How Electricity Markets Work

You are listening to Shift Key. He maps weekly podcasts about decarbonization and the shift away from fossil fuels. On this week's show, we are back to summer school and we are talking about the market that controls your power system. Why do some power plants run instead of others? Why do we use markets on the power grid in the first place and why power prices move across North America like weather systems? It's all coming up on Shift Key after this. Want to accelerate your career in clean energy? Then explore online certificate programs from our sponsor, the Yale Center for Business and the Environment. Whether you're designing policy, unlocking finance or developing impactful projects, Yale's online clean energy programs equip you with tangible skills and powerful networks and you can continue working while learning. In just five hours a week, propel your career forward and make a difference. Listeners of Shift Key and Shift Key Summer School can now get an exclusive offer. Use the referral code heatmap25 and get your application in by the priority deadline to get $500 off tuition to one of Yale's online certificate programs in clean energy. It's time to transform your career, your community, and the clean energy landscape. Head to CBEY.yale.edu to learn more. Join clean energy leaders at RE+25 from September 8th to the 11th in Las Vegas. As North America's largest event for the modern energy industry, RE+ brings together innovators across solar, energy storage, EVs, hydrogen, wind, and more. It's where industry leaders tackle today's challenges to shape tomorrow's energy solutions. RE+ is the epicenter of clean energy leadership uniting the industry to meet rising global energy demand. And listeners can save 20% with code [email protected]. [MUSIC PLAYING] Hi, I'm Robert Semyar, the founding executive editor of HeatMap News. And I'm Jesse Jenkins, a professor of energy systems engineering at Princeton University. And you are listening to Shift Key, HeatMap's weekly podcast on decarbonization and the shift away from fossil fuels. Well, today on Shift Key, we are resuming our summer series on the basics of the electricity grid. Everything from what is the difference between energy and power to what are some of the fundamental electricity units? And how should you think about them? To the history and engineering background, I'm never quite sure what to call it, Jesse, of the big thermal power plants on the grid from coal to natural gas. And then the history and engineering of our new sources of electricity, wind, solar, geothermal batteries, all of that. So today, we're going to move from some of the basics of electricity and how it works, to how some of the institutions and systems we build up around the energy work. In the next few weeks, we're going to talk to someone who actively manages a power grid every day to get into the basics of how the grid itself is a vast and humming machine operates and is managed on a minute-to-minute basis. But today, we're going to talk about the institution or the system that is behind the operation of the plan, that controls the money and the finances but behind the day-to-day operation of a grid. Today, we are talking about how electricity markets work. And my friend and co-host, Jesse Jenkins, is going to introduce this topic for us. He's going to talk to us about it. Jesse, I'm very impressed here. I know you have your lecture slides, and I know that I'm allowed to raise my hand and ask questions, but to some degree, I'm impressed that we can just do how do electricity markets work. And you're going to talk to us about it more or less off the dome. Yeah, well, we'll see how it goes. So let's start. Professor Jenkins, I want to think about electricity markets. Where should I start? Yeah, so let's start with this sort of basic markets piece of it, and then the reason I find electricity so fascinating as a sector, and why the course that I teach at Princeton is on the engineering economics and regulation of the electricity sector, is that the way we create markets and institutions to run power systems and keep the lights on and make sure we have affordable and reliable energy has to be this sort of unique blend of the physics of power plants and power transmission and all the detailed, complicated speed of light transmission of power and the kind of classic economics concepts that you might apply to create say a market for oil barrels in the global commodities market or pork bellies or whatever other commodity you have out there. And so we've gone over the last few episodes, some of the engineering basics. It's useful now to think about some of the market basics. And then we can put those back together to talk about how exactly electricity markets are structured and why that's different from other markets that you might be more familiar with. - Yes Rob. - First question. - It seems like there is another kind of bridge here between the engineering basics and the market basics, which is that when we think about power plants and the traditional mold, whether they're giant coal plants plopped in the middle of cities or the natural gas plants that we were talking about invented in the 1960s, we're talking about technologies that emerged in a utility model and a monopoly utility model, right? When single companies would build out the power grid in these relatively geographically concentrated areas and they would control everything from generation to transmission to distribution. A model that to be clear still persists across a lot of the U.S. today, including especially in the southeast. So when we think about electricity markets too, we're also partially thinking about them as historical objects, right? As institutions that developed in response to challenges and crises that emerged in the electricity sector during the 1970s and 1980s. Is that wrong? - Yeah, it's interesting though, the electricity markets are at least power pools that allowed for buying and selling a power between those utilities actually goes all the way back into the 19 teens or 20s. I think PJM, which we've talked about a few times in the show, the original Pennsylvania Jersey, Maryland interconnection actually goes all the way back to the 1920s when southeastern Pennsylvania utility now, PICO and our utility here in New Jersey now, PSC and G. Realizing that here you are across the Delaware River from one another with your own little territories and a bunch of your own power plants. And sometimes PSCG has a pretty inexpensive power plant sitting there while PICO's running a more expensive one. And their power grids are pretty well connected 'cause they're dense areas, cities right next to each other. Wouldn't it make sense for PICO to turn to PSCG's power station across the river instead and save some money? And so they set up these power pools where they started to operate their generators in a kind of combined pool to find the least cost mix of resources. And then there was some settlement to kind of pay whoever needed to get paid what it cost to run those generators. So it actually goes way back, but at the basics, you can even think about this to start from as a single utility trying to minimize its cost of production to serve its customers. So let's assume you've got a bunch of power plants that are already built and we can come back to later about what drives that investment or retirement decision about whether or not to have a power plant there. But in the short term, I've got a bunch of fixed capital, I've got a bunch of power plants sitting around and let's simplify it by saying I just have maybe three. I've got a nuclear power plant, I've got a coal plant and I've got a natural gas plant. And throughout the day, say I can produce 1,000 megawatts of electricity at maximum with all of those power plants together. But my electricity demand, of course, varies every day and every hour and seasonally. And so at some point of the day in the middle of the night, maybe it's only a couple hundred megawatts and it only ever reaches close to that thousand megawatts three, four times a year when it's really hot out and people have their air conditioning on. And in the meantime, it fluctuates up and down between then. So let's assume that the nuclear plant has the lowest variable cost, that's typically the case. It uses uranium fuel rods and those are extremely energy dense. You know, they've got like seven orders of magnitude, more energy density than a hunk of coal. And so you can generate a lot of electricity at very little cost per megawatt hour from a nuclear power plant. Even though the capital cost of building it is expensive, it's short on marginal cost or fuel cost is very low. The coal plant is sits in the middle but it's got a