PJM’s Capacity Market Is Signaling a Systemic Warning

Analysis by Arunika Chandra

Source: iStock

PJM’s most recent capacity auctions have been sending a signal that has gotten everyone worried. The capacity auction, also called Base Residual Auction (BRA), is held annually and locks-in capacity generators three years in advance. The 2025/26 BRA prices were $269.92/MW-day except for the BGE and Dominion zones which cleared at $466.35/MW-day and $444.26/MW-day respectively because of high data center demand. This was almost a 10-fold increase from the prices that cleared in the 2024/25 BRA at $28.92/MW-day. The 2026/27 BRA results that came out recently increased even further at $329.17/MW-day. This surge has gotten states worried as PJM expects electric bills to increase between 1.5-5% with higher risk of outages. The surge in capacity auction prices is a clear distress signal for the state of resource adequacy in the PJM grid.

But why is this the case? The answer lies in early retirements of thermal generators, increasing load centers, interconnection bottlenecks, as well as siting and permitting issues at state levels that will exasperate the supply-demand situation and can pose reliability concerns around the 2027-2028 timeline and onwards. PJM has already cleared 60% of the interconnection queue backlog, 90% of which is said to be Solar and Storage Interconnection Requests. Spring 2026 onwards, the interconnection queue timeline for new requests is now expected to be 18 months after the RTO moved from a first-come, first-serve serial study process to a first-ready, first-serve cluster study process.

A lot has been written already about the concerning state of the grid because of these challenges. But what if we view these auctions not just through the lens of a technical failure but also through the lens of a systemic one that encompasses the very electric market within which the grid inhabits? There is emerging interest in viewing ‘markets’ as a place that house emotions of all market participants and is rarely an autonomous and detached structure. And emotions at an institutional level form systems and systemic rules within which individual participants perform and cause varied outcomes. This school of thought is not unprecedented though. Oil Markets, the 2008 Financial Crisis, and even Healthcare Markets are all examples. In all these cases, resilience has been constantly undervalued.

While electric markets factor in continuous and reliable power supply as a cornerstone for resiliency, they are unable to adapt quickly enough to meet rapidly changing dynamics of energy demand thereby giving the illusion of sustained resiliency.  The grid too, therefore, is a social machine. And when it breaks, it breaks quietly through normalized emergency, procedural inertia, or institutional delay.

There are some PJM initiatives that help abetting this situation. FERC recently approved a stop-gap arrangement through PJM’s Reliability Resource Initiative (RRI) that prioritizes interconnection requests of 51 shovel-ready projects amounting to 11.8 GW. But it is anticipated that most of these projects will not be able to come online before 2029-30 timeframe at best. FERC also recently approved PJM’s Surplus Interconnection Service (SIS) proposal that allows generators to utilise unused portion of existing generating facility’s interconnection capacity and will be processed through a separate queue expected to take 270 days to process. This is a great opportunity for batteries to be co-located with existing solar projects. Still, the burden of processing these requests lie with PJM. With limited staff, one wonders if using the MISO-route where developers can outsource the study to third-party consultants and get results verified by the RTO could have been an efficient way to catalyse the SIS process.

Nevertheless, there is some promise in these initiatives. But it also begs the question – how did we get here? This takes us back to 2007 when FERC held a Technical Conference to examine why interconnection delays persisted across RTOs, despite earlier reforms under Order 2003 meant to open the grid to third-party generators. The goal was to understand why compliance hadn’t translated into timely processing. In March 2008, FERC put out an Order on the Technical Conference seeking reports on the status of each RTO and ISO’s efforts to improve the processing of their interconnection queues. As a result of these efforts, MISO was among the first few RTOs to be an early adopter of a first-ready, first-serve cluster study process in 2008. That paved the way for MISO to be the first RTO in 2017 to incorporate decision points and a three-phase study approach that helped keep interconnection timelines in some sort of a check. PJM filed its tariff revisions as well. They included cluster studying of queued projects among several other revisions. These changes were later accepted by FERC but formalisation and implementation of the cluster studies in a first-ready, first-serve basis within PJM didn’t happen until after the FERC Order 2023 in 2023, 20 years later. So why the lag?

A fault line: slow yet steady

Part of the answer may perhaps lie in PJM’s approach to change rooted in what is commonly known as Status Quo bias until losses from the Status Quo seem disproportionately larger than the gains. PJM’s interconnection process, though imperfect, represented a known baseline. Responses to FERC’s Orders were viewed primarily through the lens of compliance and not actual tangible change. This bias is not unique to PJM alone but has clearly impacted PJM in a compounding way. Transitioning to cluster studies introduced uncertainty and several questions with it – Would it create new disputes over cost allocation? Would developers tolerate less optionality? Would the system become more centralized and planning-heavy, contradicting PJM's market-based ethos? The deemed losses of change could have seemed to outweigh the gains fostering an increased risk-averse posture. But this very approach has put PJM on path for increased risks.

