S&P Global Ratings last discussed the PJM capacity market in July 2022, when the auction for delivery year 2023-2024 returned what even we-–a rating agency primed to be relatively conservative--considered low clearing prices. Constantly changing rules added unpredictability to an already complex market and exerted downward pressure on prices. We particularly viewed the revision to the market seller offer cap as constraining price formation. Other factors influenced that auction outcome, including the use of historical rather than forward-looking energy and ancillary services revenue offset, and the first application of the effective load carrying capacity (ELCC) for determining renewables capacity value.

There was bidding behavior at play as well. We note that, unlike physics (which is governed solely by laws), finance and economics are also influenced by the vagaries of human behavior. While gravity cannot be arbitraged away due to popularity, aversion to being shut out of an auction can result in irrational bidding. Because rules were still evolving and the timing of future auctions was uncertain, no participant was willing to bid its capacity at meaningfully higher levels, even if the generator believed prices should clear higher, for fear of being shut out altogether. So the outcome of the December 2022 auction for delivery year 2024-2025 was no better. It was evident the market needed reforms, and some have been implemented over the past 12 months.

The fact that low prices were not incentivizing new build was not seen as a crisis because the PJM then considered itself well supplied with resources. However, as a natural consequence, low prices were initiating retirements.

Then, the notion of being well supplied disappeared overnight with Winter Storm Elliott in December 2022. PJM itself noted the growing shortage in a February 2023 report where it explored a range of scenarios through 2030. PJM suggested that as much as 40 gigawatts (GW) of existing generation was at risk of retirement by 2030, which included about 25 GW of potential policy-driven retirements and 3 GW of potential economic retirements. Combined, this represented over 20% of PJM's installed capacity. PJM also remained concerned with the slow pace of new generation construction, noting that about 38 GW of resources had cleared its interconnection queue but had not been built due to external challenges, including financing, supply chain, and siting/permitting issues.

Since the last auction, the emerging dynamic of a demand surge as well as the need for decarbonization also dramatically tightened demand and supply.And so, fundamental demand/supply shifts, evolving market rules relating to the implementation of accreditation that more accurately values each resource's contribution to reliability, and behavioral economics have been influencing this latest auction. Put all of this against an increasingly steeper downward sloping demand curve (see sidebar 1) and it becomes apparent why one has to be extremely bold, or foolhardy, to forecast prices. No amount of modelling sophistication will allay the fact that all our knowledge of the capacity market is about the past, while all our decisions have to be about predicting prices of an increasingly complex market that is still revising rules.

Stated differently, capacity market forecasters are now just guessing.

We believe expectations are healthier than forecasts because they provide a vision of the future stripped of all false precision. We also believe mean-reversion is not just a statistical phenomenon, it is an economic necessity. Therefore, we tend to think in terms of long-term, mean-reverting capacity prices. Last year--when prices were $28 per megawatt-day (/MW-day)--we believed PJM's regional transmission organization (the RTO-wide pool) capacity price would eventually revert to the $100/MW-day area.

Because of constraints across the eastern load deliverability areas, we expected the Mid-Atlantic Area Council (MAAC) would mean-revert to the $120/MW-day area and the Eastern MAAC to $150/MW-day. Given the expected surge in demand from large loads and the lack of concomitant growth in supply, we now believe capacity prices will be higher for at least the next two years before transitioning to a fundamental, long-term price. Our view of what the mean-reverting price levels could be has also moved upward.

Below are our updated assumptions.

Table 1

In the following sections, we present our views on factors that drove this auction's outcome. We also examine the basis for our revised capacity price assumptions.

Even after this high print, capacity prices are about 25% of aggregate supply-related costs (see section on complementary nature of energy and capacity prices). However, we note that given the complementary nature of energy and capacity prices, we will not see capacity prices in isolation.

As the market provides further information (scenario analysis around the 2025-2026 results, etc.), we will make adjustments to either capacity prices or energy margins if we believe it's appropriate.

