Energy Transition Puts Grid Reliability to the Test
In February, PJM Interconnection, the largest wholesale electricity market in the U.S., published a report pointing to the real possibility that electricity supply in the market could fall short of the level needed for reliable grid operation in just five years. The potential shortfall is tied to the pace of the energy transition, in which fossil fuel generators, particularly coal plants, are retiring faster than they can be replaced by new clean and renewable generation.
The report has led to a flurry of activity in PJM, as its members rush headlong into a process to reform how the market manages, and values the reliability generators bring to the electricity system.
Abe Silverman, Director of the Non-Technical Barriers to the Clean Energy Transition program at Columbia University’s Center on Global Energy Policy, explores efforts underway in PJM to revise market rules to ensure that supply meets demand at all times, including during extreme weather events that have recently threatened the reliability of the grid. The resource adequacy challenges to be addressed are immensely complex and extend beyond the fundamentals of energy technology to encompass the diverse economic and environmental priorities at play in the market and, more broadly, across the nation.
Andy Stone: Welcome to the Energy Policy Now podcast from the Kleinman Center for Energy Policy at the University of Pennsylvania. I’m Andy Stone. In February, PJM Interconnection, which is the largest wholesale electricity market in the US, published a report that points to the very real possibility that electricity supply in the market could fall short of the level needed for reliable grid operation in just five years. The potential shortfall is tied to the pace of the energy transition in which fossil fuel generators — and coal plants in particular — are retiring faster than they can be replaced by new, clean and renewable generation.
The report has led to a flurry of activity in PJM, as its members rush headlong into a process to reform the way the market manages and values the reliability that generators bring to the electricity system. The hope is that PJM can revise its market rules to ensure that supply meets demand at all times, including during extreme weather events that have recently threatened the reliability of the grid. The problems to be addressed are profoundly complex and extend beyond the fundamentals of energy technology to encompass the diverse economic and environmental priorities at play in the market and, more broadly, across the nation.
Here to discuss the challenges facing PJM and the solutions that may lie ahead is today’s guest, Abe Silverman. Abe is Director of the Non-Technical Barriers to the Clean Energy Transition Research and Policy Program at the Center on Global Energy Policy at Columbia University. Until recently, Abe was General Counsel for the New Jersey Board of Public Utilities, the state’s electricity regulator, and a stakeholder in the PJM market. Abe, welcome to the podcast.
Abe Silverman: Thanks, Andy. It’s great to be here.
Stone: So you’re fresh out of the New Jersey BPU, and New Jersey is a progressive, clean energy-focused state that also happens to be one of the 13 states that is within the PJM market. I wonder if, to start our conversation and to lay the groundwork for our discussion of the resource adequacy challenges that PJM is facing, you could begin by describing the clean energy policies that states like New Jersey are pushing, and what they are not willing to compromise on in terms of clean energy adoption.
Silverman: States like New Jersey have these incredibly robust clean energy targets, and they’re being joined by a lot of other states throughout the PJM region, but also customers, large corporates and other cities and municipalities, all looking to supply an increasing amount of their energy from carbon-free resources. So there has been this real push-and-pull at PJM between states and consumers who are demanding additional clean energy options, and things like the future reliability of the grid, as we get to a more and more decarbonized grid.
So a lot of our time spent in New Jersey was actually thinking about how do we meet the state clean energy goals of New Jersey? How do we meet the state clean energy goals of Illinois and other states like Maryland or Washington, D.C., which all have very aggressive targets, and do it in a framework that keeps the lights on throughout the PJM region.
Stone: How are the states in the PJM market working to address this at this point?
Silverman: There are really three different things going on with the state policies that we’re working to integrate into the PJM market. We really see this trend where states are favoring clean sources of electricity. For example, New Jersey has its 100% clean energy by 2035 program, or its offshore wind program, or its solar program, which is trying to bring 750 megawatts a year onto the market. Those are all examples of favoring clean energy sources, and those are largely being driven by state policies in the absence of a larger, federal, cohesive policy.
So we have the favoring of resources, and then we also have a bunch of state and federal policies that are disfavoring a generation that has emissions, traditionally fossil fuel resources. And so we have a number of programs that are looking at the total carbon emissions from the fossil fleet in various states: Illinois’ CEJA program, Clean Equity Jobs Act out of Illinois has been a real groundbreaking piece of legislation in Illinois because it effectively says they have to move away from fossil resources by 2030 and really backstop a lot of that stuff with clean. New Jersey has very comparable, though somewhat less aggressive policies, like its Global Warming Response Act that put in emissions standards on existing fossil resources.
So we have that sort of second group of state policies that are leading to the shut-down, retirement — for policy reasons — of existing fossil resources. And then we have the sort of third X factor in all of this, which is: What is the load growth that we’re going to see across the PJM region? Utilities across this country have not had sustained load growth since the early 2000s. For the most part, load has been relatively steady. Certainly there are pockets where it grows, but now we have a lot of state policies that are meeting their carbon targets through a mix of clean energy requirements, but also electrification of building and transportation sectors. As we start looking at the load growth, we start to see — and I think this is going to show some of the cracks in the existing PJM system. So we have lots more clean, some less fossil, and a lot more demand. And that’s where the states are sort of leading the PJM market. And I think we’re basically seeing a catch-up, where PJM is now looking to talk about and think about how these various policies are really affecting the long-term reliability of the PJM market.
Stone: That load growth issue is very interesting, right? As you said, load growth has been pretty much flat, meaning the load has not grown. There’s no load growth, right? But you’ve got the issue of the EVs coming on. We’ve got to electrify everything, including home heating, which is becoming more of a thing. And also these data centers, right? There are certain pockets in PJM where a lot of new data centers are coming in, and the load growth in those areas looks pretty darn dramatic. So again, all these pressures are coming at a time when load growth is starting to pick up again. Is that right?