The remaining part of the answer might lie in its complex Governance process. PJM’s stakeholder process follows a multi-tiered voting structure: affiliate members vote in subcommittees, with decisions escalating to the MRC, MC, and finally the Board. States do not vote but participate through the Organization of PJM States (OPSI), which can attend meetings, file comments, and escalate issues to FERC. Final committee tallies report only percentages of Ayes and Nays, without disclosing detailed vote breakdowns. There is no right and wrong way of a Governance mechanism especially in a Quasi-Government entity like an RTO/ISO. And while these governance mechanisms have been useful in the past, stakeholder processes as conducted today can end up being resource-intensive and time consuming with upwards of 200-300 meetings per year and layers of task forces and user groups perhaps making the entire process more inefficient and overtly hierarchical. During times of low contention issues, these processes are highly effective but during an issue of high contention is when these same processes can prevent from stakeholder coordination and agreement when it is needed most. High-level stakeholder committees frequently play a role in blocking proposals from being approved. FERC already sought filings from RTOs and ISOs for enhancing responsiveness between board advisory committee and stakeholders through FERC Order 719. It is perhaps time for the commission to seek improvements in the same aspect again.

Quasi-government entities, as the name suggests, have often faced the dichotomy of displaying characteristics of both Government and Private Organizations which brings forth the question of the kind of accountability they are to follow as well as the extent to which they are to consider public interests. RTOs can be seen as this middle-party between FERC and stakeholders of which states are an integral part. While on one hand FERC lacks statutory authority to promote public welfare directly, states most certainly have that authority and purpose. It is this middle role that can sometime seem like a grey area. It is evident that RTOs and ISOs are to be deemed neutral bodies, yet they have huge power and impact on the everyday lives of millions of people. If stakeholder processes are the solution to effectively bring forth public welfare in the energy realm, the efficiency in doing so will lie in the way in which states are to be involved in the stakeholder processes and voting processes of an RTO/ISO.

Nine Governors have already sent a letter to PJM seeking the creation of a separate and formal group solely for the Governors of PJM states that will provide additional input to the Board. Further, they have sought to have the power to nominate candidates to fill two Board positions. It might be equally prudent, if not more, to seek an altogether comprehensive and bottom-up way to transform participation of all members and stakeholders within the PJM governance system.

States are responsible for the elimination or waning off coal subsidies, fostering implementation of Renewable Portfolio Standards (RPS), some renewable incentives, and retail-level rate making – all of which critically impact public welfare. Yet some might consider their ability to participate is somewhat limited. PJM is limited by multiple aspects starting with its governance system and its status quo bias. And FERC’s process for seeking reform is lengthy at best and overt at worst. This brings forth a key question – who is truly in-charge of what? And how do we get all three levels of entities to work together in the best way possible?

Backward-facing assumptions for forward-facing risks

We’re already feeling the impacts of narrowing reserves as a consequence. Take the most recent Heatwave for example. PJM met a 161,000-MW demand for power, the highest since 2011, and had to curtail exports to neighbouring regions. As we move toward an increasingly heating world, Heat waves could no longer be rare hypotheticals but instead real-time, possibly frequent, stress tests testing not just the endurance of people but of our infrastructure, including our grids. It is during times like this when systemic postures influence not just institutional behavior but also modelling inputs and assumptions. This can be understood through PJM’s summer outlook.

The contribution of the voluntary 30-minute supplemental reserve market, also called the Day-Ahead Scheduling Reserve (DASR) Market, is incorporated within the Summer 2025 Outlook. Under a conservative load forecast estimate of 166 GW (which was closer to the actual peak demand) a deficit of 1.5 GW in the DASR was already anticipated in the base scenario only to be exacerbated under low renewables and/or active gas contingency scenarios. Turn of events during the heatwave aligned with procedural code where expensive emergency resources and voluntary DASR participation particularly through Demand Response were relied upon to meet the DASR margin. It is evident that PJM already has the right kind of procedural checks and balances in place for such an event. But it is key to note that while the system did anticipate shortfall, it operated with minimal headroom, betting on voluntary behaviour, and accurate DR response — none of which are guaranteed under climate-accelerated stress scenarios. Studies already indicate that multiple and frequent requests for load adjustment can lead to Customer Fatigue. The issue of the extent of conservatism regarding reserve margins is always debatable but there is an economic rationale to be considered for some kind of conservatism as we see Forecasting uncertainties increasing over time due to climate volatility, higher load volatility due to data centers, and rising political cost of even rare blackouts, especially under high public energy anxiety.