The Latest Auction Results Were A Surprise

The high auction price strip for 2025-2026 is the proverbial Christmas gift delivered by a winter storm, just recognized 19 months later. Prices for the majority of PJM rose to $269.92/MW-day, compared with $28.92/MW-day for the previous auction. Cleared prices were even higher for Baltimore Gas and Electric Co. (BGE; $466.35/MW-day) and Dominion Energy Inc. ($444.26/MW-day) zones. In the BGE region, in-zone capacity dropped to 600 MW from 2,600 MW with the loss of the Brandon Shores and Wagner plants, which moved out of the capacity market as they transitioned toward retirement.

Dominion returned to the reliability pricing mechanism (RPM) capacity market from the fixed resource requirement (FRR) approach to avoid increasing penalties. A higher reliability requirement in the auction relative to the FRR process and demand growth in the region led to a supply shortage in that zone.

To us, the significant price increase is somewhat surprising but not unexpected directionally. The short explanation is a supply and demand imbalance that has worked itself up given no material recent investments. However, a longer line of reasoning also includes long interconnection queues. Critics also place blame on PJM's interconnection process--the PJM shut down its interconnection queue until 2025, which they claim is a result of its complex process. They believe there was insufficient future transmission planning that has fallen behind and now the problems have simply cascaded.

Fixed-income investors want predictability foremost. Although we remain disinterested observers of policy, we are not uninterested observers because policy often affects credit quality and profiles. When prices in the auction market go to a high of $270/MW-day from a low of $28/MW-day from one auction to the next, it underscores the inability of the auction to provide appropriate economic signals in a timely fashion.

So What Happened In The Auction?

While there are several factors that contributed to higher prices, Winter Storm Elliott was the most significant. It allowed PJM to acknowledge the need to factor in accreditation values for reliability more broadly.

We discuss that ELCC revision and other the major drivers below.

Peak demand and reliability requirements

The preliminary forecast peak load increased to 153.8 GW from 150.6 GW, mainly due to growth expected from data centers and related industries. We note the ELCC-driven reduction in the reliability requirement (see Appendix) is offset by a reduction in accredited capacity and an increase in demand curve pricing.

Table 2

More important, reducing the reliability requirement shifts the variable resource requirement (VRR) curve to the left (i.e., down in quantity; see Appendix). The VRR price points also increase because net cost of new entry (CONE) is tied to the ELCC of the reference resource. Specifically, the net CONE is converted to $/MW-day (unforced capacity [UCAP]) using the ELCC class rating of the reference resource rather than using the poolwide average effective forced outage rate (eFORd). This conversion results in a higher net CONE. The higher net CONE is important because the maximum capacity market price under PJM rules is 1.5x the net CONE value, which also contributed to the higher clearing price.

Installed reserve margin (IRM)

The PJM increased the IRM requirement 3.1% due to higher load shape variability than historical years. We estimate this will increase the supply requirement about 3.8 GW. In addition, the total FRR obligation--where an eligible load-serving entity meets FRRs with their own capacity resources--added another 10.9 GW. The total procured capacity in the auction, combined with resource commitments under the FRR, represents an 18.5% reserve margin. This is only 0.7% in excess of the target reserve margin of 17.8%, highlighting the rapid erosion of excess reserve margins compared with the 2024-2025 base residual auction (BRA), which cleared at a 20.4% reserve margin.

A more concerning factor for the RTO is that the amount of supply resources in the auction fell yet again this year. Supply offered into the RPM capacity market declined more than 13 GW, falling to 135.7 GW in the 2025-2026 BRA compared with 148.9 GW in the 2024-2025 BRA. This is the fourth BRA in a row in which the total capacity offered from nonenergy efficiency resources declined. The total amount of capacity in the RPM that cleared decreased to 134.7 GW in the 2025-2026 BRA from 140.4 GW in the 2024-2025 BRA.