Silverman: Yes, exactly. And the really interesting thing about load growth is PJM has identified data centers as a really large factor. And I think that is probably true, but the other pieces, I think, could be even more profound. If we — and I just look at what President Biden just did, what the EPA just did with their EV policies and the new CAFE standards. And there, they’re looking at 60% of all new vehicles sold in America by 2032 are going to be electric. That is a profound shift in the load patterns and the amount of consumption that we’re going to have in this country, if it actually succeeds.
So New Jersey actually looked at this in its Energy Master Plan of 2020 and effectively showed that we, by 2050 — so this is a long-term projection — but effectively showed that we were going to need three times as much generated capacity to meet two times the amount of load in New Jersey alone. So there is this huge, profound, titanic shift in the way we heat our houses, the way we drive our cars and vehicles, as well as manage our data. As we see these various programs start to kick in, it is really important to do a variety of scenario plans where you look at high electrification cases, high electrification of transportation cases, and so on down the list.
And of course counteracted against those load-growing policies are also load-shrinking policies. In New Jersey and other states, we have very aggressive energy efficiency targets, and that’s going to put some downward pressure on that load growth. But at the end of the day, if you think about it, we’re effectively replacing every gas station in America with electricity, and we’re going to need the infrastructure to do that. And so if you start thinking about all that energy that’s currently delivered through pipes and wires and by truck, we’re really talking about moving a good portion of that energy demand onto the electric grid instead. And that load growth piece is really fundamental in how we think about planning for a future and planning for the reliability of the grid, particularly as we move more and more critical services onto the electric grid.
Stone: Okay, so the impact of these policies at the state level, federal level, and load growth really has caught the attention of PJM. It’s very much aware of what’s going on, and on February 24th, the market released a white paper. The title of that white paper was “Energy Transition in PJM: Resource Retirements, Replacements, and Risks.” It’s actually the third report in the series on the energy transition, but this is the one we’re talking about today. The report warns of a potential shortage of generating capacity, or in the lingo of the industry, a resource “inadequacy” later this decade. Could you introduce the major findings of the report and the gravity of the situation in PJM?
Silverman: Yes, and let me even step back and talk about what is resource adequacy? When I’m explaining what I do to people at cocktail parties and such, and their eyes start glazing over, usually they really start glazing when I get to resource adequacy.
Stone: I can imagine they get really excited about this stuff.
Silverman: Oh, totally. [LAUGHTER] You know, some of the cocktail parties I go to, yes. So let’s talk about what is resource adequacy? At its heart, it’s very simple. It’s making sure that you have enough generation in the places you need it to meet customer demand for electricity. When you step back and think about the miracle of the electric grid, because it is — I’ve heard people call it “the world’s largest machine.” It is kind of amazing because other than a little bit of battery storage and pumped hydro storage, all the electricity we use is being generated instantaneously, and it all needs to stay in balance.
And so resource adequacy is just this very complicated process of looking forward, saying, “Okay, in three years, what do we think electricity is going to look like? How much generation do we need to have? What do we think the consumer demand is going to be?” And then we also have to make sure that we have a reserve margin because this stuff is too important. You don’t want — you know, if something on the system breaks, if the weather is a lot colder than we expect or a lot hotter than we expect, and electricity use skyrockets, how do we actually make sure that the system is resilient, that we have redundancy in the system? We often refer to that as “the reserve margin.” And that kicks in when things break. In fact, we don’t just look — when system planners plan out the grid and look at the resource adequacy — they don’t just look at one thing breaking. They look at two things breaking.
So you basically, in an N – 1 contingency — sorry for the jargon, but this is how we talk — in an N – 1 contingency, you take out one element on the grid. That could be a transmission line. It could be a power plant. It could be a large load. And you expect the grid to seamlessly keep working. And then you go ahead, and you say, “Okay, let’s do another — let’s take out the next most severe consequence, the next biggest thing to break.” And you take that thing out, and the grid, hopefully will keep working then, too.
So that’s really what we’re thinking about with resource adequacy. And traditionally, when we plan for resource adequacy, we ask the market to identify the least cost suite of generation resources that meets our reliability needs, including these reserve margins, in case something goes wrong. I think we actually — as state clean energy policies become a bigger driver of investment in the electric grid, we actually need to start changing that question, so not just look at what the lowest-cost suite of resources is that meets reliability, but it’s got to be the lowest-cost suite of resources that meet — reliability, because those things are must-haves — but that also meet state clean energy targets.
And the answer is going to be different if you try to do those two analyses, and so a lot of where my critique of PJM comes in is that they need to actually start explicitly accounting for consumer demand for clean energy and for state laws that mandate certain amounts of clean energy.
Stone: Abe, it’s really interesting what you just said. Reliability really has always been so much of what PJM and all the electricity markets have always been about — reliability first. Obviously economics are important, as well, but we’re talking here about kind of a new consideration, which is that this electricity be clean.
Silverman: Yes, Andy, I would even push back slightly, maybe offer a friendly amendment to what you just said. The ISOs have always said, “Reliability is job one.” I think that’s certainly true of federal and state policy-makers, as well. But there’s also the idea of cost that’s deeply embedded in reliability. So when we talk about “reliability,” we don’t say, “Reliability at any cost.” We could gold-plate the system. We could bury lines to your house. We could do all sorts of things if cost wasn’t an issue, that we don’t do.
So we say, “What is the lowest-cost system that we can get that meets certain predefined reliability metrics?” And so for the bulk power system, we’ve always sort of had this almost mythical, one-year-in-ten standard, which is we would expect to have enough generation on the grid to meet customer demand every time in ten years, except for once. And that’s the standard we plan to, and it actually drives a huge amount of the economics of the grid because we do have this target reliability metric. I think one of the things that we’re going to see over the next couple of years from PJM is really fundamentally rethinking what that metric looks like.
