Coop's Energy Transition Blog

The SetupPre-disclaimer: You know how every year in school growing up, your math teacher would teach you some rules, and then the next year you’d learn those weren’t hard and fast rules as much as general principles that might not always be true? “You can’t take the square root of a negative” in Algebra I becomes “the square root of a negative number is a sorta-half-negative-thing that you just have to imagine” when you get to Algebra II, and you just go, huh, my teachers get paid to lie to me. I’m going to take the Algebra I approach in this post, so for industry old heads who might be reading this, please know that I’m speaking in generalities intended to help establish a basic understanding for folks who are newer to these concepts. For example, if you go read the section discussing ISO/RTOs vs. vertically integrated markets and find yourself thinking, “Coop’s wrong! Not all of California is served by the CAISO grid, because SMUD and LADWP are separate balancing authorities in California that are technically part of non-CAISO WECC,” Taking this all into account, a solar 8760 will typically look like a sine wave subject to a zero lower bound, with generation each day starting up around sunrise, increasing until the early afternoon, and then decreasing back to zero around sunset. Maybe you’re generating energy at close to the nameplate DC wattage of your panels for an hour or two, but you’ll be at a lower level of output outside the prime hours, which is reflected in solar’s relatively low AC capacity factor of about 15-30%, vs. more like 35-45% for wind and possibly upwards of 60% for baseload thermal assets like combined cycle gas (CCGT) and nuclear. Capacity factor is defined as the actual annual energy production of an asset divided by its theoretical maximum annual production, i.e. its full capacity (in megawatts) multiplied by 8760 hours. So, if I’m building a solar plant that is sized at 100 megawatts AC, and I expect it to produce 220,000 MWh of energy in a year, that means my annual capacity factor will be 220,000 MWh / (8760 h * 100 MWac) = 25.1%. The higher a solar plant’s forecasted capacity factor, the more attractive it will be to build, all else equal, because you’re getting more production (revenues) out of the same equipment (capex). Of course, all else is never equal, but projected capacity factor [3] is always a key input to renewable project development and investment decision-making. [3] Or production, or specific yield – different ways of measuring the same thing. The Capacity MarketIn the capacity market, designed to meet long-term power supply needs, generators bid whatever price they want into a pool of assets that are committing to be online and available to serve load throughout a given system planning year or set of years. The auction works the same way as the energy market, with bids accepted from lowest to highest cost until the ISO/RTO has procured enough capacity to meet its reliability requirements (more on what this means below). The ISO/RTO is therefore a monopsony – a sole buyer of capacity – rather than simply being a transaction facilitator between buyers and sellers, as it is for real-time and day-ahead energy. But, since ISO/RTOs charge standard fees (called “tariffs”) for the use of the transmission lines they manage, energy users are actually the ones funding the capacity market as well, albeit through the ISO/RTO’s centralized capacity procurement mechanism. How does an ISO/RTO “meet its reliability requirements?” This is generally done by procuring an amount of capacity equal to the system’s peak demand plus its planning reserve margin (PRM), usually expressed as a percentage of system peak It’s a measure of how much “spare” capacity is able to serve load at any given moment. For example, let’s look at that same market with four natural gas power plants. Say the market operator forecasts a system peak load of 2.5 GW with a planning reserve margin (“PRM”) of 20%, meaning it’s required to procure capacity accounting for 120% of the forecasted peak demand, or 3 GW, to provide a margin of safety that ensures the lights stay on even if instantaneous load on the system is significantly higher than expected or a generator trips offline unexpectedly. Now, the 4 power plants bid their capacity in the market at a price per kW-month (they want to get paid this amount for each month that 1 kW of their plant’s capacity is made available to the grid): 1.      Plants A and B (the ones with the lower $/MWh marginal costs and energy bids) bid $0/kW-month, as they expect to make most of their money from selling energy in the short-term market given their low marginal costs, so they don’t need to make anything from the long-term capacity market to stay in the black. 2.      Plant C (with a higher $/MWh cost/bid in the energy market) bids $15/kW-month, as it needs to recover some amount of money from the capacity market to achieve its target returns for investors, since its energy price is less competitive and it will therefore be dispatched to provide energy less often than cheaper plants A and B 3.      