The return of Dominion zone

For utilities to use the FRR, they need to notify PJM with their intention to abstain from the capacity auction with a minimum four months' notice and provide their FRR plan with a minimum one month's notice ahead of the next auction. Furthermore, the FRR is a five-year commitment to stay out of the capacity auction, and utilities must maintain at least the same reserve margin as the ISO. The most recent reserve margin established in the 2025-2026 capacity auction was 17.8%.

Given its rapidly rising demand growth from datacenters, the increasingly capacity-short Dominion zone returned after three years to the RPM capacity market (from the FRR approach) to avoid paying increased penalties. This also pressured prices.

ELCC accreditation

Due to outages during Winter Storm Elliott, a technical change was introduced relating to the amount of capacity that generators could bid in. The FERC approved changes related to the ELCC that were designed to reflect the true availability of resource types. This was done largely to address reliability inadequacies seen during the storm, when a significant portion of PJM's gas fleet experienced outages due to a lack of winterization.

For the latest auction, PJM has accredited a poolwide average UCAP of 79.69%. The UCAP represents the installed capacity adjusted for weather conditions. Effectively, the PJM assigned a 20% eFORd to the system, compared with the 5% eFORd % in the previous auction. This reduced the amount of supply that can be bid in. While the accreditation better aligns incremental market pricing with resources' incremental reliability value (and exerts upward pressure on prices), load now pays only for UCAP within PJM's market construct (i.e., (1+IRM) x UCAP). Notably, even nuclear capacity did not fully clear--its accreditation was 95%.

That said, the capacity accreditation reduces the total supply of UCAP. If a unit's UCAP decreases under the new accreditation approach, its fixed unit costs will not. It will still need to recover the same total dollar amount for the equivalent offer under the new accreditation. For instance, consider a 10 MW UCAP resource that previously bid in at $50/MW-day, reflecting costs of $500/day. If it's now accredited at a 5 MW UCAP, the offer would be adjusted to $100/MW-day, resulting in the same total $500/day amount. Thus, moving to a marginal ELCC capacity accreditation approach shifts supply curves up and to the left (i.e., up in price and down in quantity) relative to the supply curves seen under the former capacity construct. The accreditation methodology affects natural gas plants' and solar generators' UCAP the most.

Retirements Have Continued Even As New Capacity Additions Have Struggled

We believe the surge in prices across most of PJM's footprint is spurred primarily by decreased supply offers into the auction due to next thermal installed capacity changes from generator retirements. About 6.6 GW have retired or have must-offer exceptions (signaling intent to retire) compared with the generators offered in the 2024-2025 BRA. The retirements were a result of policy rules that included New Jersey's carbon dioxide rules, Illinois' Climate and Equitable Jobs Act, coal combustion residuals, and effluent limitation guidelines. In comparison, new capacity additions were only 864 MW, which mostly included uprates (754 MW). We believe most of the shift up the supply curve occurred because of the removal of this retiring capacity.

The Reliability-Must-Run (RMR) Units Were Significant To This Auction Result

The must-offer exceptions include Talen’s Brandon Shores and Wagner plants, which were withheld from the auction given their ongoing RMR proceeding. Whether a RMR unit participates in the capacity market depends on the unit's specific arrangement with PJM. So far, while PJM has designated RMR units as needed for grid reliability purposes, it has not required them to make commitments in the capacity market. PJM can dispatch and schedule Brandon Shores and Wagner for reliability purposes, subject to their specific operational restrictions.

Table 3

RMR units participating as energy-only resources likely affected the auction price formation. In fact, a Maryland Office of People's Counsel study concluded that had the Brandon Shores and Wagner units remained as supply-side resources in the capacity market, it would have shifted the supply curve to the right and moved the demand and supply curve intercept point down and to the right, resulting in a lower clearing price by as much as $100/MW-day. To be clear, that analysis assumed these units would have bid $157-$165/MW-day in the auction, which is Wagner's current RMR ask and more than double its offer price in 2024-2025. The capacity price in the BGE locational deliverability area (LDA) in 2024-2025 was $73/MW-day, and most of Brandon Shores' and Wagner's MW cleared as capacity resources in that delivery year.