But I think the other piece of it — sort of beyond just the scenario-planning and the loss-of-load expectation kind of risk calculation that people in my world do all the time and argue about all the time — I think the question is shifting, and this is the piece that I think is absolutely critical as we move forward with the energy transition. We need to talk about not just what is the loss of load expectation from the traditional grid, but what does it look like in this new world we’re creating, where there are more variable resources, where climate change-driven severe weather patterns are more pronounced? I think if we’ve learned anything from the reliability crises over the last couple of years, it’s that when Mother Nature fights the grid, Mother Nature is going to win. And we can do what we can to keep the lights on, but at some level, when we see ever-increasing hot weather, ever-increasing cold weather and more severe cold weather, as well as increased storms, that’s all going to drive the future reliability planning that we do.
So anyway, just coming back to the point, I do think that resource adequacy has always been about setting the standard and then finding — and then the markets find — the least-cost means of meeting that standard. And so that’s kind of how the two really do work in parallel.
Stone: So let’s go back to this issue of the February 24th report, because simultaneous to the release of that report on the same day, the PJM Board of Directors, alarmed, I think, somewhat by what that report came up with, issued a letter to PJM stakeholders that initiated a process to address the resource adequacy concerns, again that were identified in that report. Could you tell us about this process and its goals?
Silverman: So there was this fascinating report coming out of PJM, where they looked at the future 2030 grid and kind of had an “Oh, shoot” moment. That may not be the word they used, but that’s the word I’ll use. And so in this report, they lay out really four fundamental things that have them concerned. One is that they are seeing decreasing reserve margins for the first time in PJM’s history. PJM has always been blessed with a very large, healthy reserve margin. In fact, for the last several years, a lot of the sort of zeitgeist has been around the fact that PJM has too high a reserve margin. And now to have PJM coming in and saying, “Actually we’re not sure we have enough generation resources in 2030” was a sea change in the way PJM has been talking about these things.
So there’s this declining reserve margin for the first time. What’s driving that? It really is a combination of two factors, both having to do with existing fossil resources and whether they remain in the market. First the PJM study noted that there are about 20 gigawatts, so that’s 20,000 megawatts — think of that as 20 large coal facilities — 20 gigawatts of generation that was going to be required to retire for state and federal environmental permitting reasons.
On top of that, they were also looking at approximately another 20 gigawatts of dispatchable fossil resources that were retiring for economics in that 2020 to 2035 timeframe. Really for the second time in its history, PJM was seeing a large number of resources retiring, the first being associated with coal retirements, set to be at 2010, associated with the Mercury and Air Toxics rules out of the Environmental Protection Agency.
So for the second time, PJM is looking at the retirement of up to 40,000 megawatts, 40 gigawatts of generation, over the next decade. And so that had them very concerned. And just to give you a bit of context, the entire PJM system, including reserve margin, is just under 200 gigawatts.
Stone: So we’re talking 20% of the market, right?
Silverman: We’re talking 20% of the market, right. For anybody who works in the clean energy transition, they probably hear that number, and they say, “Well, that’s not fast enough.” But to utility planners, they think, “Oh my God, what are we going to do? How are we going to replace those 20 gigawatts?”
So that was problem number one that PJL identified in this report. The second is the load growth. And we’ve sort of already talked about that, but they in particular cite this concern over new data centers, largely in the Washington, D.C. suburbs, on the Virginia side, where they have identified a large amount, in the 10-gigawatt range, of new load coming on it. I think the data centers are obviously a big issue. I actually think electrification of medium and heavy-duty vehicles may be at least as significant, as well as, of course, all the electrification of the light-duty vehicles and housing stock, as we move more towards heat pumps and away from gas furnaces. All of those load-growth issues are really critical and can drive a huge swing in what the future needs of the grid are.
So then we get into — you know, those are kind of the macro pieces. And then PJM sort of highlights two additional problems. These get a little bit wonky, but they’re absolutely critical. One is that PJM spends a lot of time talking about the rate at which new generation resources are coming into service. And this is often lumped together or called the “interconnection process,” which is really the process of plugging in a new generator into the grid. Fossil resources go through the interconnection process. Renewable resources go through the interconnection process, and it is the key gating item to bringing a new resource onto the grid.
So PJM currently has over 200 gigawatts, 200,000 megawatts of generation in their interconnection queue. Almost all of it is renewable clean resources — wind, solar, about 30% is batteries. All of these things are sort of coming in, fighting to take market share from the existing incumbent resources. Now that sounds great, and if all of those resources, if those 200 gigawatts of generation were coming through the interconnection queue, actually hooking up to the grid, I don’t think that we would be having this conversation right now. But what PJM identified is that something less than 20% of the resources currently in the queue are actually reaching commercial operation. In fact, the story is even more grim in the last couple of years, where there are even fewer resources coming on and actually interconnecting.
So I think PJM, giving this in the light most favorable to them, they were very concerned when they’re sort of looking at this inflow/outflow analysis and saying that the outflows seem pretty certain, because they are being driven by state laws.
Stone: Meaning the coal retirements?
Silverman: The coal retirements, and some gas. We can’t just talk about coal. There are gas standards, as well, that are coming. And so you have all those exiting resources that are there, and yet the flow of new resources to replace them is coming in at a trickle. And this is a result of a complicated set of policies, but effectively, the PJM interconnection process has ground to a halt in the last couple of years.
PJM used to process a couple of hundred interconnection requests a year. Large central station generators would come in. They would go through a fairly orderly two to three-year study process. Then they would connect. These days, PJM is just being bombarded on all sides by new resources coming into the queue, and as a result, the process has slowed down dramatically. In fact, PJM is just part-way through its new process, trying to streamline the interconnection rules and improve them, but even under the new, improved rules, we’re still not seeing the number of resources come onto the grid that we need to effectively backstop the departure of the fossil.