slightly lower capital cost than the nuclear plant and a slightly higher fuel cost. And then let's assume the natural gas plant is the most expensive one to operate with the highest fuel cost. Historically, that's been the case, although in recent years as natural gas has gotten cheaper, sometimes older, less efficient coal plants are more expensive than natural gas plants. But that's the typical mix of thermal power plants. Nuclear coal gas in order from cheapest to most expensive. So what do I want to do if I'm in utility trying to minimize cost as demand goes up and down? Well, you can basically think about it as I'm just going to start turning on the cheapest power plant first until I meet my demand. So if demand is only 200 megawatts or less, all I need is my nuclear plant. If it's between 200 and 450 megawatts maybe, then I need to turn on my coal plant. And then above that, I also have to run my gas plant, something like that. And so as demand goes up or down, I'm basically turning on and off generators to keep the most inexpensive ones from a fuel or variable cost perspective running first and most often. And so that's what leads to this concept of base load generation, by the way. It's not a feature of a power plant. it's actually a feature of the demand or in electricity. The electricity parlance, the load, for some reason we've decided to call electricity demand load, just to make things more engineering-y. But basically, the idea of a base load is that's the level of demand below which your demand never falls. So you know, at the lowest point of the year, your demand is a couple hundred megawatts. That's your base load. And then it makes sense to have a power plant like a nuclear plant that runs basically all the time to meet that demand because it's the most inexpensive generator. And then you kind of layer other generators on top of that that are a little bit more expensive that are designed to meet that intermediate load and that peak load. You have the sine wave of electricity demand through the day, right? Yeah, it's a typical, yeah, it typically looks like that, like some kind of wave. You know, it rises during some kind of shape, right? People get up in the morning, they start making coffee or whatever. Then there's a little, you often a little morning surge and then they go to work and people start turning on their conditioning or the heating and the winter and the demand starts to build over the course of the day, in the summer, it typically peaks in the afternoon on a hot day when, you know, or in the evening, even five, six p.m., when people have just gotten home and are cranking their own air conditioners. But the office is still open. There's still people there. There's people working late or people cleaning and people on the factory shift still. And so they're still, like you get the combined maximum demand from both commercial and industrial and residential sectors. If it's an electrically heated region, then it might be the middle of the night on a cold winter night instead. So that's the kind of basic, kind of engineering economics of it. So if I'm just a utility operating on my own, I want to basically run my fleet on what we call economic dispatch, which is rank ordering them from cheapest to most expensive on a fuel or a variable cost basis and trying to maximize my use of the less expensive generators and only turn on the more expensive generators when I need them. That introduces this idea of a marginal generator, where the marginal generator is the last one I turned on that has some slack to move up or down as demand changes. And what that means is that if I have one more megawatt hour of demand in that hour or over the five minute period or whatever, or one megawatt hour less, then I'm going to crank that one generator up or down. And so the marginal cost of that megawatt hour of demand is the variable cost of that marginal generator. So if it's a gas plant that can turn up or down, say it's $40 a megawatt hour to pay for its fuel, the cost on the margin of me turning on my lights and consuming a little bit more is that that one power plant is going to ramp its power up a little bit or down if I turn something off. And so the way we kind of identify what the marginal value of supplying a little bit more electricity or consuming a little bit more electricity is the variable cost of that last generator, not the average cost of all the generators that are operating because that's the one that would change if I were to increase or decrease my output. Does that make any sense? It does. In other words, the marginal cost for the whole system is a property of the power plant on the margin, which I realize is topological, but basically the marginal cost for increasing output for the entire system by one megawatt hour is actually a property of the one plant that you would turn on to produce that megawatt hour. That's right. Exactly. And that can change over the course of the day. So if demand's really high, that might be my gas plant that's on the margin, but if demand is low or in the middle of the day, that gas plant might be off and the marginal generator during those periods might be the coal plant or even the nuclear plant at the bottom of the supply curve. So that kind of general concept of how I would operate things as an individual utility can expand to think about how I might pool a whole bunch of power plants together owned by different people across a region, going all the way back again to this idea of Southeast Pennsylvania and New Jersey, utilities trading power across the border. Eventually, that expands into a much bigger pool. The New England utility started doing this not too long after, maybe in the 50s, or cut in Texas started doing this, right, is the Texas grid evolved. This was sort of a function of the fact that power grids started out as basically centered around individual cities, run by individual monopoly utilities. And then as the grids got bigger and bigger and they started to become more strongly interconnected with one another, then it made sense to potentially pool and share or buy and sell power across the borders of those individual utilities. And so that's the antecedent to today's modern electricity markets in what we would call competitive or restructured markets, where we've actually stripped that generation function usually away from the traditional monopoly utility entirely. And we have independent competitive generation companies that can operate kind of anywhere within that region. They might own one power plant or several. They could have a sister company somewhere else across the world or the country or not. It doesn't really matter. They might even be part of the same parent company as one of the transmission utilities, but the generation companies are supposed to operate independently from that. And so if you want to organize all those different generators together, what principles can we use to make sure that an electricity market yields the same outcome as that kind of omniscient utility would if it were controlling everything. And the trick is to basically continue to follow that concept of economic dispatch of running the least expensive generator that we can first. And then the next one and then the next one and so on until we meet the demand across now the whole region instead of each individual utilities region. Rob? I have a question. When we were talking about thermal power plants in the second lesson of 15 summer school, one of the properties, or maybe it was the first lesson, one of the properties we talked about is that when you turn the lights on in your home or you plug in your eB or you start an appliance or you turn on your air conditioner, there is a plant somewhere that has to rev up, it's turbine spins just a little bit faster, all the turbines in the system spin just a little bit faster to meet your demand. Is that captured by this production function or are those, no, is there like a certain level of tiny fluctuations in demand that don't actually get captured in this minute to minute production basis? Yeah. So that's a great question. No, that kind of physical inertial response doesn't get captured and that's one of the ways in which power systems are different than any other commodity market, because what I just described to you of like lining up a supply curve from the cheapest to the most expensive, that's how oil markets work, right? The global oil price is set by the marginal producer somewhere in the world, not the average of all of the producers out there, not the cheapest producer. So if I'm Saudi Arabia and I can produce, you know, oil at $10 or $20 a barrel, if the going price is 60, because that's what it would take to get the marginal producer to produce one