The futility of attempting to pin down numbers in a forecast becomes apparent when 2 GW of generation can affect the outcome of the entire RTO-wide auction this meaningfully.

What Does This Mean For Generators' Credit?

Broadly, all generators benefit, but the impact is muted for some nuclear operators because of state or federal subsidies that already provide an uplift (or conversely do not provide an uplift because of a regulatory compact to serve load at predetermined prices in exchange for a floor price protection).

The price outcome was the most beneficial for Talen. The company ended up clearing 6.8 GW, resulting in $670 million of annualized capacity revenues. More important, this excludes the Brandon Shores and Wagner RMR plants in the BGE zone. In the last deliverability period, the company's capacity revenues were only about $180 million, inclusive of the RMR plants. We also note Talen has requested $378/MW-day for Brandon Shores and $157/MW-day for Wagner in its RMR price negotiations, compared with the BGE zone clearing at $466/MW-day.

Vistra cleared 10.25 GW at a weighted-average clearing price of $273.45/MW-day, aggregating more than $1.0 billion in capacity revenues. Therefore, we expect about $6.0 billion of EBITDA for Vistra in 2026. In PJM's previous capacity auction, Vistra cleared 6.95 GW at a weighted-average clearing price of $43.25/ MW-day, equating to about $110 million in capacity revenue for the 2024-2025 planning year. We note that Vistra bought Energy Harbor in February 2024which owns about 4 GW of nuclear capacity in PJM.

Constellation Energy cleared 17.5 GW in 2025-2026 compared with 18.725 GW in the 2024-2025 auction. Of note, the company cleared 600 MW less nuclear and 625 MW less non-nuclear in the 2025-2026 auction compared with 2024-2025. The nuclear reduction is due to PJM's new accreditation system, which affected all generation types. The UCAP for nuclear was 95%, which led to lower cleared MW. The higher capacity prices apply to about half of Constellation Energy's PJM fleet, but with an offsetting nuance around the Illinois carbon mitigation credits and production tax credit floor. We see no upside to 6.2 GW of capacity under the carbon mitigation credit plan but potential upside for the remaining nuclear fleet (9.35 GW), especially those in the eastern PJM zone (5.8 GW). The 1.95 GW of non-nuclear capacity has unrestricted upside.

We expect PSEG Power to clear about 3.7 GW. Despite the PTC floor, we forecast a modest uplift to its EBITDA because its nuclear units experience a PECO basis discount relative to prices in Eastern MAAC.

What Does This Mean For Future Auctions?

Yesterday, PJM published parameters for the December 2024 RPM auction for delivery years 2026-2027. For this delivery year as well, demand and reliability requirement are higher than the July 2024 auction by 3.3 GW and 2.8 GW, respectively, and installed reserve margin is now 18.6% compared with 17.8%. The poolwide accredited UCAP is also modestly lower at 78.98% from 79.69%. All these revisions point to higher price outcomes, and there is really nothing in the latest parameters that lead us to expect prices to recede materially from the 2025-2026 levels. However, as a rating agency, we tend to be adverse selective in our assumptions.

We believe the following reasons will likely result in elevated prices that eventually mean-revert, but to a higher level than we previously expected.

Time to auction

Some reasons why this auction priced higher are apparent: the market was overreacting to low prices of the past three auctions, as well as to the uncertainties of constantly evolving rules. But another reason was simply that the PJM got itself off its three-year schedule. The primary purpose of the auction is to provide visibility (price signals) that incentivize new build when needed for reliability. That can't happen unless there are about three years between auction results and in-service dates that allow for financing to be obtained and construction completed.

PJM is trying to get back to its three-year forward schedule by May 2026, but there will still be three more auctions before it does so (December 2024, June 2025, and December 2025). This leaves minimal time for new capacity to affect outcomes or for load to respond. It also does not provide time for any rule changes to sway price outcomes. In our view, it's possible the 2026-2027 auction price will clear at high levels (and potentially higher than our assumptions).