If the forecasted shortfall is framed as a manageable gap with fallback options (like expensive emergency procurement or voluntary Demand Response), then not acting preemptively doesn’t feel like a loss, it feels like “avoiding unnecessary cost”. This is called Loss Aversion – a cognitive bias that influences people or systems to find the psychological pain of losing is twice as powerful as the pleasure of gain resulting in people trying to avoid losses much more than trying to avoid equivalent gains. The cost of deploying higher margins and not purely relying on voluntary behavior and emergency response systems seem higher than they might be especially when compared to the benefits. But if it were framed as a reliability risk that could trigger emergency pricing or public backlash, the system might act sooner.

It is also evident that there has been a gross error in accounting forecasted load demand. PJM had forecasted summer peak load to be 154 GW while the actual peak load on the 23^rd and 24^th of June 2025 was almost 162 GW and 162.4 GW respectively. The 2025 Long-Term Load Forecast highlighted – “No explicit assumption is being made about Climate Change, and thus the assumption is that future weather will resemble past weather”. This is even though in 2022, a technology company for energy and water resource management called Itron that was engaged by PJM recommended that forward-looking temperature trends through climate models will need to be incorporated to improve accuracy of PJM’s long-term load forecasts. One wonders if this might be an issue of probability weighting - a bias that reflects how people often overweight small probabilities and underweight large ones. PJM deferred action perhaps because climate-driven extreme heat was construed to be a comparatively lower-probability tail event in the short term when in fact it is increasingly becoming a new central scenario. PJM might also be falling prey to historical normalcy bias resulting in a planning framework that is democratic in procedure but is still risk-averse to change and therefore, technically overconfident creating a false sense of adequacy. The question then becomes – can resource adequacy planning evolve fast enough when the data we use is backward-looking, and the risk is forward-scaling?

What now?

Considering the problem of today, utilities and Load Serving Entities (LSE) have the option to shift to a Fixed Resource Requirement (FRR) alternative. An FRR is an option through which utilities and LSEs can opt out of PJM’s capacity market, committing instead to self-supply via bilateral contracts or owned assets while being subject to penalties by PJM in the event of short-supplies. States with strong vertically integrated utility traditions (like Virginia and parts of the Midwest) tend to have higher FRR engagement. These decisions are often as much about legacy baseload plants of coal and hydro technology, long-term expensive fuel contracts, regulation, and politics as they are about economics. While some of these LSEs have come under fire for high capacity prices because of their expensive fuel contracts, PJM’s recent price trends in capacity auctions can help them justify their procurement philosophy.

This also seems to impact the predicament that deregulated LSEs and utilities are currently in. With PJM’s high BRA auction prices signalling a dire need for new generation to enter the market, even deregulated LSEs and utilities are in a dilemma wondering whether they should re-enter the generation owning business that they long left behind to favor deregulation. Even if these LSEs adopt a generation owning business model, PJM’s past actions of penalising LSEs with FRR units sets a dangerous precedent. For context, PJM shifted from using an average to a marginal method for valuing how much different resources contribute to reliability — meaning each new megawatt is now judged by its incremental value rather than class averages. While this approach reflects real-time system needs more accurately, it significantly lowered the credited value of renewables and, without a transition period, penalized utilities that had already planned their capacity using older assumptions. This goes against the very premise of the Order 2000 aimed for the creation of RTOs and ISOs that explicitly stated that “the existence of RTOs has not and will not… interfere with traditional state and local regulatory responsibilities.” Yet, lack of a transition period puts that very autonomy at risk. Moreover, if discriminatory transmission access is the concern, then the persistent delays in PJM’s interconnection queue have had far broader discriminatory effects impacting all market actors, not just FRR participants. The asymmetry here is not just technical, but jurisdictional.

High ELCC Storage technologies are perhaps the bridge between states RPS obligations and PJM’s reliability needs. But as mentioned earlier, interconnection is the immediate bottleneck. It is also important to note that states have jurisdiction over electric sales at a state level that FERC does not have limiting its jurisdiction in taking steps for any immediate resolution to bring more storage assets online beyond what is possible through the RRI. But storage can aid transmission planning as well which brings forth the question – Can FERC enforce a separate interconnection and storage related order that aims to accelerate interconnection of storage particularly? Would storage then be viewed as a generator or a reliability and transmission aiding tool?

The situation today is a consequence of systemic fault lines across risk preparedness, technical planning aspects, governance and jurisdiction, as well as market ethos. The Eastern interconnection, therefore, is in a reactionary state – not designed to align with forecasted potholes along the road. Until that changes, the system will be at a constant risk of being two steps behind the problem. And all these impacts will increasingly be felt most by vulnerable communities for whom an increase of even a few dollars can put them either in a state of debt or risks to health due to dramatic shifts in weather.