more. I'm going to sell mine for 62, because if you don't buy it for me, you got to buy it from that guy. And so we see this in all kinds of markets, markets for corn, markets for copper, markets for oil, where the lower cost producers effectively make more money because we have a uniform clearing price for that commodity market that reflects the cost of the marginal producer. What you're pointing out is that actually on a physics basis in a very short time frame, every generator that synchronized to the grid actually contributes a little bit to that marginal supply. And so that's different than oil markets. There's no physical response like that, because the oil market isn't synchronized at the speed of light the way that the power grid is. So that actually, that inertial response and even some of the frequency regulation and reserve products that act very quickly, those are priced separately outside of what we would call the traditional electricity market. Those are called ancillary services. And in the case of inertia, they're typically not paid for. It's just something that all the physical generators, the thermal generators are generators with synchronized turbines will provide as a basis of physics and it's been an ample supply. And so we haven't really paid anybody for it. Things like the ability to track the frequency regulation signal that's required to correct for small little adjustments. We pay for that in a separate product called frequency regulation or something like that. There are lots of different names for it in different markets. And then we even have a set of backup generators who are kind of there holding on short of their maximum production in order to potentially step in in case something goes wrong. And if the generators that we're dispatching as part of the economic dispatch, say if the nuclear plant fails or the gas plant has its transmission line go down, you need to have be able to instantaneously or close instantaneously ramp in new generation to make up for that. And so we also have to pay some generators that wouldn't otherwise either turn on or would turn on but produce at their maximum a standby fee effectively to standby with spare capacity to ramp and have it online or quickly able to turn on to ramp up in the case of contingencies or emergencies. Those are called reserve products typically or ancillary services. All of that sort of unique to the power market and outside of the core electricity markets. And all of that is critical to keep the grid running, but it isn't a lot of money. The small fraction of what we pay in our electricity bill is for all of those ancillary services. So they're really important per unit of service provided. They're really valuable so that people doing them are very happy to do them because they get paid well. But they're not very big markets because you just don't need a lot of that. What you need is a ton of energy to meet all of the demand of the aggregate region. And that's what the general electricity market is designed to supply. So this concept of a supply curve and then a marginal generator is the sort of basics of how we organize competitive electricity markets where as demand goes up or down at differing increments in the case of the U.S., the real-time markets are cleared in five minute increments. Basically every five minutes generators are dispatched up or down to meet the demand and whatever the marginal cost of the last generator that's needed to meet demand sets the market. price. Now, I should say accept that the transmission grid makes all this a lot more complicated and we'll come back to that after we've got this concept down because it means there often isn't just one marginal generator. When the grid is congested, we end up with more than one. But let's put this grid aside for now and assume that the grid is like a copper sheet, we can instantly transfer power from any point in the market to any other. Then what would set the marginal price for everybody, the market price would be the marginal cost of the last generator to dispatch. Everyone gets paid that same uniform clearing price and that might sound dumb, like why am I overpaying for those cheaper generators that could produce for less than that? I'm paying $60 per megawatt hour because that's what it took to pay for the marginal generator, but I got all these other ones around who might be a lot cheaper. It might be $10 or $20 or $30. Well, the reason we do that is twofold. One, it's what we call incentive compatible. As long as I'm paid exactly what it costs me to generate or more, I'm perfectly happy to turn on and run my power plant. There's an economic incentive for all of these competitive individually owned generators to follow the least cost system dispatch if we clear it that way. If we don't do it that way, if we do a pay as bid market, then each generator basically has to guess how high they can bid before they're not called on anymore. If the marginal generator is $60 a megawatt hour and I'm below that, I should be bidding more. I have some market power. I can bid up until I'm $60, but what if I get it wrong and I bid $62? Instead of picking me in the market, you pick that more expensive generator. Instead of me, now the cost of the market has gone up because you've picked a more expensive generator. That's like this pay as bid market idea is not incentive compatible because it encourages me not to tell you as a generator, not to tell the market or the system operator what my actual cost of production is. Whereas if you're paid as cleared, I don't really have an incentive to tell you I cost something different unless I think I'm the marginal generator that's setting the price. As I can tell you, I cost 20 bucks and as long as the market is large enough that I'm not the marginal generator, I'm still going to get paid at least 20 bucks and oftentimes much more than that and that generates what we call "inframarginal rent" or "gross profit" that I can use to pay off my investors, to pay my fixed costs that are not reflected in my fuel costs, property taxes, staffing, maintenance, things like that. And so that inframarginal rent is actually really important source of revenue for these expensive generators that have high fixed costs. Do you want to build critical skills that are transforming the clean energy sector? Then discover the Yale Clean and Equitable Energy Development Online Certificate Program from our sponsor, the Yale Center for Business and the Environment. This fully online five-month program is designed for working professionals who want to build projects that serve both people and the planet. You'll dive into energy fundamentals, project finance, and key phases of development and operations while learning how to center community engagement. In just five hours a week, you'll learn from leading experts, develop practical skills, and gain actionable insights. You'll engage in video lectures, live sessions, and a real-world group project and Yale will ensure that you have the expertise to advance clean energy solutions that benefit communities and drive economic resilience. Listeners of Shift Key and Shift Key Summer School right now can get an exclusive offer on these programs. Use the referral code to keep map 25 and get your application in by the priority deadline to get $500 off tuition to one of Yale's online certificate programs in clean energy. So transform your career, your community, and the clean energy landscape. Head to cvey.yl.edu to learn more. It does seem like there's an important point here for topics that we discuss on Shift Key all the time, which are number one, this is why solar and wind, especially solar, I think which so clearly has a zero fuel cost, can be quite profitable, especially in grids that don't have a lot of solar on it, because at the middle of the day when it's just cranking out electrons, it's still making whatever the marginal cost of production is. That's right. Yeah, and so they basically come in, so if you think about renewables with zero marginal cost, they basically come in at the bottom of that supply curve and they push the rest of the supply curve out further. So if you stack them up horizontally with the capacities on the bottom and the marginal cost on the y-axis, you're adding something at the very bottom and the bottom left of that chart and you're pushing the supply curve further out. So that does lower the marginal clearing price, which is partly why we talk about the value decline that wind and solar experience because the more they produce, the more they're pushing the more expensive generators off the market and then they're getting lower and lower clearing prices. Unless a winder solar plant is the one on the margin because we're curtailing winder solar