There is no meaningful incremental dispatchable build that can bid into the next auction

It is not possible to bid in a combined-cycle gas turbine (CCGT) into the December 2024 auction and bring it online by mid-2026 unless it is already under construction. We believe the only new build that can potentially bid into the 2026-2027 auction is Clean Energy Future's (not rated) Trumbull CCGT (940 MW) being built in Ohio.

High prices signal the need for new generation but do not guarantee it

This situation reminds us of the Electric Reliability Council of Texas' (ERCOT) generation supply in the wake of the winter 2021 storm, referred to as Winter Strom Uri. Energy prices hit the systemwide offer cap of $9,000 per megawatt hour (/MWh) for three continuous days but provided little incentive for generation. There was simply no generation available to kick in. Price signals provide appropriate incentives only when there is generation available to respond.

The significantly higher capital costs now exceed $1,500 per kilowatt (/kW) in some eastern locations but still average $1,250/kW in the broader footprint for CCGT builds are deterrents for generators to commit funding In addition, new Environmental Protection Agency (EPA) rules require new CCGTs to sequester carbon emissions by 2032 or operate below 40% capacity factors. These factors provide little incentive for sponsors to contemplate baseload build unless there is longer-term price visibility at higher levels.

We believe more peaking units will come online in the near term, and that peakers now have capital costs of $900-$1,000/kW with 12-15 month construction time frames. Finally, even if investors are somehow convinced to finance projects on the back of merchant energy revenues, new entrants will also be delayed by the lengthy interconnection queue. The PJM queue is dominated by resources with low accredited capacity value. In the near term, we believe PJM's main focus will be to incentivize higher ELCC resources and improve queue efficiency.

Recognition Of Accreditation Underscores That All MW Are Not Equal

The effective revision of eFORd to 20% from 5% is true for the system as a whole, but some resources were more insulated than others (nuclear, coal). The ELCC accreditation for solar is 8%-13% and onshore wind is 34%. While battery storage has been assigned an ELCC of 78%, it is applicable to a 10-hour battery (effectively 32% prorated for a four-hour battery).

We believe weak outcomes in future auctions could have implications for the PJM market because market participants can now consider alternatives. For instance, unregulated nuclear units could withdraw from PJM's capacity auction by participating in an FRR approach, assuming a load-serving entity initiates the withdrawal of load from the capacity auction and contracts with that nuclear generator bilaterally to include in its FRR plan (i.e., to fulfill the capacity obligation for this load via FRR). Should more nuclear units commit to a similar transaction, the tighter the rest of the PJM market becomes. We note that this leaves little time before the December 2024 auction for 2026-2027 deliverability for any large load to procure resources as part of the FRR to meet existing and future demand.

Separately, but related, PJM recently issued its critical issue fast path proposals that include moving to a seasonal (summer/winter) capacity market construct as well as improving capacity accreditation to better reflect resources' contribution during risk periods.

The Complementary Nature Of Energy And Capacity Markets

Traditionally, new CCGTs bid into auctions to lock in capacity revenues to obtain financing. However, because of recent auction delays, energy revenues have become a larger proportion of a plant's levelized cost of energy revenue requirements. In theory, it does not matter how a plant is compensated for its return on (and return of) capital. However, capacity prices are more visible and therefore considered more "bankable". As a result, sponsors and lenders typically prefer capacity to comprise 25%-30% of total gross margins.

That said, we want to put systemwide costs to load from these capacity prices into perspective. If we were to convert prices from $/MW-day to $/MWh, capacity costs will still be a small proportion of the bill. For instance, for a 90% capacity factor nuclear unit, the $270/MW-day price translates into $12.5/MWh (and $15/MWh for a CCGT operating at 75% capacity). That is still about 25% of aggregate supply-related costs.

We note in this auction, capacity prices are above net CONE in all regions. Therefore, on average, an investor would expect to be made whole for this delivery year without other nonenergy revenues. In fact, in the BGE and Dominion zones, the price egregiously exceeded gross CONE. That means a new plant is being made whole (to be clear, just for that year) even without energy revenues. Coincidentally, energy prices are already starting to decline after the results of the auction.

