The Promise and Perils of Exclusion in Electricity Governance
Why the rules that make the grid governable are also the rules that can slow it down
Photo by Adrià Cerezo Bertran on Unsplash
RTOs are governance institutions that transform a reliability commons into a rule-defined club. That single claim explains both their success over the past twenty-five years and the intense strain they are under today.
My new article in the Journal of Institutional Economics asks what sounds like a narrow, technical question: what kind of institution is a Regional Transmission Organization? But the answer turns out to matter for some of the biggest questions in energy policy right now — why the interconnection queue has ballooned to over 2,600 gigawatts of proposed projects, why capacity market prices in PJM have hit historic highs, and why integrating data centers, batteries, and distributed solar into the grid is so much harder than it should be.
The short answer is that RTOs are clubs — but clubs of a very unusual kind. And understanding why changes how you think about what’s going wrong and what reform would actually require.
Reliability Is a Common-Pool Resource, Not a Public Good
Electricity reliability is routinely described as a public good, both non-excludable and non-rival. That framing understates the governance challenge.
Under normal conditions, reliability does look non-rival: when the grid is stable, one user’s benefit doesn’t diminish another’s. But this is misleading, because it only holds until the system comes under stress. Under congestion, reliability becomes rivalrous; one participant’s actions constrain what others can do. That combination of conditional non-excludability and latent rivalry and congestibility is the defining characteristic of a common-pool resource (CPR), in Elinor Ostrom’s framework.
Physics make this reality inescapable. In an AC grid, Kirchhoff’s Laws mean that electricity flows along all available paths — electrons don’t obey contracts. Every injection and withdrawal affects system flows throughout the network. This physics gives reliability its CPR character: it’s not that we lack the legal tools to exclude people from the network administratively, it’s that the physical consequences of participation can’t be isolated to the participants who caused them.
Ostrom’s great contribution was to show that CPRs don’t have to be privatized or placed under centralized state control in order to be governed sustainably. Through empirical work across fisheries, irrigation systems, forests, and other settings, she found that communities can develop durable, self-organized governance arrangements, if those arrangements satisfy certain institutional design principles. The core idea is that effective CPR governance requires clearly defined boundaries, rules that align benefits and costs, collective choice arrangements that give affected parties a voice, monitoring, graduated sanctions, and nesting across governance scales.
When regulatory restructuring separated generation from transmission and opened wholesale markets to broader participation in the 1990s, reliability could no longer be governed through internal hierarchy within vertically integrated utilities. The number and heterogeneity of actors using the network expanded enormously. A new governance architecture was needed, one that could specify participation conditions, performance obligations, and cost responsibility across a much larger and more diverse set of participants. RTOs are that architecture.
Bringing Buchanan into the Picture
Ostrom explains how common-pool resources can be governed. James Buchanan’s theory of clubs helps explain the specific logic RTOs use to do it.
A club, in Buchanan’s framework, provides partially excludable, congestible goods. More members spread fixed costs — a benefit — but also create congestion — a cost. Optimal club size balances those two margins. What makes clubs distinctive is that they govern access through membership rules rather than market prices alone. You have to qualify and pay dues, not just show up.
RTOs look remarkably like clubs. They spread the fixed costs of dispatch, settlement, and market administration across many participants. They use locational marginal pricing — essentially a per-use fee that reflects marginal congestion costs — to allocate scarce transmission capacity. And they condition participation on meeting technical and financial standards: creditworthiness, telemetry capabilities, compliance with dispatch instructions, adherence to performance obligations.
But here is the critical departure from Buchanan’s model, and the institutional pivot on which most of my argument turns.
In Buchanan’s framework, clubs are replicable. If a club’s governance becomes stagnant or exclusionary, dissatisfied members can exit and form another club. Competition among clubs disciplines governance. In theory, a population would naturally partition into clubs of roughly optimal size, and no club could sustain inefficient rules for long because the exit option is real.
