require a lot more coordination across all of these regions. And so what they layered onto this is something called the Energy Imbalance Market. And this is voluntary, many utilities have to join it and they do so voluntarily. And if you go look at the map, you could see there was the first one and then another than another, another. Now it's covering a very large fraction of the total Western (interconnect) this year, I guess, the schedules of utilities joining. And what is the Energy Imbalance Market? Well, it's not quite the real-time market that I was describing before, but it's pretty close. Because it's 15 minutes ahead. And then you have all of these participants submit offers and bids and schedules and so on. And then they have a system operator coordinator who calculates the economic dispatch across all of these bids and everything across the whole energy imbalance market, which includes California, and they calculate locational marginal prices and charge and pay according to these locational marginal prices for these schedules 15 minutes ahead. And they're trying to extend that to a more forward market or possibly go closer to real time. And the system operator for California is the system operator for the energy imbalance market across the whole west. So they have this essentially real-time market, or close to real-time market, which is coordinated in the same way and uses the basically the same principles. And it was driven by renewables. So this is a very powerful empirical argument that says, no, you’ve got it exactly backwards. We it's not that we need to get rid of this. We need to strengthen it and expand it. Because the basic design works in theory and it works in practice, and all the other things that we have tried – I don't know how many of your listeners know but PJM tried something else first. And it didn't work. And New England tried something else first, and it didn't work. And California tried something else first, and it didn't work. And Texas tried something else first. And it didn't work. And then they all had to revise and reform and go through the agony of that process to get to bid-based security-constrained economic dispatch with locational marginal prices, which is now done across all of these. And New York started this from the beginning, the Midwest started this from the beginning, the Southwest Power Pool when it converted to an RTO started with this, so it's that's why it's now true in all of these markets in the United States. And there are conversations going on and other countries around the world. They started it a lot earlier and New Zealand and Chile. So this model is ubiquitous, and successful. I call it successful market design – SMD, that you remember from your FERC days.
EMP: Don't give me PTSD.
WH: (laughter) But it is the successful market design. And it's the only design that marries the economics and the engineering that's internally consistent. Now the opposition to this and the criticisms of it fall into I would say broadly speaking, three categories. So one category is to get confused about what's going on. So this is some of the writing that you’ll see that will imply that the single market price applies across the entire region. And the answer is no, that's not true. That's the copper plate story. And no, we don’t know how to do that. That's just a mistake. And we've tried to correct people who make that mistake. Another kind of problem that you run into, which is a little bit harder to deal with is – it’s sort of incoherent, just say, well, I don't know, why don't we try this? Why don't we try that? When we you know, still there's no systematic analysis to support it that’s just a lot of assertions. There's a third kind of criticism, which I think is much more relevant. And that would be a criticism that there are tradeoffs. So if we could do something that was a little bit more in one direction or another, we'd give up some of the efficiency that we're talking about, for example, in real time, but it will be made up by increased liquidity, you know, in the market, or an easier way to do forward contract hedging or something like that. And as long as you aren't giving up too much, you know, you don't cause the lights to go out or it is very expensive, then, well, maybe that's the right trade off. That's a that's a more coherent argument. And that's an argument which I would say is the conversation they're having in Europe today. They understand the theory that I described. But then they say there are other things we have to think about, the political constraints and blah, blah, blah. The problem with those arguments is that they're not empirically based in the sense that it can be really bad. So you can have very serious problems if you don't have an efficient market design in real time. And once you have that, then everything flows from it. I haven't seen an argument where the tradeoff was worth it. So it's, and it isn't that difficult to do. And once you once you do it, it becomes not so controversial. It's just straightforward and it works. And the fact that we tried, you know, slightly different, four different models in the RTOs in the United States, and they all failed, and they all switched to this one, and then all the others used this one for the beginning, that tells you that's an empirical statement, which is pretty powerful.
EMP: What really puts the nail in the coffin of this argument that you cannot use LMP to get into a clean-energy grid is the fact that throughout the West, you've got all kinds of entities that at one time were opposed – adamantly opposed – to RTOS are now adopting the LMP market specifically to facilitate the market entry of renewables.
WH: Right.
EMP: So there is an issue called “missing money.” And I guess that's where capacity markets come in? Can you explain where the missing money is and why we need capacity markets within this LMP construct?
WH: I think a full disclosure here. I can explain the theory behind the missing money story.
EMP: But you’re not an advocate of capacity markets.