However, if this "above-market" signal cannot incentivize new build, it would imply that sponsors are either unwilling (e.g., because of perceived risks related to carbon sequestration) or unable (prolonged delays in interconnection queues) to respond. That would mean that the market is truly broken and demand would be forced to respond (i.e., the AI/datacenter growth narrative will temper and growth will depend on whatever supply the power markets can provide).

There Are Reasons For Prices To Decline

Prices could retreat in the near term, but we believe this would occur from a political backlash or regulatory intervention. Longer term, there are other reasons that we list below.

Regulatory and political response

We note systemwide capacity costs to load will climb to $14.7 billion, the highest they have been (the previous high was $10.9 billion in the 2018-2019 auction). Notably, this cost is meaningfully higher than the $2.2 billion in the previous two auctions. Therefore, we believe there will likely be some political backlash and court challenges given customer price impacts. The high prices will likely drive stakeholders to advocate market changes and possibly out-of-market interventions.

ELCC accreditations could increase

This factor was included in the upsides as well, but we note it here because it has the potential to be a double-edged sword.

The ELCC value was determined for the July 2024 auction based on 20 years of historical performance during periods of system stress. PJM intends to use a rolling 20-year average moving forward. In fact, it has already released preliminary ELCC class ratings for 2026-2027 through 2034-2035. These ratings are nonbinding and are for indicative purposes only but do show improvement for certain asset classes (CCGT's increase to 85% from 79%).

It's important to recognize that even as UCAP has declined about 20%, the installed capacity doesn't actually disappear. The entire installed capacity is available to bid into the day-ahead energy market, so the preliminary accreditation can change. PJM considers enhancing its accreditation approach to ensure the reliability contribution of each resource is accurately determined. Based on the nature of the calculation noted in the ELCC section above, specifically its use of historical performance data, it would seem that it takes multiple years of significant overperformance across the fuel types to move the average that goes into the calculation. However, we believe this will eventually occur (the New York ISO often tweaks its locational capacity requirement parameter for zone J). Should the ELCC accreditation value be revised upward in later years, we expect that to weaken prices.

The steepness of the VRR curve

We placed the steepness of the VRR curve in the downside risk instead of the upside for the following reasons:

• A seamless market should return one RTO-wide clearing price, implying no physical or structural/regulatory constraints across the grid. However, past PJM auctions have invariably resulted in price breakouts in several regions because of issues ranging from transmission congestions to higher compliance costs in regions dominated by fossil generation. This auction ended up with the same price except for two regions, where either supply is substantially short (BGE: retirement of key units) or is short because of out-of-control demand (Dominion: the datacenter pull).
• PJM rules are such that there is a zero-sum outcome: assets that clear at a higher price in an LDA have to bid in at $0 for the broader RTO zone, making the RTO supply curve more convex--and resulting in lower prices. So typically, the breakout in one LDA lowers the price of the RTO.

However, something quite different happened in the 2025-2026 auction. We've been watching Maryland demand/supply for some time. With Brandon Shores and Wagner out, the BGE zone is now so short that it effectively borrowed supply from the RTO, pulling the entire pool's price up. This was the perfect storm--the decline in unforced supply and the increase in installed reserve margin brought the supply curve to the steepest portion of the VRR curve. Conversely, a slight increase in supply would also make clearing prices work their way down the curve.

What Does This Mean From A Credit Perspective? Takeaways For Market Participants

Capacity price forecasting is often difficult or ambiguous. That said, market observers can probably get capacity moves directionally correct.

Issuers (or their bankers) will likely be tempted to releverage existing transactions or increase the debt proportion when capitalizing new transactions in the belief that prices are higher for longer. They might well be right. But because S&P Global Ratings' analysts have seen these types of fluctuations before, we believe prices always mean-revert. Some sponsors take distributions when prices move up but rarely inject equity when prices decline. We view favorably the ability for an entity to deleverage before paying distributions. A track record demonstrating the ability to pay down debt matters, especially for single-asset financing. We will differentiate between financings that show a track-record, compared with others that releverage on the back of one strong year in our ratings assessments.