Now this is a really critical point, because PJM itself says that under a high new-entry scenario — so where the interconnection issues are resolved or ameliorated — there is actually sufficient generation coming onto the grid to replace the departing resources. So in a lot of ways, this report that PJM put out has two — if I were in journalism, and I was writing a headline for this report, which I know would be read by exactly three people — but it would say, alternatively, “PJM Predicts Resource Adequacy Shortfalls in 2030.” That is one framing. The other framing, the one that I actually happen to subscribe to, is, “PJM Identifies Need to Improve Interconnection Process to Avoid Future Reliability Issues.” Two totally different framings of exactly the same report.
And then I’ll even go one cut down. This is where we’re getting really into the deep cuts here. We’re maybe not even on the B side anymore. We’re on the C side. But there is this issue that all the system operators across the country are dealing with, which is how do you accredit resources for capacity purposes? And the science tells us, and a lot of engineering tells us that as you get more and more solar or wind resources coming onto the grid, their effective load-carrying capability goes down.
And what does that actually mean? It just means that if the sun isn’t shining over New Jersey on a particular day, all the solar resources in New Jersey are going to have a subpar day. The same thing for offshore wind. If the wind isn’t blowing on any particular day, you’re going to have less wind on that day. This is not rocket science. It’s very clear. But what we’ve found, and we see this as it has become a very important issue in places like California that have much higher renewable penetrations than we do in the PJM region, is that the production of variable resources is codependent across a large geographic area. And so to account for that, PJM is looking at adjusting the capacity accreditation.
So think of it in another way. A megawatt of solar coming onto the grid does not have the same dispatchable characteristics as a megawatt of coal leaving the grid or a megawatt of natural gas leaving the grid. Now, footnote: We also see a lot of those same interdependencies on the gas system and the coal system, as well. Failures of the natural gas fleet are also very correlated across particular kinds of weather patterns, and so we see a lot of these resources actually showing up. You know, when you sort of step back and think about it, in the modern grid, they have a lower capacity factor or a lower capacity value than we may have been ascribing them for the last couple of years.
So when you combine these four factors — decreasing amounts of fossil generation, the discount that’s being applied to new clean energy resources through the ELCC methodology, the problem with the interconnection queue choking off new entry — then you add in load growth. And I think that’s where all of a sudden PJM and the Board came out with this study and said, “Okay, we need to fundamentally step back and think about what we’re doing here.
Stone: So just to reiterate what you’ve been saying, one of the points here is that one of the challenges we’re seeing is that it’s going to take a lot more renewable energy, wind and solar, to replace the — say the coal and to some extent the natural gas generation that’s retiring. And that’s due to this concept known as “capacity values,” right?
Silverman: Yes, absolutely. Because if you think about running the grid, it’s much easier to run the grid when you have a whole bunch of fossil resources that basically turn on and off. And then when you start injecting large amounts of clean energy resources, you basically have a probabilistic weighted average of the amount of energy that they can produce, and so you effectively say, “All right, solar is on about a third of the time, so we’re going to give it a roughly one-third value.” So if you are taking a whole bunch of fossil off the grid and replacing it with clean, you actually need on a nameplate capacity to have a lot more clean resources coming in.
And so one of the points that the PJM report made and that the Board of Directors’ letter said is they noted that, “Okay, we have 200 gigawatts of generation in the PJM queue, but once we start discounting that and looking at the implied actual capacity value, that may not even be enough.” And then particularly, when you start looking at the problem with getting those resources through to the grid because of the interconnection queue and the lack of transmission planning, that we actually are not seeing enough generation coming online to replace the stuff that’s leaving.
And so particularly when you sort of think about 200 gigawatts of generation in the queue, you discount that to maybe 30%. That’s 90 gigawatts. And then you discount that again by the sort of 5 to 10% of projects that are actually making it through the queue. And then you start having a real potential mismatch between the flow of resources coming into the grid and the ones leaving the grid. And that’s really where the PJM Board sort of pulled the fire alarm.
Stone: Okay, so the future of PJM looks very uncertain, potentially scary. Summing up: More generation is leaving than is quickly replacing that generation that’s being lost. Now, the future again looks uncertain, but one of the additional plot twists here, we’ll call it, is that there are already reliability and resource adequacy concerns that exist right now, day one — the day that we’re having this conversation – in the PJM market. PJM, just in December, narrowly averted a crisis during Winter Storm Elliott. That was the cold snap around Christmas, when a whole lot of generators in the market failed to perform. What happened then, and what does that mean about the base point of reliability we’re working from going forward?
Silverman: Oh, that’s such a great question, Andy, because the first thing I want to note is the PJM analysis suggested that we have this problem in 2030. I think a lot of us who actually step back and look at the data have raised some pretty serious concerns with whether the sort of very grim case that PJM is presenting is actually accurate. A number of parties filed letters actually with the PJM Board, critiquing the PJM analysis, noting that some of the retirement predictions as a result of state public policies may be overstated. They didn’t necessarily account for reliability safety valves that were incorporated into some of those state policies, including ones in New Jersey. So there was a real sense that perhaps the situation wasn’t quite as grim as PJM was portraying it.
The second critique we really saw coming out of particularly the clean energy advocacy folks was that PJM is assuming not only that 25 gigawatts of generation retires for state public policy reasons, but that another 15 gigawatts is retiring due to economics and is expected to leave the grid. One of the things that this community pointed out — and I have to say I have some sympathy with this argument — was that PJM is ignoring the fact that its markets will show higher prices in the future if we start getting into these scarcity conditions, and that that will affect the willingness of generators to stay in the market. And so when you sort of ignore that impact on generator bidding behavior and retirement plans, you’re painting a very grim picture that maybe things aren’t as bad as they are.