WH: I was approached when I was working with the cooperating utilities that were in PJM, everybody except PECO at the time. And we were trying to advance this model that I described, and we lost incidentally, because PJM started with something else that was recommended by Enron. I was asked in one of the prep meetings what we could do to put in a capacity market into this system. And I said, well, the capacity market is completely inconsistent with the theory of the case. And it won't work. And this was a, you know, not a public conversation, but I was a participant so I can say what I said. And the response was, well, we need one because the regulators want it because they think they need it. So I said, well, I don't think it's going to work. I'm busy. I don't want to get involved. (laughter) So I abstained from participating in those conversations and my theory was is that it would collapse it of its own weight because it didn't work and it would go away.
EMP: Certainly it's starting to collapse of its own weight, even if it hasn't quite gone away yet.
WH: It never seems to go away. We keep trying to fix it. I'm a capacity market skeptic but let me give you the underlying theory here.
EMP: Okay.
WH: What you have to worry about. It's kind of it's kind of a copper plate theory. And so it says, well, we need to have enough capacity around in order to maintain certain kinds of reliability standards. But if we have these market-clearing prices, and we're worried about politics and monopoly and all the other things, we can't let the prices get to be too high. So we'll have some kind of a price cap on the system so it doesn’t get is too high. And then we'll let it freely fluctuate under this price cap and then it turns out that that doesn't generate enough money to pay for the investment in the generating plants.
EMP: Well, if I can interject, the missing money is due to price caps, artificial price caps.
WH: That's one argument.
EMP: Okay.
WH: Another argument is, well, if you take the extreme version of marginal cost pricing, and you ignore the demand side and you say, well, we'll set the price at the most expensive plant that's running. How does that plant ever make any money? So that's part of the question. So because the price is never above their variable cost of operation, so they can't make any money. So either way you look at it, which is, of course a conceptual mistake. You don't want to set the prices like that. You want the demand-supply interaction. And you also want to account for scarcity conditions and operating reserves. And there's, you know, that's a whole another issue here are these ancillary services. So it's an idea, you know, that there might be forces in the market design which prevents you from making enough money to effect investment. And the easiest case was the plant that’s got the most expensive operating cost – think of that as an old-fashioned peaker and then you say, well, how do you pay for the peaker. Well you can’t under that marginal cost, you know, literally, the lower price. So and that's a payment which affects everybody, because it affects the gas plants, the combined cycle plants it affects the coal plants it affects the nuclear plants it affects the renewables. It affects everybody because prices are suppressed for them too during those scarcity events that affects everyone. So you end up with a story about how you won't get enough capacity. And now we need something to make up for the “missing money” that has been asserted in these ways. And we get these capacity markets where people want to identify capacity that is going to be needed. And that raises a whole set of other questions. And I'm trying very hard to be neutral and objective here.
EMP: There's no need for that. You're not a journalist.
WH: Now, there are two issues here that I think are hidden in this conversation quite often, which is, one is a lot of discussions about capacity and capacity markets are extreme case stories. So this is one-day-in-ten-year reliability. We need to have enough generating capacity at the right locations in order to serve the load that makes sure that that capacity is deliverable to the load so that we can meet our one-day-in-ten-year planning standard. That's a really hard problem to solve. My view is that we don't know how to do that. Because if you think about what it's saying is that we can identify the generators and we can identify the load, we're going to match the generators to the load. And we just decided we couldn't do that in real time. Because the contract bypass model doesn't describe how the system works. And it doesn't account for all of these interactions. And if we can't even do it in real time, knowing as much as we know, how can we do this months and years ahead, as required by these capacity markets? So I think that's actually a problem we don't know how to solve and can't solve and we shouldn't kid ourselves. The second problem that people are worried about is having enough capacity so that prices don't get to be too crazy. So we don't get into extreme – we reduce the number of times to get into extreme conditions where prices are very high. And that's a, if you think about that, that's a hedging problem. And, and now I go back to my the fourth principle I talked about earlier, which is that customers care about power delivered to them, not power a producer generates. So if we had a copper plate system, well, that would be okay, because those would be the same thing. It wouldn't matter. And