Term loan Bs (TLBs) were introduced in industries with volatile cash flows, such as merchant power, as a self-help approach to financing. The premise of a TLB is to swap default risk for refinancing risk. TLB agreements often include a cash flow sweep mechanism that requires projects to apply excess cash flows to prepay debt after annual project-related costs are covered. Given that power prices are subject to varying economic cycles, TLBs typically reach their maturity date with a debt amount that needs to be refinanced (ideally less than 50% of the original amount), and this is often refinanced with another TLB.

In the current interest rate regime, lenders are earning high yields and appear to be happy doing so. They have made covenants lighter, lowered and slowed excess cash flow sweeps, and, in some instances, waived excess cash flow sweeps in their entirety. Given higher auction outcomes, should transactions be leveraged on the back of higher price assumptions, the lack of excess cash flow for the benefit of debtholders could be deleterious for credit quality. We think a term loan B should do its job it was intended for-—substitute default risk with refinancing risk. It should not become a 'loan to own'

Finally, when policy starts affecting credit quality, we do take note. Debt investors require predictability that the PJM capacity auction has not lately provided. In fact, we have noted in the past that, all else being equal, we typically rate projects in the PJM higher than those in ERCOT because of the three-year forward capacity market construct that provides visibility (see attached research). Should capacity auction outcomes remain volatile, that may no longer apply.

Appendix

Basics of a capacity construct

In a capacity construct, we plot a downward-sloping demand curve, also called a VRR curve, against peak projected demand in MW (as a percent of peak load). At its starting point, generation capacity supply is below demand. In this hypothetical region of generation supply, reserve margins are at or well below a target level, and the demand curve provides additional incentives for new plant construction, typically at about 150% of the net CONE. The demand curve starts dropping only after the demand/supply situation reaches target reserve margins, and it continues its measured decline as reserve margins move from scarce to adequate and eventually to surplus levels. Thus, the downward slope of the curve represents the value of resources beyond the installed reserve margin.

Along the y-axis, prices for supply offers are lined up (these are bids that merchant generators make for supplying power). This supply curve is convex in shape as bids increase for providing incremental generation megawatts. The point at which the supply offer curve intersects the demand curve becomes the auction clearing price.

Salient capacity market terminology (for context) includes but is not limited to:

• Gross CONE: Revenue requirement on a levelized basis that supports new generation. This is the annualized total cost of new generation capacity.
• Net CONE: The measure of revenue deficiency between the cost of the marginal unit and what it is unable to receive in the energy markets. An asset should not "double dip" in the energy and capacity markets, and the net CONE calculation indicates that the asset has not. It equals gross CONE minus the energy revenue offset. Because of this offsetting feature, capacity prices become linked to energy markets. However, an issue that has been switched about some in past auctions is the reference years to be used for calculating the energy margin offset (historical or future years).
• Avoided cost rate (ACR): Because of market power limitations, generators are usually capped at bidding their ACRs into the auction--i.e., fixed O&M costs net of energy revenues. In earlier auctions, these could be adjusted to reflect capital costs, including environmental retrofitting costs (but now those are bid in energy prices). As a result, 80%-85% of the bids in an auction essentially become noncompetitive bids (i.e., price-taking bids), and only the remaining 15%-20% of the highest-cost marginal units eventually influence the capacity price. Generators can certainly bid an amount lower than their ACR, which often happens if they wish to ensure their capacity clears in the auction.

Related Research

• Capacity Market Assumptions For Power Corporate And Project Financings, Aug. 28, 2024
• Evolving Term Loan B Market For U.S. Merchant Power Industry, Sept. 28, 2023
• Power Markets Update: PJM's Capacity Market Auction Lays Another Egg, July 12, 2022
• Power Tennis, Or Power And Tennis: Serving Up The Limitations Of An Energy-Only Market, Sept. 9, 2019