and we could increase its production, which does happen sometimes now in markets where we have an oversupply of winder solar at certain periods, then the market price is going to be above zero dollars because it's going to be set by some thermal power plant, some colder gas or nuclear plant. And then that extra money that I'm making there, that's what pays for all the fixed costs of that wind farm or that solar farm. And so as long as I make enough money from inframarginal rents, from those periods when I'm not the one on the margin to cover my costs, then it's a good investment. And that kind of links that idea of expected inframarginal revenue or expected gross profits that I earn when I'm generating at below the marginal price. Those expected revenues are what in theory are supposed to support investment in new capacity. As long as I think I'm going to make enough of that money to cover my investment and return my profit margin, I'm going to build another power plant. And if I don't, I'm going to not build another power plant. And that's what sort of, if in theory sends the long-term signal that we need to drive investment and retirement decisions as well. So then that links to this idea of, you know, you might have heard this when I was describing it, is the plants with the lowest variable cost tend to be the ones with the highest fixed costs. There tends to be this inverse relationship between fuel cost and efficiency and capital cost. So a wind or solar farm is 100% capital cost effectively. It has no fuel cost, very little variable operation and maintenance, especially for solar, maybe tiny little bit for wind. And so it's the bottom of that short run supply curve, but it's maybe the most expensive per megawatt or has a very high capital intensity per megawatt. Then you've got say nuclear or hydro, they're a little bit higher variable cost in the case of nuclear, also high capital cost. Then coal and combined cycle natural gas plants kind of sit in the middle where they have a modest fixed costs to build and medium or intermediate fuel costs because they're relatively efficient in the case of a combined cycle plant or they burn relatively cheap fuel in the case of a coal plant. And then the kind of high end of the curve tends to be less efficient, open cycle or combustion turbine gas plants or even steam driven gas plants like coal plants that have been converted to run on gas and diesel generators reciprocating engines and things like that. And those have much lower fixed cost per megawatt, but are less efficient in case of diesel burn more expensive fuel. And so they're the high end of that curve. And so you have this inverse relationship between fixed cost and variable cost. And that ends up working out really well for the short and long run picture because in the short run, the plants with the lowest variable cost operate more because they're the ones that gets dispatched first. And they tend to earn a larger informational rent or revenue because they're below the information. They're not on the margin, right? That's what informational means. They're below the margin. And so they earn a lot of that revenue on a day-to-day basis in the energy markets. And therefore they can afford to have a high fixed cost that is paid off by all that revenue. Whereas if, yeah, whereas if I'm a peeker, say like a really, you know, I'm a diesel generator that only turns on 10 hours a year. And I'm close to the marginal cost because I'm one of those expensive generators that turns on. I might not make a ton of money from my energy sales through the year. And so in order to justify that investment, I have to be really cheap on a capital cost basis as well because I don't earn a ton of money in the energy markets in excess of my variable cost. And I don't operate that often and I'm pyer up the supply curve. So my informational rents when I do operate are lower. And so this kind of inverse relationship between fixed and variable costs and then the way the short run market works by ordering only based on variable cost actually works out where it equilibrates in theory anyway. There's risk and there's uncertainty and there's lots of complicated factors on top of that. But in theory, the sort of short run and long run decisions are linked through this idea of informational rent that you earn as part of this supply curve with a single uniform clearing price. It's reminding me of trains in 19th century America. I realize it's a ludicrous but or I think trains in all economic contexts actually meet these requirements. But I'm thinking I know trains in 19th century America the best is that like 60 or 70% of the cost of basically running a train line was not capex but was like locked in. If you were running a train it didn't matter how full you made the train that was going to cover once you decided to run send a train down the tracks you were spending 50% of the cost exactly. Yeah, I fixed costs. And so the train companies came up with all these mechanisms to try to basically have as high a capacity factor as possible because their profit It depended on running that equipment absolutely to the max as much as possible. as they could. It's actually quite analogous here where you want to run the power plants with a relatively high fixed cost. They need to run all the time, and because they have low variable cost, they're incentivized to. Incentivized to you by this market design. In theory. Yes, exactly. In this market design. And then, so if you think about that, if that makes sense, this sort of idea of a supply curve, hopefully you have it in your head, this sort of stack of generators getting more and more expensive, you can also, to be clear, it's like wind and solar down at the corner, right? Yep, exactly. And then, and then, and then, and then clear, wind and solar hydro at the bottom, then nuclear, then combined cycles are cold, depending on which is cheaper in your region, then open cycle gas turbines, then diesel and oil fire, and then at the top, it's like, and then demander, right, demanders, and also like, the generator that's sitting next to a hospital that is meant to be built. Yeah, and which is actually often, yeah, which is actually often bid through a demand response program. So a lot of what we call demand response, meaning reductions in electricity consumption are actually self generation. So there are people turning on a generator at their hospital or their factory, and kind of reducing what they're drawing from the grid instead. So it looks like a reduction in demand from the grid, but it's actually being met by a very expensive smaller, inefficient generator. But because that generator already exists for backup power purposes, I don't have to justify all of its cost. Based on those few hours of operation, I just get sort of extra revenue that way to lower the cost of my back. And that's also why local air pollution gets so much worse on the absolute holidays. Yes, that's right. Yep, because we're running all those things. Plus a bunch of bunch, plus a bunch of those peaker plants, the most expensive variable cost ones. Again, they have to have the least lowest fixed costs. So they're run on those only on those really hot days. That means they have to be pretty inexpensive, which means on terms of capital cost, which means they can't afford to invest in expensive air quality controls, like air pollution control mechanisms, the way a coal plant that runs 70% of the year might be able to justify investing in sulfur scrubbers or things like that. And so they also tend to be very polluting because of that, because they can't support the economically justify the sort of high fixed cost of pollution controls when you don't run them very often. So yeah, all that contributes to pretty nasty air pollution during peak demand periods. So if you have this supply curve in your head, you can think about two, I think, really important dynamics that have been going on in electricity markets lately. One is the role of natural gas prices in affecting electricity prices. So whenever a natural gas plant is the marginal generator, what determines its price is basically the price of fuel times its efficiency. And so if the price of fuel doubles, then the marginal cost of that generator also doubles. And so if you have a whole bunch of gas generators in your supply curve, that's basically making your supply curve steeper when the price of gas goes up and flatter when the price of gas goes down. And that directly translates to in the case of higher gas prices, higher and more volatile electricity prices because as demand goes back and forth as horizontally on that picture, it's intersecting with a steeper supply curve. So the price is moving up and down vertically on the supply curve chart