The last piece of it is I think a lot of us, particularly who have been on the ground, trying to get clean energy built in PJM, have really seen the impact of the supply chain, the pandemic-associated supply chain barriers. And so what PJM is really extrapolating from in terms of how many projects are going to make it to commercial operation was really affected negatively by the pandemic and by the supply chain. So if you’re trying to draw — like at the beginning of all those investment podcasts that I listen to, past performance is not indicative of future performance. And so you were a little bit concerned that PJM is taking a worst-case scenario by looking at the completion rates over the last couple of years, even though those have been pretty dramatically affected by both the interconnection delays at PJM, but then also compounded by these supply chain issues.
And so when you take a step back, I think there’s significant question as to whether the reliability issues that PJM identified in its paper really warrant the kind of response of the road they’re going down. Except then, of course, they have to step back and say, “As somebody who has worked in state government and also frankly at FERC and in the private sector, you have to plan for a worst-case scenario.” And so I think we all fully support what PJM is doing here. We’re just not — I think a lot of us are not quite as convinced that the situation is as dire as PJM says it is. But all that said, I absolutely support looking into it and giving it some serious thought.
So now let me turn to really what your question was about the storm Elliott in December. And I have to tell you, I was in New Jersey government at the time, and we started getting phone calls and emails and texts from really across PJM and across our regulated utilities, all suggesting that there was a very serious reliability event. I will never forget the morning of Christmas Eve, we were up visiting family in Brooklyn, and I had to step out. I went to our local coffee shop and was taking phone calls, increasingly frantic, about: Is the system going to survive? So what actually happened?
Well, the weather was the key driver. And it wasn’t the coldest day we’ve ever seen, but it was a day where the temperature dropped super rapidly. And so you had this evolving condition on the grid where PJM’s load forecast — you know, their prediction of how much electricity was going to be needed the next day — was too low. You had these rapid drops in temperature. You had a couple of nuclear units that came offline. So as you think about: What do we design our grid to? It’s an N – 1 or N – 2 contingency. But here we had, I don’t know, N – 3? And then on top of that, we started seeing really an unprecedented level of failure on the natural gas system. Books will be written, at least by the North American Electric Reliability Council and people who read that kind of material on exactly what happened. But here’s what we know.
There were certain natural gas faculties — so you had this sort of unprecedented number of things that happened that day. First, like I said, load was higher than expected. And then you had a couple of nuclear units drop. And then you had failures of natural gas-fired generation resources. And this is where you get really complicated into the root cause analysis. I’m not necessarily the expert on that, and I’m not going to opine on that. But what we effectively saw was a combination of natural gas supply dropping, including things from the Marcellus Wells that were shut in or otherwise unavailable. So you had less —
Stone: All due to the severe weather, right?
Silverman: All due to the severe weather. You had less supply coming onto the pipeline system, and then you had a number of generation resources, particularly gas resources, failing. And there was something at its peak, like 40 gigawatts of gas that did not show up when dispatched. So that’s a huge portion of the PJM system. Think 200 gigawatts total on any given day. You usually have 140, 150, 160,000 megawatts showing up. And here we had 40 gigawatts of gas failing for various reasons. And there is ongoing litigation right now, as we speak, at Burke, about whether — It’s a lot of finger-pointing. The gas-fire generators say, “Well, we couldn’t get the gas because of inflexibility of the natural gas pipeline system.” They’re saying, “Hey, PJM didn’t tell us to turn on in time.” There’s about 2 billion dollars in penalties hinging on exactly what happened and who is to blame. And you could have a whole series of podcasts on that.
Stone: Which is a ridiculously high sum. Two billion dollars in penalties is completely unprecedented.
Silverman: It is high. We have to think about context, remember? Because when ERCOT finished up with Storm Uri, they had in the tens of billions.
Stone: Okay, so PJM yes. But ERCOT —
Silverman: Yes, two billion dollars is a lot of money. I think it represents about one year of capacity revenues. It’s a significant amount, and I think there will be a lot of question as to whether it was the right amount, and whether there should be a rejiggering of the incentives and the penalty structure, but suffice it to say that when these resources were needed, they did not show up. And so this has really, I think, added to the sense of unease as we look to this clean energy future where we have more intermittent resources. We’re losing some of these fossil and gas resources, and the fossil and gas resources just aren’t as reliable as everyone has always thought they were.
So when you add those things together, particularly with the specter of climate change and weather that’s getting worse over time, we have to step back and say, “Okay, are our markets actually giving us a suite of resources that is reliable?” Now the amazing thing is — actually not that we were close to failure — but that the PJM system survived, because it is, I think, a real testament to the incredible hard work that all the grid operators at PJM did, a testament to demand response and state policies that were driving conservation. And it’s partially due to luck, frankly. And you saw that down in the Duke Service territory. In other parts of the country, they were actually having a lost load.
And so as we think about the Elliott storm, and then the Uri storm in Texas and the polar vortex back in 2014 — these were all examples that when the weather gets super cold, problems just start materializing. And each time we go through one of these extreme weather events, and we do our root cause analysis, we come out with a series of recommendations. We, the royal “we” here — FERC, NERC, and other entities all sort of go through and say, “Okay, here are the ten things we need to do to stop this from happening again.”
Stone: Well, I think it’s interesting in PJM — this was supposed to be solved, right? So in 2014, there was the winter polar vortex, a similar situation of extreme cold impacting the system. You had generation outages. And in response to that, PJM responded with something it calls “capacity performance,” which is a system of requirements and of essentially carrots and sticks to ensure that generators are going to be available to operate, even under extreme conditions. Those reforms, based upon the results that we saw, what happened during Winter Storm Elliott, those reforms were essentially, it looks like, a failure. Is that too strong language to use?