Plant D (with the highest energy cost/bid) bids $25/kW-month, as it needs to recover almost all of its target annual revenue from the capacity market because it is so rarely “in the money” (dispatched) in the energy market – it only turns on when scarcity conditions drive very high energy pricing, so it’s dependent on “getting paid just in case it’s needed” via the capacity market. Again, the market operator evaluates the bids in order of price, and here we see the 3 GW reserve margin target is hit at the price of $15/kW-month, which Plants A, B, and C will receive. Plant D does not “clear” the market as its offer is above the market-clearing price. If this happens a few years in a row, Plant D will be decommissioned and will exit the market, since its marginal cost is too high to be competitive in the energy market and its annual revenue requirement is too high to be competitive in the capacity market. Renewables in the Capacity MarketSo what happens to the capacity market when we add the 1 GW solar asset onto the grid along with the four 1 GW gas assets? It depends on what market you’re in, but generally ISO/RTOs recognize that 1 GW of solar does not provide the same capacity to the grid as 1 GW of gas, because you can turn gas on and off whenever you want (with some limitations), while you only get solar when the sun is shining. Therefore, the solar asset shouldn’t get paid for its entire 1 GW size, even if its capacity bid clears in the market (i.e. is less than or equal to the marginal bid that sets the price for the whole market). Instead, ISO/RTOs will do what’s called a “capacity derate.” Typically this is done by multiplying the asset’s nameplate capacity (the theoretical maximum it can deliver to the grid) by the asset’s effective load carrying capability, or ELCC. ELCC is expressed as a percentage of nameplate, and is set by each ISO/RTO to capture different technologies’ (solar, wind, battery storage, natural gas, coal, nuclear) ability to serve load during peak demand periods. This is sort of like a capacity factor, except it measures an asset’s actual production as percentage of theoretical output only during peak demand periods, rather than throughout the whole year. How meaningfully do ELCC-driven capacity derates affect renewable projects’ revenue potential? Well, research from the National Lab of the Rockies (known as NREL in happier times) puts current marginal solar PV ELCC – the value for a new solar project coming online now – at or below 10%, meaning that if you have a 100MWac PV plant connecting to the grid this year, you’ll only get paid for 10 MW worth of capacity. This varies by region: The authors do note that marginal ELCCs for solar PV are higher than 10% in the Great Plains ISO/RTOs, SPP and MISO, where wind is the dominant renewable resource and there is less solar to cannibalize capacity revenues by reducing system net demand (gross demand less variable renewable generation) during sunny daylight hours – meaning that the peak demand periods used to determine ELCC occur when the wind isn’t blowing. Since wind generation is generally stronger at night, that means SPP is more likely to be undersupplied during the day, so the marginal daytime capacity added by solar in SPP is actually significant, for now. SPP PV ELCCs are currently between 56% and 74% for the summer 2026 capacity period, although no one expects these elevated values to persist for long, as substantially more SPP solar will come online within the next few years. The bottom line is that, while capacity revenues are often a big part of the economic equation for dispatchable technologies like natural gas and nuclear, no one really expects solar or wind to make much money from capacity, unless they’re paired with battery storage. So when IPPs try to make a renewables project pencil out, they’re relying on selling energy, often bundled with per-MWh renewable energy certificates (RECs), to provide 100% of meaningful near-term income. Then there’s additional upside when the initial PPA ends and asset “goes merchant,” as its debt will be largely or entirely paid down at this point and it can capitalize on potentially higher spot prices by selling into the liquid wholesale markets, assuming it’s located within an ISO/RTO. Capacity is only a meaningful economic driver for wind or solar projects that are paired with battery storage, since a 2- or 4-hour battery massively enhances the dispatchability of a project relative to a solar- or wind-only asset. What Else?Here’s some stuff we didn’t touch on in this basic overview. Maybe we’ll come back to it in a future blog? Maybe I don’t know enough about any of these things to actually write said future blog? TBD. For now, in case you want to plug these into your AI chatbot of choice (I recommend this one): • The relationship between apparent power, real power, voltage, and amperage • Locational marginal pricing and congestion charges in liquid markets • Why are solar project capacities larger in DC than in AC, and what is inverter clipping? • Why do certain renewable projects (especially wind) bid negative prices rather than zero? Why do they sell energy to the grid in certain negative-priced hours rather than curtailing? How do these questions relate to the federal Production Tax Credit (PTC)? • Seasonal and monthly capacity pricing • What does it mean that the ERCOT and AESO markets are “energy-only?” How do these grid operators meet their reliability and reserve margin targets without a capacity market? • What does it mean that the SPP capacity and CAISO resource adequacy markets are “bilateral?” How do these grid operators meet their reliability and reserve margin targets without a capacity auction? • The detailed mechanics of virtual PPAs and the concept of basis risk • Where does battery storage fit into all this talk of capacity factor and ELCC for renewable projects? That’s all for now! Thanks for reading. -Coop Standard Disclaimers• This blog post is written in my personal capacity and reflects only my own thinking, research, experience, and opinions. • This blog post is not sponsored by, endorsed by, or affiliated with my employer, although its content may be informed by some non-confidential aspects of my work. • This blog post is not investment advice and does not constitute any offer, solicitation to offer, or recommendation of any investment product or security. • This blog post is 100% human-researched, drafted, and edited. AI was not used in any capacity. • Want to weigh in with your thoughts? Propose a correction or improvement? Find me on LinkedIn or shoot me an email: cwetherbee at alumni dot gsb dot stanford dot edu.