much more for each megawatt of changing demand. Whereas if gas prices are really cheap, then there's a lot less incremental difference between one gas generator and the next. The supply curve gets flat and the variability in wholesale prices or electricity prices is less volatile and lower when gas prices are low. And so this is a feature we've seen play out multiple times in different cycles throughout the last 15, 20 years of electricity markets in the US and Europe is natural gas prices are pretty volatile, relatively speaking. They go from couple bucks to eight bucks back down to three, you know, all over the place. And those drive very substantial changes in the supply curve of the electricity market. And so electricity prices tend to be quite tied to underlying gas prices partly again because the gas generators often set the marginal price just because of the position they play in our supply curve. Now generating over 35% of US electricity nationally and higher shares in certain regions. And the fact that they have this intermediate fuel cost that sits them right on the margin where the demand often falls. And so they're most often setting the marginal price. And so our electricity markets become even more sensitive to the price of gas. Then if you were to just look at the average cost of generation across the fleet, which might be less sensitive. And then the last thing I think we've just we've hinted at before is the role of renewables in driving down prices. So whatever that supply curve looks like, however steep it is because of the price of gas or not, renewables are when they're generating like their variables. So throughout the day they're there and then they're not right. And so they're also moving the supply curve out and back from the origin whenever they're around. So if you have a big surge of solar production in the middle of the day, you can think of it as either shifting the supply curve way out or it's the same effect actually taking out some demand because they're the first ones to meet the demand at the lowest because they have zero marginal cost. So it also has the same effect of shifting the net demand that all those thermal power plants have to meet in towards the origin by subtracting away the solar or wind. And that's exactly what that infamous duck curve that some of our listeners are probably having in their mind from California of the net demand for electricity as solar ramps up in the middle of the day, which creates this huge kind of belly of the duck in the middle of the day. That's what that's depicting is that net demand because that net demand is what determines how much fossil generation or thermal power plant generation. And increasingly now storage dispatch we need to meet the demand at any given time netting out the fact that the renewables are the first ones to always get consumed because they have zero marginal cost. And so renewable output throughout the day will actually shift the supply curve around at the same time that demand is shifting around. And they have the effect of driving down prices all else equal when renewable output is high. And effectively increasing prices when renewable output is low because the supply curve gets closer to the origin or the demand gets effective demand gets higher. This is easier with PowerPoint slides in my lecture notes, but hopefully you've got some visual. No, no, no, I think this makes I think the idea is very intuitive that basically like that solar sitting there, the sun shines on it, it's going to generate electricity. And so we're going to use that first. You're going to use that first. All of that goes unless you have so much of it as they know that you can't use it in California that you can't use it all. We can complicate this further with the physics in a moment, but before we do that, I wanted to just make sure you don't have any other questions about. I just I guess the one question I'd have is when you talk about the curve getting steeper, that's because it gets increasingly as gas is expensive on the spot market. Yeah. Yeah, so say I have a power plant that's 50 percent efficient. That's probably typical for a combined cycle plant somewhere in that range. And gas prices are $4 an MMB2. I don't know, I have to do a math in the head about how many BTUs you need for megawatt hour. My heat rate is seven. So I need four times seven. I need $28 worth of fuel per megawatt hour that I consume. Yeah. So the efficiency part's kind of hard here because as we talked about in our energy units class, all these things are denominated in different metrics and natural gas prices tend to be denominated in million British thermal units of heat content as opposed to per megawatt hour in Europe. You actually they they're smarter on all things units related. They use the metric system and they also denote their fuel prices often in megawatt hours. Maybe it'll be easier for me to just do that conversion. But say it costs I have a fixed efficiency of my power plant. And when gas prices are $4 an MMB2, my marginal cost for production is about $28. If gas prices double to $8, now my marginal cost for production also doubles. It's now $65, $55, right? $56. So now my marginal cost for production doubles as well. And it's now 56 dollars. So if you think about where that where I sit in the supply curve then, the height of my bar and that supply curve, the part that indicates my marginal cost for production is now twice as high. That's what drives the steepness of the curve. But it's not just one generator. It's all of the gas generators. And the ones that are higher up in the curve are less efficient. So they go up even more than the ones that are lower in the curve. So it just makes the whole thing more steep. You know, it increases more rapidly as a function of how much demand I have to serve. And obviously if if gas prices increase, they're going to have a bigger effect on the less efficient generators, cost of production than the more efficient ones. And so that's what yeah, tilts on the curve faster. Efficiency matters because it's purely a measurement of how well you can convert your fuel into megawatt. Into power, exactly. Yeah. So unfortunately, we have to talk about a heat rate, which is how much heat content of fuel I have to consume per unit of electricity. That's the inverse of the efficiency actually. And it's measured in MMB to use per megawatt hour because that helps you figure out how many MMB to use a fuel I have to burn to get a megawatt hour of electricity. But if you think about the efficiency of a power plant being 50 percent, then the cost of producing a megawatt hour is going to be the cost of two megawatt hours worth of fuel, right? Because I have to burn two megawatt hours of fuel to get one megawatt hour of electricity. If my efficiency is 33 percent, then now I have to burn three megawatt hours of fuel to get one megawatt hour of electricity. And so my marginal cost of production is going to be three times the fuel price. Join clean energy leaders at our EP. plus 25, happening from September 8th to the 11th in Las Vegas. As North America's largest and most comprehensive event for the modern energy industry, RE Plus brings together innovators and decision-makers across the full clean energy spectrum, including solar, energy storage, EV infrastructure, hydrogen, wind, micro grids, and more. RE Plus is where the industry's rightest minds gather to tackle today's most urgent challenges and shape the solutions of tomorrow. So come to the epicenter of clean energy leadership and drive collaboration and innovation to meet the world's rising energy demand. As a shift key listener, you'll save 20% on a full conference pass by using the code HeatMap20, register now at re-plus.com, that's re-plus.com. Okay, let's get into some physics of the grid, which is that the grid is not actually a big copper sheet, electrons can't zoom around from place to place and from power plant to user as they might in a kind of imagined market. So how do we reconcile the fact that the grid is actually a very concrete or maybe not even concrete, a very steel and copper set of structures that unite and link certain plates with other certain plates rather than a big copper sheet? So there's two things that we have to account for now when we think about the grid. We have to account for transmission losses. The fact that as you send power over power lines, we lose some of that electricity, primarily to heat from resistance losses or inductive losses in the lines. Transmission lines are very efficient, that's why we use them there. The losses on a transmission line go down in inverse proportion to the voltage squared. So this is why we have very high voltage transmission lines because if you double the voltage, you