Silverman: I think you need to step back and ask the question: What would it have been if we hadn’t had those reforms? Because at some level, the system did hold up.
Stone: But a lot of generation that was supposed to be reliable did not show up, right?
Silverman: So what capacity performance does is it provides a financial incentive for generators to make investments. None of these generation resources will ever be one hundred percent reliable. That’s just the world we live in. What do generators do in response to a threat of penalties? They actually do make investments in things like weatherization and identifying where the problems are in their own systems. And that can be fuel procurement, that can be switching to backup fuel. I know my former company, NRG Energy, went through a very traumatic event in ERCOT in 2011, where a number of our generating resources failed due to cold weather. The company made significant investments and modernized its operating procedures to try to avoid having those problems occur again.
So I think you do need to ask the question, did generators respond to the CP market mechanism in a way that made it better than it otherwise would have been? And I think you have to look at this, and you want to really step back and say, “Do that type of analysis. And how were generators responding to those price signals?” Now that said, I think we all recognize that having 40% of your generation fleet on a cold day fail is simply not acceptable. And then you sort of get into what happened between 2014, when, as you say, capacity performance came in heralded by many stakeholders as a real evolution in the capacity market design. So what happened between then, when generators were making investments in reliability, were buying additional pipeline capacity on the grid, were really taking efforts to weatherize their coal plants and gas plants. So what happened between that 2014 period and 2022?
I think part of it was there were no capacity performance assessment hours for almost a decade. So I think it’s fair to ask the question: Did some of the generators get complacent, because they had not seen the carrots and sticks kick in for so long? Did that make them more comfortable? Were there things that they stopped doing, in terms of having extra shifts or doing extra preventative maintenance before a cold snap? So here we have a very large amount of penalties that kick in, and if it’s really only going to be once every ten years, then I think it’s almost human nature that we forget the lessons of ten years ago, and we’re more focused on the economics of today and yesterday. And so I think there will be a movement to reform the capacity performance mechanism, to have it be a little bit more spread out, maybe trigger more often, so that you have these incentives, and you get to see how your generation fleet is performing without making it quite such a draconian penalty on days when you don’t perform.
Stone: Let’s jump up to the present day, from 2014 to April of 2023. The process that PJM has initiated and is underway right now to address some of the resource adequacy concerns that the February 24th white paper identified for later this decade. The process that is undergoing right now is called “the Critical Issue Fast Path,” and the goal is to come up with some solutions and submit those solutions in a plan to the FERC by October 1st of this year. And as I understand it, this has to be done quickly because things move slowly in the electricity market, as we have already talked about. It takes time to plan and build new generation, to interconnect generation, et cetera. So we need a head start, okay?
There was a meeting yesterday that you attended, where the stakeholders in PJM started to talk about what reforms need to be made. And those comments that happened yesterday were based upon some initial proposals that came from PJM itself several weeks ago. Tell us, what are some of the initial solutions that are being discussed? What did you see and hear in yesterday’s meeting?
Silverman: Yes, it’s really interesting. I think we’re seeing this desire to move away from the one-in-ten-year reliability standard, and more to talk about expected value of unserved energy. So what’s the difference there? Unserved energy is supposed to put a magnitude calculation into our future reliability planning. Because a one-in-ten-year standard mostly talks about how often an event is going to happen, but then doesn’t really talk to us about, doesn’t really inform how severe that event is. So I think there’s a real trend and interest in moving to an expected unserved energy calculation, which looks at the magnitude. How severe is the event? How many customers were actually affected, if we do lose the system due to reliability event?
There are a lot of different proposals, and they get super wonky. But one of the really interesting ideas is this focus on winter reliability, because I think we’ve all seen — whether it be the Storm Uri in Texas or Elliott or the polar vortex — the grid is increasingly getting stressed during the winter. When I was first in this industry, the focus was really on those hot summer days.
Stone: When everybody has A/Cs running, right?
Silverman: When everybody’s got their A/C running. It’s cranked up. You’ve got a huge load on the system. And I think what we’ve actually seen is that most of the vulnerabilities on the grid tend to happen during the winter because of the freezing of equipment and the kind of unexpected impacts of cold weather on coal piles and natural gas distribution facilities — pipelines. There’s been this really switch from focusing on the summer, to focusing on the winter, and recognizing that our existing system overvalues summer compliance with reliability metrics, and undervalues winter.
The clean energy transition is only going to exacerbate that concern because things like solar — they’re great on that hot, peak summer day. Distributed solar takes thousands of megawatts off basically its negative load from PJM’s point of view. And it really helps the system a lot. In the winter there’s less solar. That will be somewhat offset as we bring in increasing amounts of offshore wind and other things. But particularly, that load growth, as we electrify our heating — building heating in particular — you’re going to start seeing more winter load.
And so a lot of us in the industry, when we look into the future, we see this flipping. Winter is now going to become the peak load period and certainly the biggest challenge for reliability. So a lot of the options that PJM was looking at all center around how do we get more of the price signal for being reliable into that winter period?
Stone: So the carrots and the sticks, or maybe the carrots in this case, for reliability become even juicier, right? To incentivize the generators to really be available. I think it’s interesting. I listened a little bit to the meeting, and the Market Monitor, which is the organization, the body that actually monitors the PJM market to make sure it’s run efficiently, et cetera, also brought up this idea that fuel supply should be firm for all resources — meaning guaranteed delivery of fuel, for example, gas plants. Is that something else that would be part of the solution?
Silverman: This is one place where I think I have to respectfully disagree with Dr. Bowring.
Stone: He’s the Market Monitor.