we could write contracts for generators to produce power at their location. And that would provide a hedge for customers. But that's not the reality. And what we should be focusing on therefore is what's most important, which is the volume and prices of power delivered to customers. That’s the volume and prices of power produced by generators. And then if we could create those hedging arrangements and people signed up for them or were required to sign up for them, that would create a market and generators could arrange their own forward contracts and they could do all the kinds of things that are going to be a risk-reward tradeoff that they will make in making those decisions. But are decisions that could not be satisfied in a regulatory proceeding, you know, that you wouldn't be able to provide the kind of scrutiny for them to say that this is really going to be sufficient. And so I think the whole resource adequacy hedging story and investment story is misplaced. We should be focused on the customers, not the generators, and because our objective course in the end is what happens to the customers, not what happens to the generators. And the generators can work out market arrangements using the bid-based model and financial transmission rights and intermediaries and hedging contracts of their own and so on in a way that's going to be much more complex than the regulators could manage. And then the regulators can use – my favorite example is basic generation service in New Jersey, which is not unique, but it's one of them. And I've looked at it in the past and occasionally catch up. But basically, what that service is, it's for small customers, and small commercial and residential customers. It's an opt out system. So you're – if you don't say anything in New Jersey and you're a homeowner, then you're in it. If you don't want it, you have to tell somebody, and then you have to meet certain requirements like you have a meter and so forth in order to get out. But you can get out if you want to. So it’s an opt out system. And then for the ones that are included, these small, commercial and residential customers, it's a rolling three-year contract where for the full requirements, and so every year they have a new auction for delivered power at a fixed price for the next three years, and they do it for one-third of the load. And then next year, they do another third and the next year, staggered, you know, so it goes forward so that means that it smooths itself out. So the price is fixed for a year but not for two years. And it's only partially fixed for two years and so forth to three years and eventually it gets back to market conditions and whatnot. So that's a kind of reasonable, smoothing, hedging focused on delivered power. And the people who are winning these awards, some of them own generation and some don’t. And the ones who don't are arranging other kinds of contracts that the regulators don't have to know about and don't see. And that system works. It's worked for almost 20 years, I think. It's uncontroversial, as near as I can tell. You never hear about it. At least I don't hear people...
EMP: It's not controversial unless you're a retail power supplier in New Jersey
WH: Oh, well that's true. That's a fair point.
EMP: So, it seems to me that the capacity market is designed to protect the most expensive resources on the system. And if we...
WH: No, it's designed to protect everything.
EMP: Okay. Well, if you have price-responsive demand, doesn't that take away the need for certain capacity resources on the grid? If you've actually got the customers responding to prices in the marketplace, as opposed to what you described in New Jersey, then then we can have a more efficient market.
WH: Right. And I'm all in favor it. But price-responsive demand and inclusive in the market to charge them for what they take and so on. And that's all terrific and the mandatory hedging of the service in New Jersey, only applies to the residential and small commercial. There's a very large volume, all the industrial and large commercial customers who are on the market, and they're going to have to make their own arrangements. And they could be the source of price-responsive demand. In principle, you could imagine going even further with the residential and commercial and we probably will eventually someday when we get all the technology in place and the institutions, but, so I'm all in favor of price responsive demand where I mean, I'm not saying to word demand response. And I'm not saying that on purpose, because I don't think that's the right formulation of the problem. So price-responsive demand where you charge them for what they take, as opposed to demand response where you pay them for what they didn’t take. The latter is rife with all kinds of problems that I've written about elsewhere.
EMP: We won't bring Jon Wellinghoff into this discussion. (laughter) Let's – we're running out of time here. So let's touch on the last topic that I wanted to bring up with you and that's the operation of state subsidies of specific resources – you have states that are subsidizing specific resources within that state, but they're part of an overall regional market. And that became highly controversial within this whole capacity market system that you're not supportive of. But you in your paper with Scott Harvey seemed to imply that with LMP you can incorporate state subsidies into that. So you're not necessarily opposed to state subsidies in the marketplace.