get one quarter of the losses. So if you have a very high voltage line, you have very, very small losses, typically on the order of like one percent of all power generated is lost in transmission lines. But on the margin of power lines are heavily loaded, power losses go up quadratically with power flow. And so when a power line is heavily loaded, the marginal loss could be relatively high. Maybe it's 5, 6, 7, 8, even 10% on a lower voltage line. So we have to account for that. If I'm trying to consume one more megawatt hour of electricity, the marginal cost of supplying me at my location has to account for which generator is going to produce that megawatt hour and then how much is going to get lost between where that generator is and where I am. Because if we lose 10%, then if I want a megawatt hour, you got to generate 1.1 megawatt hours. And so I'm going to have to pay for that in my marginal cost of production. So that's the first dynamic it's relatively straightforward. We have similar things in many markets for trade, right? You got to pay the fuel cost to power your ship, to get your ship from port to port. You got to pay the oil price for your, you know, you got to pay for diesel for your trucks. You got to pay for compression losses in a natural gas pipeline, right? Because compressors consume gas to drive the gas pressure, you know, that's typical. That is less typical is how congestion or constraints on power flows work in the powered system. If I'm driving my car down the road and I hit a traffic jam, Google or Waze or whatever can tell me, okay, get off at the next exit and go around the traffic jam, right? And so the road system has an additive capacity of all of the parallel paths to get from here to there. Because at the maximum, we everybody could spread out across all these parallel routes and get from point A to point B. Sometimes it's just one bridge or two bridges in a tunnel to get into Manhattan and you get a lot of congestion. But in most cases, there's an alternative route. And even in that case of the Manhattan bridges, they're all, they can all be operating at maximum capacity simultaneously. You drive up to the tap and see, yeah, or you drive to the tap, see whatever you do. You take an account, the added distance to get up there, but yeah, you can go around it. That is completely different in the power sector. Because electrons or I should say electrical power that transmits along lines doesn't flow in accordance to choice. You're not directing them down lines. They flow based on the physics, kind of like water flowing downhill with gravity. The electrical energy follows the path of least resistance or in the case of alternating current lines impedance, which is the alternating current equivalent of resistance in a direct current line. So impedance is just how hard it is for power to flow down an alternating current conductor. And if you get to an electron or an electrical current gets to a fork in the road, a split down two parallel paths in the power line, it will split an inverse proportion to the impedance of each line. Meaning it'll flow. If I have one line that's twice as easy to flow down than the other, then two thirds of the power will automatically go down that easier to flow line and one third will go down the other one. It's not a choice. It's a question of physics and the relative impedance of the lines, which are basically fixed properties of the conductor and how the lines are laid out and all that. And so that's the equivalent. And so what happens is whichever one of those two lines gets congested first, meaning which everyone gets close to its maximum operating capacity, that will constrain flows on both lines because I can't say the first line is at 100 megawatts and that's its limit. And I try to send another megawatt down the parallel paths. Two thirds of it will go down the first line and one third of it will still try to go down the second constrained line and it will hit that maximum and it will overload that line. And then that line will fail. And then all of that power that was going down the line will immediately shift to go back down the first line, which is now going to be over its capacity and we have a cascading blackout. So we can't let that happen. What happens when the power line fails? Power instantaneously rerouts along parallel paths. Oh no, I mean physically the transformer blows up. Oh, it could be a variety of things. Yeah, it could be the conductor sags too much because it's too hot and it's too old and so it hits a tree below it and starts a fire. Like it's happening in some case in California or a tree falls on it or a windstorm breaks something and it falls down or a transformer somewhere along the path overloads. Or just we try to send too much power down a line and it physically anneals. It falls apart because the heat generated by the current flow destroys the conductor. There's lots of different ways it can fail. We never want that and we never want to be anywhere close to that happening. So we operate the grid in a very conservative way called N minus one contingency usually or N minus one security constraints where the grid should be resilient to any one transmission line failing at any time. So the power flow that we're that the system operator is going to allow has to be simultaneously feasible across all of these possible contingencies where they take one power line out at a time because they don't ever have time to react in the case of an actual failure. The only thing that can react are automatic switches switch gear. We have these protections that will isolate lines when they detect a surge so that it doesn't get overloaded. But they won't have a choice to reroute power from line A to line B. If line A fails, it will immediately try to flow down line B. And the only thing stopping that is a switch that operates at the speed of physics right at the speed of light. The system operator doesn't get any say in the matter. Backing up though, what this means is that think about that example of driving into New York City. You've got multiple bridges and tunnels. What that means is that the first one of those routes that gets to its maximum say the Holland tunnel gets fully congested. No one else can go into Manhattan period, right? That's it. Nobody can go over any of the other bridges or tunnels. That's the effect that a constrained transmission line has on power flows across all of the other parallel paths in the network between point A and point B. And that might not just be two lines. It might be dozens of lines across a hundred mile front in the power system, all of which form these parallel loops because the power grid is often a meshed network. And so there can often be very lots of different parallel paths to get from point A to point B. And again, whichever one of those paths gets constrained first will limit all of the rest of them to send more power. So that's what we call grid congestion. It's not just a phenomenon of a single line unless it's a radial network or that's all that's connecting point A and point B when it's a mesh network which most of our grid is at the transmission level. You get these big fronts of congestion that will segment the market into different chunks. On one side of the congestion, we might have a cheaper generator on the upstream side that could be supplying that next marginal megawatt hour of demand. But it can't get through that congested front to the city where the demand is. In other words, these big price shocks move through the mesh node like a weather system. Yeah. In some ways, it looks a lot like that. So what results from all this is what's known as a locational marginal prices, where instead of a single clearing price for everything, every price, every location of the grid has a different price. But each one reflects the marginal cost of supplying power at that location, which has to account for losses and whether or not the generator is on the right side of the congestion and not to meet my demand. And if you look at these maps, these heat maps that the system operators or markets produce of locational marginal prices, they have to do look like a weather map with these big fronts. Arm-front or a cold front moving across the country. You see these big bands where on one side of the band, like frequently is the case in Iowa or whatever with lots of wind power. The price might be zero or even very close to that because you've got lots of wind and maybe even your curtailing wind power and the price is zero. And then right on the other side of that constraint in say Chicago, the price might be $500 a megawatt hour. because you're firing up backup generators or running demand response to meet the demand in that location, and you can't move