Silverman: He is the Market Monitor. Joe — everybody in the PJM core community calls him Joe. I mean, he is an institution, absolutely brilliant economist who has been a very effective market monitor for PJM now for a number of years. On this issue and several others, I respectfully disagree. It is very difficult to say that you want firm fuel supplies without taking into account the financial impact of that requirement. So what happens when we start putting firm fuel requirements on? One, you’re simply increasing your reliance on the natural gas system. And as we saw during Elliott and a number of other events, there are failures of the natural gas system that occur during cold weather, just like there are for other technologies.
So even if you have firm transportation on an interstate pipeline that is still no guarantee that it’s going to get to you — in particular, and I won’t go into the crazy wonkiness of this. But even if you have firm transportation, gas service, and even if the gas is available, you still have to tell the pipeline that you want that gas generally 24 to 48 hours in advance of your needing it. So that’s wonderful if you sort of anticipate a heat wave coming through during the summer, but it’s much less effective during the winter if temperatures start dropping.
I think the firm fuel requirement is maybe seductive in the sense that, “Oh, if everybody had the right contractual relationships with their fuel supplier, we wouldn’t have these problems.” But I think the cost implications of doing that are going to make it completely infeasible. And I actually don’t think it would get you all that much because we still have these other factors out there that are really compounding the problem.
I will say, and I think one theme you heard throughout yesterday’s meeting was the need to move from sort of annual metrics to some sort of hourly capacity value, such that you’re not just looking at uncorrelated outages across the entire year when you do your capacity purchases, but that you’re actually focusing and looking at what’s available on a cold winter’s day, between the hours of, say, 6:00 and 9:00 a.m. That may require a different shaped capacity product in order to ensure reliability than we traditionally plan for. And I think that’s a really interesting theme that we’re going to see coming out of the PJM process.
Stone: At the risk of making this conversation even wonkier than it already is, one of the interesting things also is the capacity market. The auctions for capacity are three years ahead of when that capacity is going to be needed. So there are some reasons that are not exactly the case right now, but the market will work that out. But actually I did also hear some proposals that the three-year ahead capacity auction be moved much closer to the date when that capacity will be needed, like a year or 18 months out. Can you talk about what that might offer in terms of reliability, if that’s a solution that might come through?
Silverman: This is an on-going national debate. Different parts of the country have different capacity market structures. PJM, in New England, they do it the same way, has this effectively three-year forward market. And it basically says, “All right, generators, you tell us who’s raising their hand to say that they will be here in three years?” And if not enough people raise their hand, then you go in and contract with new resources to come on and meet that requirement in three years. I happen to think that that is an effective system, or at least that it can be. And particularly, as we start focusing on the clean energy transition, states and other buyers who are really prioritizing clean energy, need to be able to know that the clean energy and the resources needed to back up that clean energy are available three years in the future.
So I actually think there’s a lot of merit to the three-year forward market. That said, New York has a month-ahead market that sort of is a year-ahead market in varying ways, and the complexities are really intense, so I won’t get into them. And you can make that work, too, because you’re sending a very clear price signal that drives forward investment.
In California, they use a combination, a laddered procurement mechanism, where they have one-year procurement, and then they have showings in two and three years. ERCOT, they have no capacity market. All the money is put into this energy market shortage price signal. At the end of the day, if you create a structure where generators see that their total revenue from the energy market, plus the capacity market, plus any ancillary services or renewable energy credit value exceeds their cost of bringing a new resource into the market, they will do so, right? Plus their profit and risk premium.
So I think any one of these various structures could work, but they have to be very carefully calibrated to make sure that consumers are getting the best possible deal for their money, while also sending a very clear price signal to generators who want to make an investment. I tend to think that that three-year forward is about right. As you, one of my major concerns actually with the interconnection debacle — I will use that term — at PJM is that even if prices start creeping up in the PJM capacity market for three years forward, you can’t get a new project through the interconnection process in time to have it actually show up and meet that need. And that’s a real problem because all of our reliability metrics are premised around this idea that new entry could come in and satisfy the reliability need and help discipline pricing behavior, and discipline exercise of market power because, hey, if I’m an incumbent generator, and I put my price up too high, somebody else has three years to come in and undercut me. And that’s a really important facet of the PJM market.
So I’m not saying that one of these proposals to shorten that period couldn’t work. I absolutely think that it could, but it would be a really fundamental rethink of how the PJM market works. It would come with a host of market power and design implementation details that I think are even actually beyond the scope of this current PJM process.
Stone: It’s interesting you just mentioned this interconnection process and what a problem it has been in the PJM market. I just have to note, it’s conspicuously absent, this interconnection issue, from all the discussions that are going on in this fast-track process right now. So you can shore up the reliability of existing resources, but again, if we have so much coal retiring, and we need to replace it with new, optimally clean resources, one of the core problems here is how do we get those resources into the market on time, quickly? And this process really isn’t looking at that.
Silverman: What you say is exactly right, and one of the things about this process that makes me absolutely crazy is there are two things that are missing. One is the interconnection process, and then the second is whether the solutions coming out of one of these new capacity market designs meet and are consistent with state clean energy policies.
So let’s talk about the interconnection for a second. As we look at the Inflation Reduction Act and the financial incentives in clean energy that are coming in through the IRA, we should be seeing just an off-the-charts, massive deployment of capital across this country. And the interconnection queue is one of the largest non-technical barriers to getting that money spent and getting clean energy deployed. As I think about my daughter and the future of the planet, we cannot lose the fight against climate change because we can’t get interconnection studies done. That is simply unacceptable.
And yet we see this. It’s not just a PJM issue. There is a very good study saying that we have 2,000 gigawatts — that’s a huge amount of generation — currently sitting somewhere in an interconnection queue across this country, and that projects are taking on average five years to get through the interconnection process. This is in addition to permitting and actually building the darned project. So we have this fundamental problem that the system that we design to connect new generation is not adequate to the clean energy transition.