Transmission systems don’t work that way. High fixed costs, network externalities, siting and permitting constraints, and the institutional realities of cost recovery make it effectively impossible to build competing regional grids. The transmission “club” cannot be replicated.
That non-replicability changes the institutional logic in a deep way. In a replicable club, competitive pressure from outside disciplines the club’s governance. In transmission, governance institutions must supply that discipline from within, which means that the design of those institutions is not merely a question of administrative preference; rather, it’s the primary determinant of whether the system remains both reliable and innovative over time.
Exclusion as Institutional Technology
The word I use to describe what RTOs actually do is exclusion: the bundle of rule-defined participation rights and obligations that governs access to the network and the market platform built on top of it.
RTOs can’t make electrons excludable; the physics won’t permit it. What they can do is formalize institutional exclusion: you may participate in centralized dispatch, settlement, and market coordination only if you meet certain criteria. Creditworthiness. Metering and telemetry capabilities. Compliance with dispatch instructions. Capacity accreditation standards that determine how much of your output counts toward peak reliability obligations. Performance penalties when you fail to deliver.
These rules are what make the reliability commons governable. They’re also what can make the club hard to enter.
This tradeoff is the tension at the center of the paper. The same rules that protect reliability can become barriers to innovation. And in a non-replicable club, where the exit option is constrained, there’s no competitive pressure to self-correct. Here, Albert Hirschman’s Exit, Voice, and Loyalty offers an important frame that I didn’t develop fully in the article: when exit is blocked, voice and institutional reform must substitute as disciplining mechanisms. That constraint puts an enormous amount of weight on the quality of the stakeholder processes and regulatory oversight through which rules are revised. When those processes are slow or captured by incumbents, governance stagnates even when the case for reform is widely understood.
The Pacing Problem in Practice
Most RTO participation rules were designed in the 1990s around large, centralized generators. The resource landscape today is completely different: variable renewables with near-zero marginal costs, battery storage with fast-ramping capabilities, distributed energy resources that aggregate thousands of small devices, flexible demand, and now enormous data-center loads that are reshaping regional supply-demand balances almost overnight.
When technology changes faster than institutional rules, a pacing problem emerges. Governance arrangements that once enabled competition can begin to impede it. And in a non-replicable network club, the main mechanism for resolving that mismatch — competitive entry by alternatives to the existing club — isn’t available. Adaptation has to occur through rule revision within the existing governance architecture, which is a slower and more contested process.
This shows up most visibly in interconnection queues. By 2024 the active queue across U.S. RTOs had grown to over 2,600 gigawatts of proposed generation and storage capacity, more than twice the existing current U.S. generation fleet. The overwhelming majority of new requests are for solar, wind, and storage. Many projects spend three to five years in study backlogs before receiving an interconnection agreement, and a large share ultimately withdraw. The queue is a governance bottleneck that determines which technologies can enter the reliability club and on what timeline.
FERC Order 2023’s move to cluster studies — evaluating groups of proposed projects together rather than sequentially — represents an attempt to redesign how exclusion is administered at the entry stage. Whether it works depends on implementation, and on whether the cost allocation rules that assign network upgrade costs across projects are revised in ways that don’t simply recreate the same backlog under new names.
Capacity markets face analogous stress. PJM’s forward capacity auctions, which procure commitments from generators to be available in future delivery years, produced historically high clearing prices in recent auctions, triggering political backlash and FERC-approved price collars. Those high prices are partly a function of slow entry: when the rules governing who qualifies to participate haven’t kept pace with the changing resource mix, supply lags, and scarcity pricing follows in a non-replicable market.
Data Centers as an Institutional Shock
The relatively sudden growth of large-load demand from data centers has amplified every one of these dynamics.
Data-center loads are large, lumpy, highly reliability-sensitive, and growing faster than the governance processes designed to integrate them. Some developers are exploring co-location with generation or novel interconnection structures that sit at the edge of existing rules. That raises fundamental questions about the boundary of the reliability club: Are these loads fully inside? Do they pay proportional cost shares? Do they impose reliability externalities on other participants, and if so, how are those externalities priced and managed?