another electron from where it's cheap and abundant to where it's needed because the grid can't do it. - Is it wrong viewers can't see me. I'm kind of looking, I'm making visible thinking faces. It's actually kind of a similar thing because it's price pressure. Like the reason fronts develop in the weather system is because of differential pressure resulting from the Coriolis effect, and there's more air in some place. And what's actually happening is the exact same phenomenon but with electrons in a fixed environment. - Yeah, it's a good analogy, I think. Yeah, that sort of makes sense. Yeah, and again, it's all connected at the speed of light because it's all a synchronized grid. And so these things can develop very quickly and they come and go very quickly. And so this leads to a lot of volatility well in the locational marginal prices that's larger than what you would get if you had a copper sheet grid. And places that tend to have lots of demand or places that are hanging on to the edge of the grid like say Long Island or Cape Cod or other peninsula that are sort of on the edge of a much larger grid. Those tend to be the places that experience that congestion most often and have higher prices because of it. Because they're basically constrained off from accessing cheaper generators that could flex up or down to meet their demand, but they're too far away on the other side of a congested path. And so they can't actually send their power from here to there. And this is one of the values of transmission expansion is it allows us to overcome those congestions and access a much broader market for cheaper power and to equilibrate the prices so that we don't get these big price diversions is that occur when congestion occurs and splits the market into two halves. And so whenever there's a congestion in the market we get basically two, we get, if there are end congestions, you get n plus one marginal generators in the market because each of that little zone of the market has one generator that's able to flex up and down and the market price reflects that different marginal cost in each place. It's a little more complicated because sometimes you actually have multiple generators that have to change because of those complex power flows as well like in order to get one megawatt down this path you actually have to turn a power plant down a little bit so you can move another megawatt over there to avoid these constraints. So you can get very complicated and weird locational marginal prices that only really make sense if you trace everything very carefully in physics. And in fact when I try to teach this and others do like once you get beyond like a three node network with like a triangle like network with parallel paths it's almost impossible to like actually make sense of what's driving these locational marginal prices. And so we use complicated optimization software actually to do this. This is something my advisor Ignacio Perez-Arriaga actually worked on pioneered after he earned his PhD at MIT was that you could use the same engineering software that was used to operate a least cost power system meaning I'm gonna try to meet all the demand in the grid subject to the constraints of all the physics of the power lines while minimizing the total cost of short run generation. And if you run that thing, you actually the marginal prices at every location that you have a demand constraint falls right out of that optimization software. It's called the dual value or the shadow price of each constraint which basically tells you if I were to change my demand constraint by one unit what would the effect be on total system cost? And that's exactly what the locational marginal price means. It means if I were to consume one more megawatt at any location, what would its total effect be on the cost of generation across the system accounting for all the losses, all the complicated power flows and all the different possible generators that could be affected by that. And so fortunately we don't need to actually know exactly which generator is moving. We just need all their bids to tell us what their marginal cost is. We need to know where they are. And then the system operator or market operator also needs a physics model of the power flows on the grid in order to clear that market. And that's what our regional transmission organizations or ISOs do. And by doing that, they're simultaneously finding both the least cost possible generation mix and a physically feasible one that doesn't blow up our grid. And then they pay everybody what they're owed based on their marginal, based on the locational marginal price at their location. Because that's the value that they are producing when they inject one more megawatt hour into the grid. Do the big monopoly utilities like Southern Company, Duke, do they have power markets within them at this point? Because it's just the better way to run. So they do what are often called balancing markets where they don't fully optimize like that across the whole region with everybody submitting their bids to a central generator. They know their bids, like they know their marginal cost of their generators. And what they do is they usually run a market on the margin, on the edges, right? OK, I'm a little bit long. I think I'm going to be a little bit long this hour. Anybody a little bit short and wants to buy a little bit of my extra power. And then somebody else out there is like, oh, yeah, actually, I could turn off one of my expensive generators if you've got a cheaper one. I'll buy a little bit of that on the margin. And that's what they do in the southeast. In fact, Southern Company runs a combined dispatch for a bunch of other smaller ones to make sure that they're doing that more effectively. In the west, the California independent system operator, which runs all of these organized, you know, the whole market for the California system, for most of the California system that isn't publicly owned, they also run on behalf of a whole bunch of other transmission utilities across the west, what's known as the Western Energy and Balance Market, where they are kind of bilaterally trading between individual utilities. But it's done at that utility by utility level, not at the generator by generator level, the way a fully competitive electricity market would. I should also add, maybe add for our European friends or anybody who wants to, is curious about how Europe does this. They do it a little bit differently. They've decided that they would prefer to have a single unified European market that also includes the UK and other non-EU but allied countries like Switzerland. And in order to have a market that large, 'cause the European system is huge, it's far bigger than any one of our regional transmission organizations like Germany or France are about as big as PJM or ERCOT or any one of our individual regional grid operators. So they can't actually solve, or at least they couldn't until recently, actually solve that physics-based transmission constrained dispatch for the whole system at once. And so they decided to simplify things and run what are known as zonal markets. So they only capture flow constraints between very large regions, usually countries, whole countries. Some countries along skinny ones like Italy and Norway, they split up into multiple zones north-south 'cause they know the transmission system is often constrained along those paths. They basically have one big area and they run a power exchange that's actually separate from the system operations that is purely a commodity exchange. And they only capture constraints on interchanges between the regions that are approximated constraints on how much you might be able to trade between, say, France and Iberia or Northern Norway and southern Norway. And then those trades aren't actually physically possible, often because they're ignoring all of the transmission constraints. And so then they go to a secondary market where each of the individual transmission operators is responsible for fixing shit in their territory. And so they have to redispatch all the generators a little bit in order to make all of the dispatch feasible and people get paid in a separate imbalance market for going up or down in that market as well. And so it's a little less efficient way to operate the market on a kind of grid by grid basis or region by region basis. But what it allows them to do is have huge liquid markets that go across wide regions and they prefer that kind of broad gains from trade from having a really large, easy to participate in market, relative and then incurring the cost of redispatch relative to what we do, which is having several smaller markets that are kind of perfectly dispatched within their market. But you