WH: Some subsidies are better than others. So you have to think about what the problem is that you're trying to deal with. So I'm not surprisingly a carbon tax advocate. So if the problem is carbon emissions and associated greenhouse gas emissions, then you should focus on that and tax them in order to internalize the cost. And then that flows into the cost of operating the different kinds of plants in very complicated ways that you don't have to control. It just works through the system naturally, it affects the dispatch, it affects which generations, so therefore, which generations run which loads participate it affects the locational prices and everything and all works right in that system and will have the effect of providing the most efficient way to reduce carbon emissions in the model that we already have. Subsidies, on the other side, where you're trying to say, okay, we're going to subsidize this renewable plant and that, you know, maybe you specify where it's going to be in my state, in the northern part of my state, but not whatever you're doing, are a different matter entirely. In the first place, they typically treat the system like it's a copper plate, right? So if I generate – if I have a renewable plant that I produce power and must be reducing production by, on average, maybe, or you know, at the margin or something, carbon emissions, and so there's, they tend to think about it as a kind of one-to-one story, but it's actually not true. So if you go, and it's for the same fundamental reasons. So PJM actually publishes now, I haven’t looked at it recently, but they started a couple years ago, where they would calculate every five minutes. So if you generate additional renewable energy at this location or megawatt hour, how much does it reduce carbon emissions on the grid? And the answer is the marginal effect and they calculate that number and you say, is that the same everywhere? And the answer's no. It's not. And you say is it different by a lot? And the answer is, well, for the extremes, you know, it's probably a factor of two or three. So in terms of the amount of carbon emissions affected by the renewables at that time. So the average of the margins can differ by quite a lot. And so this is very inefficient to go around subsidizing things when you don't actually know what you’re doing. And then you start characterizing, you think about the whole economy running that way. You know, we're, you're constantly trying to manage everything along this chain that's very complicated. And I think that's a hopeless task. There's some evidence of this. So one of the things that comes out of all these climate models and climate analysis, at least of the sort of Bill Nordhaus type, is that there's a social cost of carbon which is the same everywhere in the world. So that means that, in theory, if we were a global social planner, we should have the same price everywhere. At a minimum, we should have the same price everywhere in the United States. Well, at a minimum, we should have the same price everywhere in PJM. And so then you go and you look at what's actually done through all of these subsidies and the market monitor for PJM publishes a report every year about the market and they have a table in it and it says the implied social cost of carbon of this REC [renewable energy credit] subsidy program in this state and that state and solar and wind and so forth all across – they go through across all kinds of technologies and across locations. And you get prices which differ by a factor of, I don’t know about 50 or something like that, from high to low here. So it's not even close to being what the theoretically efficient solution would be and trying to modify all of these subsidies so that the implied price at the margin or the subsidy is the same across all of these regions is a well, I think it's hopeless. So what I would do is not start there. I would start with the problem, and the problem is emissions and I would charge for emissions. Thank you very much. Now we have this problem goes away. Right? Because the implied price for all these different things will then be whatever you said is the social cost of carbon and the carbon tax. That's what it will be. And that's what we want. So I think it's a mistake to go the subsidy route. And the carbon price solution is – it'll have its own problems. Nothing works perfectly. But I think it's much more – I mean it almost goes back to you want a market economy or you want a centrally planned economy. And I'm a markets guy.
EMP: So in short, you're saying that if you've got a social price of carbon then you don't need the subsidies.
WH: Right.
EMP: Is a tax better or should we have a cap-and-trade program for the emissions?
WH: Well, in the classroom, the cap-and-trade and taxes look very similar and in fact theoretically could be made identical. It’s just a question who gets the money? So that's in terms of the price of the efficiency and the technology choices and so, in practice, we have an entirely different situation. And what we find is, in the theory of cap and trade is, well, we can, we can pass out the ownership of these rights. And that's a way of getting people to accept it and we'll end up with a good outcome that will be similar to a tax but it makes it more politically acceptable and politically feasible. And I was very receptive to this argument many years ago when we were first looking into it in this country. And we saw it with sulfur and it worked pretty well. And then we haven't been able to get it to work very well for carbon. But what do we know about the cap-and-trade so far? Well, what happens – I mean California is a good example – is that if you don't get exactly the right amount of permits, then the prices can be highly volatile. And we don't like that. So then we start trying to monkey around with how many permits there are going to be in order to manage the price. And then there’s a lower bound? There's an upper bound and there's, you know, we have all these euphemisms for it. But basically what we're doing is coming in the backdoor and creating a carbon tax because what we're targeting is the price, not the volume of permits. And that's good because we have a lot more understanding of what the appropriate social cost of carbon is than we do the volume of permits.
EMP: So the tax is a cleaner, more elegant solution.
WH: Right. And that's and it's turns out with to the extent that they're different, the cap-and-trade and the tax, the tax is better.
EMP: I don't know if there's anything else you want to bring up for the good of the order before we sign off?
WH: Well, no, I think we've covered a lot of material and the – we didn't talk about operating reserves demand curves, one of my favorite solutions.
EMP: That's another podcast.
WH: Right. So but I'd be happy to discuss further and glad you're doing this.
EMP: Well, I appreciate your generous allocation of your time and hopefully we will have an opportunity to do this again.
WH: Thank you.
EMP: William Hogan, Harvard University, Kennedy School of Government.
# # #