Capacity's Broken Promise

PJM cleared 144,000 MW of capacity for the 2024/25 delivery year, implying a comfortable 22% reserve margin above projected peak demand. On paper, the math worked. In the June 2025 heat event, it nearly didn't.

Actual available capacity fell to 139,000 MW as temperatures climbed. Forced outages removed 8,400 MW of cleared capacity from service. Demand response underperformed by 600 MW. Temperature-induced deratings cut another 3,100 MW. The reserve margin that appeared adequate in auction results nearly vanished when it mattered most. Load shedding was narrowly avoided.

The episode crystallized what grid operators have suspected for years: capacity markets are paying for the wrong thing.

The Physics of Failure

The core problem is methodological. Current accreditation credits resources based on nameplate capacity minus historical forced outage rates. The approach assumes past performance predicts future availability—an assumption that fails precisely when reliability matters most.

Forced outage rates during stress conditions exceed historical averages by two to four percentage points, according to RTO post-event analyses. Gas plants derate as ambient temperatures rise. Demand response participants fail to curtail when their own cooling loads spike. The resources that clear capacity auctions aren't the same resources that show up during emergencies.

Where 12,100 MW of Capacity Vanished

Where 12,100 MW of Capacity Vanished. Source: PJM June 2025 heat event data. Forced outages caused 69% of the capacity gap between cleared and available resources.

CAISO's Effective Load Carrying Capability methodology offers an alternative framework. ELCC calculates each resource's marginal contribution to reliability through probabilistic analysis, weighting performance during loss-of-load-probability hours. Under this approach, four-hour batteries receive 85% to 95% capacity credit reflecting consistent availability. Gas peakers receive 75% to 85%, reflecting the temperature sensitivity and forced outage patterns that June 2025 exposed.

The gap between current accreditation and stress-tested performance isn't academic. S&P Global projects PJM capacity prices will remain in the range of $225 to $275 per MW-day through the end of the decade—pricing that bakes in scarcity conditions the current methodology failed to prevent.

Three Paths Forward

Reform proposals cluster around three approaches, each with distinct winners and losers.

Stress-based accreditation would replace historical averages with emergency-condition performance. PJM's pending capacity market filing proposes ELCC adoption, which would shift billions in capacity revenues from underperforming thermal units to batteries and other resources with reliable stress-condition availability. Generators with strong summer performance stand to gain 15% to 25% in capacity value; those with weak performance face corresponding losses.

The Accreditation Gap: Paper vs. Reality

The Accreditation Gap: Paper vs. Reality. Source: CAISO ELCC methodology. Gas peakers face 20-25% capacity credit reductions under stress-tested accreditation, while batteries maintain 85-95% credit.

Seasonal capacity requirements would disaggregate annual constructs into summer and winter obligations. MISO's seasonal proposal recognizes that reliability challenges differ by season—a gas plant with fuel supply constraints may fail winter peaks while providing summer value; wind resources with strong winter output but weak summer contribution could clear winter auctions while failing summer requirements. ISO-NE implemented a seasonal construct in 2025, with early results showing improved matching between capacity supply and reliability needs.

Enhanced performance penalties would increase the stakes for emergency underperformance. PJM's current regime reaches $3,500 per MWh for underperformance during Performance Assessment Hours—a penalty structure implemented after the 2014 Polar Vortex. Critics argue the rates haven't driven sufficient reliability improvement, as Summer 2025 demonstrated. But raising penalties risks driving marginal units into retirement rather than reliability investment, potentially worsening adequacy.

The International Evidence

Other markets have grappled with these tradeoffs. The United Kingdom's capacity market applies de-rating factors based on historical stress-condition performance, reducing capacity credit for resources with poor emergency availability below what nameplate suggests. Australia's National Electricity Market uses a Reliability and Emergency Reserve Trader mechanism to procure emergency reserves when market-based capacity appears insufficient—layering insurance procurement atop energy-only design.

Germany relies on strategic reserves held outside the market. Reserve plants cannot participate in energy markets during normal conditions but activate during emergencies, providing insurance without distorting normal price formation. The approaches differ, but share a common recognition: nameplate capacity and reliable availability aren't the same thing.

What's Retiring Through 2025

What's Retiring Through 2025. Source: FERC. Coal dominates near-term retirements, but gas losses include major assets like Utah's 1.8 GW Intermountain Power Project.

Who Wins, Who Loses

Capacity market reform redistributes billions annually. The political economy is predictable. Existing thermal generators resist reforms that reduce capacity revenues for their fleet. New entrants and clean energy advocates support reforms that reward performance over incumbency.

State interventions add complexity. The MOPR wars of 2018 to 2022—triggered by state-subsidized resources participating in capacity markets—remain incompletely resolved. Any reform must navigate the tension between RTO-level market design and state energy policy.

For load-serving entities, capacity procurement strategies require recalibration. Contracts structured under current accreditation rules may become misaligned with reformed market design. The prudent hedger is already modeling ELCC scenarios.

What Comes Next

Reform proceedings are active across RTOs. PJM, ISO-NE, MISO, and NYISO each have pending dockets examining accreditation methodology. FERC's technical conference on capacity accreditation will gather stakeholder input, with comments due in the first quarter of 2025.

The timeline matters because the underlying stress is intensifying. Federal regulators warned earlier this year that high temperatures and data center load growth will push 2025 summer consumption above the past four summers. Meanwhile, FERC data shows 60% of planned capacity retirements through 2025 are coal units, with natural gas comprising another 27%—including the 1.8 GW Intermountain Power Project in Utah.

The reform trajectory will determine whether capacity markets can adapt before the next stress event. Summer 2025 provided a warning. The question is whether regulators will act on it before the grid delivers a harder lesson.