know, PGM and MISO may not talk well to each other or may not be making the best use of their transmission interchanges. So we have chosen to fragment our markets and operate them in different ways that reflect different institutional preferences, I guess. - It's so cool. - There's some politics now to the zones. So in Europe, there's been a big pressure because of the rise of renewables to change to nodal prices, to use these locational marginal prices that are node by node basis, location by location. Because there are some very frequent congestions within some of these big broad regions, like in Germany, moving power from the North Sea where there's lots of wind power to the South where there's all the Bavarian industry and towns and whatever is often constrained. Or in the UK, moving power from Scotland to London is often constrained. But if they were to go to the locational marginal prices, what that would mean is more expensive power in the consuming places that are on the other side of those constraints. So like London or the Bavarian factories and less expensive power in the places where the wind farms are injecting into the grid. And so politically, they don't want to do that. They're happy to have this sort of uniform socialized price. So that's another reason why Europe has been in America. And I realized that utility pricing varies by state. But is it generally the case, let's say in PGM or New York or MISO or California that your local nodal price has some effect to the price you pay as a consumer? or it is basically everyone in New Jersey or everyone in Pennsylvania pay the same utility rate. - Yeah, great socialized all of that. - And in fact, I guess that's good segue, 'cause the answer is more of the latter. In the US, unless you're a large industrial consumer that's directly participating in the wholesale market, you tend to be settled on a load zone price, which is similar to that European market price I was just talking about where they take all of the locational prices in a given transmission system, a transmission owner's territory like PSCG or Con Edison or whoever else, and average it out across that area so that everybody is paying the same per megawatt hour across the whole utilities territory. That's again a choice to socialize the power prices across a broader region and say it's not your fault, you happen to be on the other side of some substation that tends to be more congested than me and so you're paying more. And there's again, there's some good equity reasons why you might wanna do that, right, to make everything fair. It does however blunt the accurate market signal, which is to say if I'm trying to think about like at my house or my business, does it make sense to put in a battery or put in a solar system? It would actually be better for people to know when they're making that decision, are you in one of those expensive nodes or one of the cheap nodes? Because if you're one of the expensive nodes, you should get more value for your solar system and you should be encouraged to do that somehow through economics and if you're in the less valuable location you shouldn't and we lose that signal at the level of retail pricing for the most part and in competitive retail markets, the retailer themselves tends to average those prices out amongst all their customers. But if I'm a large customer like Princeton here, which runs its own power plant and is like a tens of megawatt load, we participate directly in the PJM market and so we're actually seeing those real nodal prices on a regular basis and can operate our whole complicated energy plant system very effectively and very efficiently to arbitrage against those prices to lower our cost of energy production and by doing so, lower the cost of the PJM system as a whole. And so you might imagine a world in the future where automation and distributed generation distributed storage, thermostatic control devices of our smart thermostats or whatever. All of that is automated in a way that makes it as easy for my household to act like Princeton's central plant and participate more directly in the grid. And if that's the case, then the value of sending more granular locational prices all the way through to consumers starts to get a lot more important than if we can't do that, if nobody can really respond to the price anyway, so why send it to them. And that's one of those big shifts that I think has been slowly occurring and mounting over time is that 20 years ago, it didn't matter if you knew that you are at a high price note or a low price note or your system, you're at a location that tended to have very congested prices in the afternoon and so power in the afternoon was more valuable than some other time of the day, 'cause you couldn't do anything about it. But now, there's lots of ways in which individuals, even down to the household level, could be responding to those prices, and so the opportunity cost of not passing them through demand at a more granular scale is getting larger and larger. And it's prompting some people to revisit how we might deal with that. - It would be a triumph, I think, for the country's dads of all genders. You know, dads telling their families that they're not gonna raise a thermostat and that they should instead put a sweater on, I think. - Well, that's an even, that's the classic demand response, right, like the age of human Carter, it's like put a sweater on, right? But now it's not. I mean, the demand response options now if prices are high are that my charger that I'm not even thinking about slows the rate at my car is charging from seven kilowatts to one kilowatt and then back up again later when the price is lower. And I still have enough to get to work in the morning 'cause it knows how much I need and it's handling all that on an automated basis, right? Or my smart thermostat is pre-cooling my house by a couple of degrees because it knows the prices about to get high and then it turns it down later, it turns the thermostat down later in the hot afternoon. And I don't really notice, I notice that the temperature is varying a couple of degrees and that's what it always does 'cause my thermostat isn't that precise anyway. So there's lots of automated ways in which we could be responding to these things now, but-- - But if you could, what I'm saying, Jesse, I hear all of that, but what I'm saying is that if there was some way to return some share of the savings from that demand response to the consumer, just imagine the entertainment value to the nations. - Yeah, so running around with your app being like, "Oh, power prices are high." - Well, just even watching it all happen on your phone as you saw the sense pile in, it would be like getting the push notification that you were gonna save 50 cents on your power bill. - Exactly, that's-- - It is unfortunate a little bit like day trading penny stocks. It might not be worth the squeeze, right? But you just might not be worth the squeeze, but in some cases, it is because electricity prices can go to several hundred dollars in megawatt. - I think we found the right way to get the country's kind of dopamine poison zoomers who love sports betting and gambling and getting on their phone. - They need a little push app on their phone. - Exactly. - All our listeners out there at DraftKings, remember that it was our idea and we deserve a little bit of the credit for this. - I'd really like to be able to manage my daily, I'd like to be able to pay my electricity bill through DraftKings and manage the power system in my house through DraftKings. - Hey, that's not a bad idea. - That's what I want. And I don't see why I can't have daily fantasy power system on my phone. We need to wrap this episode up and we haven't even gotten to long cycle prices yet. - Don't worry, there's always takes more than one lecture in my class anyway, so. - Okay, I have some questions, but I think we should just come back. Thank you so much for listening. We will link to the other episodes of Shifke Summer School in the show notes. I'll say I learned a lot from this. If you learned something and you think you have friends who might enjoy it too, send it to them. Spam your group chats with it. And also if you enjoyed it, feel free to rate this episode. In fact, please do rate this episode in whatever podcast app you use, be it Apple Podcast, Spotify, Overcast, Deezer, The French One. Any of them, we appreciate it. It means a lot for us. Shifke is a production of Heat Map News. Our editors are Nicolores Shella and Jolene Goodman. Audio engineering is by Nick Woodbury and Jacob Lambert. Our music, as always, is by the inimitable Adam Kromelau, who might be listening to this show right now. Thanks so much for listening and see you next week. (gentle music)