These aren’t primarily engineering questions. They’re institutional questions about how boundary rules are drawn, who gets to participate in drawing them, and how quickly they can be revised when circumstances change.
Polycentricity: Enabler and Constraint
RTO governance is polycentric. Authority is distributed across multiple semi-autonomous decision-making centers that interact within a broader institutional framework. RTO stakeholder committees, independent boards, FERC oversight, NERC reliability standards, state public utility commissions, and adjacent RTOs with their own rules and seam-coordination challenges all share jurisdiction over different pieces of the governance architecture.
Polycentricity can enable experimentation and learning. Different RTOs can pilot participation models for storage, aggregated DERs, or large loads, and successful designs can diffuse through observation and emulation. CAISO’s DER Provider model and NYISO’s storage participation framework are examples of this kind of institutional experimentation at the regional level.
But polycentricity also creates veto points. Rule changes that would lower entry barriers for new technologies — revised accreditation methods, adjusted telemetry requirements, reformed cost-allocation frameworks — have to navigate multiple venues: stakeholder working groups, RTO boards, FERC filings, reliability standard reviews, and sometimes state commission proceedings as well. When costs and benefits are unevenly distributed across those venues, and when incumbents who benefit from existing rules have disproportionate influence over the process, reform can stall even when a strong case exists.
In a replicable club, that stalling would be disciplined by competitive exit. In transmission governance, it isn’t. The polycentric structure provides checks on unilateral decision-making, which is valuable, but it also raises the transaction costs of adaptation in ways that can amplify the pacing problem precisely when it’s most acute.
What Would Reform Actually Require?
The article’s practical implication is not to eliminate exclusion, which is unavoidable for reliability in a congestible network, but to design exclusion that is contingent, evidence-based, and systematically adaptable.
That means a few things concretely. Boundary rules should be treated as deliberately revisable rather than as fixed templates inherited from the restructuring era. Tiered participation structures, in which access rights and obligations scale with a resource’s system impact, with clear pathways toward fuller participation as compliance is demonstrated, can substitute for the missing competitive discipline that replication would otherwise provide. And RTOs should adopt explicit protocols for periodic rule revision so that the exclusion regime remains adaptive as technology changes, rather than leaving reform to episodic conflict and prolonged stakeholder combat.
This last point connects back to Hirschman. In a setting where exit is constrained, the quality of voice mechanisms becomes essential. If the collective-choice arrangements through which boundary rules are revised are captured by incumbents with strong interests in slow-walking change, the adaptive capacity of the institution degrades even when the case for reform is compelling. The governance of governance, how rules about rules get made, turns out to matter as much as the rules themselves.
The Larger Point
RTOs have delivered real benefits: improved dispatch efficiency, expanded trade, congestion pricing, and substantial decreases in wholesale energy prices across the regions they serve. The institutional architecture that produced those benefits is genuine and worth preserving.
But that same architecture is now under strain. Exclusion is both promise and peril: the mechanism that makes a reliability commons governable, and also the lever through which incumbency can harden and innovation can slow. Data-center growth, renewable integration, and digitalization are both engineering challenges and governance challenges.
Because the transmission network cannot be replicated, institutional design is the primary margin of adaptation. The debates over interconnection reform, cost allocation, capacity markets, and large-load participation are, at bottom, debates about how to recalibrate exclusion in a non-replicable network club.
If we get that recalibration right, RTO governance can remain both reliable and adaptive. If we don’t, we risk turning a once-innovative institutional architecture into a procedurally congested bottleneck — not just for wholesale market efficiency, but for the energy transition and for whether the grid can accommodate the digital economy that increasingly depends on it.
The full article, “The Promise and Perils of Exclusion: Using Institutional Design Principles and the Theory of Clubs to Analyse Regional Transmission Organization Governance,” is published open-access in the Journal of Institutional Economics.