I respect that PJM didn’t necessarily want to tackle that huge problem while it was also doing the sort of more mundane pieces of the capacity market reform. But it misses this key component of getting this done. And even PJM’s own study looking at that 2030 scenario says that if we could get enough new generation hooked up, it largely alleviates the worst of the reliability concerns. And so we just have to, as a society, figure out how do we break through this really artificial barrier, where we cannot get stuff built? And we’ve just got to build it, and we’ve got to get it connected. And so interconnection reform is, to my mind, just a huge issue that is left unaddressed through this process.
In fact, I’m not the only one saying this. There was a coalition of environmental groups who wrote to the Board of Directors of PJM in response to their initial letter and said, “Where is the interconnection reform?” Because if you’re telling us — and particularly when I was in my seat as a state regulator — here you have PJM telling states and everyone else, “We need more resources on the grid.” And so when the state shows up to bring new resources, whether it’s storage, which is a very effective capacity resource, or any other technology, including fossil technologies, and yet you go to get in line to solve this 2030 reliability problem, and PJM says, “Oops, sorry. There’s no room at the inn, so come back and see us in five years, and we’ll have a study for you.” That is a fundamentally broken piece of this equation and one that I would really like to see PJM — and frankly maybe the leadership needs to come from FERC or Congress to step in and say, “We need really fundamental reform. We need to build out the transmission system even ahead of generators showing up, so that we’re not — we don’t just have this five-year delay in everything we’re trying to do.”
Stone: PJM has a lot of committees and groups and task forces that are established to address/tackle/find solutions to various challenges that exist in the market. And one of these task forces is called “the Clean Attribute Senior Task Force.” And it’s looking at how clean energy market solutions may address the capacity market problems we’re seeing. Could you talk a little bit about this task force? How is it coordinated with, or not, or aligned with the overall process that’s going on around resource adequacy in PJM right now?
Silverman: Yes, this gets to this question that we’re asking our markets to solve. The CAPSTF, as it’s affectionately known, is about how do we make sure that whatever solution for reliability and economics that PJM markets develop also meets state clean energy targets and private demand for clean energy? So what we effectively have is right now, we have clean energy investment incentives through renewable energy credits for tax policies that are showing one set of price incentives and driving investment in clean energy. And then you have PJM, sort of sitting over on the other side, continuing to run its markets effectively in exactly the same way, to identify reliability solutions and resource adequacy solutions that continue to look only at economics and reliability.
So the CAPSTF is all about how do we actually get our markets to ask the question that society needs markets to answer? What is the lowest cost suite of resources that maintains system reliability and meets clean energy demand? In a lot of cases, that’s coming from states that are mandating clean energy requirements. In a lot of cases, it’s coming from private individuals or corporations that are asking for clean energy.
And so the CAPSTF is all about trying to get the market to help us answer that question. And so I think if you look again at what the — this is using acronym soup — but the CIFP, the Critical Issues Fast Path process is talking —
Stone: That’s the main thing we’ve been talking about today.
Silverman: The main thing we’ve been talking about is asking the question: How do we make the capacity markets work better? But they completely ignore the fact that we’re asking the markets to come up with a reliable solution that has to meet state laws and that has to meet voluntary demand for clean energy. And so if I have a critique of the PJM process thus far, it’s that the questions about how we bring clean energy into our long-term planning processes need to be front-and-center in all of these various processes — along, of course, with the interconnection stuff we talked about before.
Stone: Okay, so the ten million-dollar question at this point is we’ve got this process, the Critical Issue Fast Path, that is supposed to deliver some sort of proposal to the FERC by October 1st. Given all of these moving parts, given that we’re not looking at interconnection issues or that PJM and the stakeholders really aren’t focused on that at this point — given the inherent messiness of the PJM stakeholder process, do you see a solution reasonably coming out of this process by October?
Silverman: I think the answer is going to be yes. Now to me, the question is actually less about whether they — and you know, it may be delayed by a month or two — but in FERC and PJM world, October, November, and December — it’s all effectively the same. The question is, is it going to be, on a 10-point scale, where 1 is status quo, and 10 is the most wonderful market design you’ve ever seen, that meets every state’s clean energy policies, that ensures perfect reliability and puts a unicorn in every kid’s bedroom — where is it going to be? Is it going to be at that 1 level or at that 10 level?
My sense is where they are right now, and given this incredibly tight timeframe that they’re working on, I think they’re going to be shooting for that 3 to 5. So something that is a definite improvement over the status quo but leaves a lot of issues left to be answered for future proceedings. And it’s not entirely clear what happens when this stuff gets to FERC, because these are very controversial issues. I really worry that if PJM comes up with a bunch of capacity market reforms that completely ignore the issue of whether the reliability solution they’re developing is simultaneously feasible with state clean energy policies, you set up this very interesting dilemma for FERC, where they’re looking at potentially approving a process that is not taking into account state laws. And I think that will be just a fascinating development to see what happens.
But yes, I would say given the focus and the need to address these things quickly, we will. And just to put it into context, remember: The reliability issues that PJM is identifying come up in the 2028, 2029, 2030 and beyond timeframe, which is both five to seven years from now, so it seems like a long time. But it’s also incredibly quick, if we need to be readjusting our deployment of capital, because generation resources, even ignoring the interconnection queue, take a long time to get built.
So if we need to get state laws changed, if we need to get investment signals changed, if we need to have private capital responding to those investment signals, it is absolutely critical that we set up that paradigm as soon as we possibly can.
Stone: Abe, thank you very much for talking.
Silverman: Oh, my pleasure. Thank you very much for having me.
Stone: Today’s guest has been Abe Silverman, Director of the Non-Technical Barriers to the Clean Energy Transition Research and Policy Program at Columbia University.