Somewhere in the control room of a gas-fired peaker plant that runs maybe four hundred hours a year, a finance director is staring at a spreadsheet that has nothing to do with fuel costs. It has everything to do with an auction she probably lost.
Capacity markets exist to solve a problem that pure energy markets cannot: they pay generators not just for the electrons they produce, but for the promise of being available when demand spikes. The logic is sound. A grid that only rewards output will, over time, starve itself of the reserve margin that keeps the lights on during a heat wave. But the structure of those auctions, the geometry of how bids clear and who gets paid, does something else entirely. It quietly sorts generators into survivors and the condemned.
The clearing price is a blunt instrument with sharp consequences
Most capacity markets, including PJM's Reliability Pricing Model covering much of the US mid-Atlantic and Midwest, run as descending-clock or sealed-bid auctions against a sloped demand curve. The grid operator estimates how much capacity it needs, sets a curve that pays more per megawatt as available supply shrinks toward that target, and generators submit offers. When the auction clears, every resource that clears receives the same single clearing price, regardless of what it bid.
That uniformity sounds fair. It isn't neutral.
Consider two generators: a 600-megawatt combined-cycle gas plant built fifteen years ago with its capital costs largely recovered, and a 200-megawatt wind farm that has just come online and is still servicing project-finance debt. The older plant's cost to stay available is low, so it can afford to bid near its floor and still profit handsomely if the price clears high. The wind farm needs a minimum revenue to cover its fixed obligations. If the clearing price lands between those two thresholds, the older plant collects a windfall and the wind farm collects the same number on paper but still may not pencil out across a multi-year investment horizon. Same price. Opposite fates.
Now flip the scenario toward retirement. A coal plant built in the 1970s faces mounting maintenance costs, environmental compliance capital expenditure, and a heat rate that looks embarrassing beside newer combined-cycle units. Its offer into the capacity auction has to reflect those real costs, which means it bids high. If it clears, it survives another commitment period, typically one to three years. If it doesn't clear, the revenue gap may be enough to tip the economics toward early retirement. The auction didn't explicitly choose to close that plant. The structure just made staying open uneconomic.
This is the mechanism. Not a policy decision, not a regulator's order. An arithmetic outcome.
The minimum offer price rule and the retirement it accelerates
The sharpest edge in modern capacity market design is the Minimum Offer Price Rule, or MOPR. Introduced to prevent heavily subsidized resources from suppressing the clearing price by bidding near zero, MOPR floors set a minimum at which certain generators, often those receiving state clean energy subsidies, can offer their capacity. The intent is to protect market integrity. The effect is considerably more tangled.
A nuclear plant receiving a state zero-emission credit, for instance, might carry an all-in cost of roughly $150 per megawatt-day to stay open. Without MOPR it could bid at $10, clear easily, and collect the market price on top of its subsidy. With MOPR forcing it to bid near its full cost, it competes on equal terms with unsubsidized gas and may not clear at all. If it doesn't clear, the subsidy alone might be insufficient to keep it operating. The very rule designed to prevent market distortion can accelerate the retirement of a low-carbon resource that the state's own policy was trying to preserve, which is roughly the regulatory equivalent of prescribing a cure that neutralises the original medicine. PJM's prolonged fights over exactly this tension, involving plants in Illinois and New Jersey, consumed years of Federal Energy Regulatory Commission proceedings for good reason.
The practical upshot for any generator is stark: capacity market rules are not background conditions. They are the terrain on which survival is decided. A plant that would be profitable under one set of auction parameters can be uneconomic under another, with nothing about the physical plant changing at all. That is not a market efficiently allocating resources. That is a rulebook reshaping an industry, and it deserves to be read as such.
Ask yourself whether the engineers who built a combined-cycle unit in the early 2000s imagined that its commercial life would one day hinge on a FERC order redefining offer floors. They did not. They imagined fuel costs, load growth, maintenance schedules. The auction architecture they now inhabit was not yet invented.
This means retirement decisions are partly jurisdictional artifacts, a point that receives far less attention than it warrants. Two identical gas turbines, one in a capacity market region and one in a regulated cost-of-service state, face entirely different existential pressures. The turbine in the regulated state gets a rate case: slow, negotiated, visible to regulators and intervening parties. The turbine in the market region gets an auction: fast, mathematical, and easy to misread as the neutral verdict of supply and demand when it is, in fact, the verdict of a particular set of design choices made in a particular decade by particular committees.
The generators that retire before their engineering lives are up are not usually broken. They are on the wrong side of a clearing price that nobody designed to be a demolition order, but that functions as one all the same. The distinction between an intended consequence and a structural one offers cold comfort to the finance director updating her decommissioning timeline.