Few commodities swing like electricity. Day-ahead prices can sit placidly at $30–40/MWh for weeks, then print hundreds or even thousands of dollars per MWh during a severe heat wave — for illustration, real-time prices in some markets have exceeded $9,000/MWh during extreme scarcity events. This is not a market malfunction. It is the market doing exactly what physics demands of it.
The physics of volatility
Three physical facts combine to make electricity prices uniquely jumpy:
It cannot be stored economically at grid scale. (Batteries are changing this at the margin, but the stock of stored energy remains small relative to total demand.) Every megawatt consumed must be generated at essentially the same instant. There is no inventory to draw down when supply runs short — the buffer other commodity markets rely on simply doesn’t exist.
Demand is extremely inelastic in the short run. When temperatures hit triple digits, air conditioners run regardless of the wholesale price — almost no end user sees that price in real time, and even those who do can rarely curtail instantly. Demand doesn’t politely decline as prices rise; it barely moves at all.
Supply is lumpy and steeply tiered. The last increments of available capacity — aging peakers, emergency demand response, imports over constrained ties — cost far more than baseload generation and exist only for a few dozen hours a year. The supply curve goes nearly vertical at the top.
When demand pushes against available supply, there is no warehouse to raid and no demand response to soften the blow. Price is the only rationing mechanism left, and it rations by rising until enough high-cost supply appears or enough voluntary curtailment materializes. The spike is the system balancing itself the only way it can.
The value of lost load: what scarcity should cost
Economists frame the efficient scarcity price around the value of lost load (VOLL) — the economic value consumers place on avoiding a blackout. Estimates typically run into the thousands or tens of thousands of dollars per MWh: an hour of lost power costs a semiconductor fab, a hospital, or a household far more than the few dollars of electricity it would have consumed.
The logic: if the market price during scarcity reached VOLL, two good things would happen. Consumers with the ability to curtail would face the true cost of their consumption and cut back voluntarily. And generators would earn, in those scarce hours, the revenue needed to justify building peaking capacity that runs only rarely. In the textbook, scarcity pricing simultaneously rations demand and finances supply. Reality, as always, is messier.
Scarcity pricing by design: the ORDC
Market operators formalize scarcity economics through the operating reserve demand curve (ORDC). The mechanism: as operating reserves shrink toward minimum requirements, the operator adds an administratively determined price adder to energy (and reserve) prices. The adder is small when reserves are comfortable and rises steeply as they thin — for illustration, from near zero at healthy reserve margins to thousands of dollars per MWh as reserves approach emergency levels.
The ORDC does two jobs. Operationally, it ensures prices reflect tightening conditions before an actual shortfall, giving demand response and flexible resources the signal to act early. Economically, it manufactures the scarcity revenue that peaking resources need — the “missing money” discussed in our capacity markets piece — within the energy market itself. Every U.S. organized market now uses some form of ORDC or scarcity adder, though the calibration differs enormously: ERCOT’s is the steepest, reflecting its energy-only design; eastern markets pair milder scarcity pricing with capacity markets.
Price caps: the political economy of spikes
If VOLL-based pricing is efficient, why cap prices at all? Because electricity is an essential service, and prices in the thousands per MWh sustained over days create genuine hardship and political crises. Most eastern markets cap energy offers around $1,000–2,000/MWh; ERCOT’s cap has historically been far higher at $9,000/MWh.
Caps are a tradeoff made visible. A low cap protects consumers during events but deepens the missing-money problem, requiring capacity markets or other out-of-market payments to sustain investment. A high cap sharpens investment incentives but concentrates financial risk — retailers and utilities caught short during a multi-day event can face existential losses, and the political backlash can itself threaten the market design. There is no technocratic optimum here; every cap level is a societal choice about who bears the risk of extreme weather.
Anatomy of a spike: how events unfold
Price spikes follow a recognizable anatomy. Stage one: tightening fundamentals. A heat wave or cold snap drives demand toward seasonal peaks while planned maintenance or forced outages reduce available supply. Day-ahead prices rise as the market anticipates stress — this is the orderly phase, where forward prices do their job of signaling.
Stage two: real-time stress. Actual conditions diverge from forecasts — demand overshoots, a large plant trips, wind output drops unexpectedly. Five-minute real-time prices spike as the operator dispatches ever-more-expensive resources and the ORDC adder kicks in. These spikes are often brief and localized: a single constrained zone, a handful of intervals.
Stage three (rare): sustained scarcity. Forced outages compound — plants that were supposed to return don’t, fuel supply tightens (gas pipelines constrain during extreme cold), and reserves erode over days rather than hours. Prices stay elevated across many intervals, the ORDC adder persists, and emergency procedures begin: public conservation appeals, then, in the extreme, controlled outages. This is the phase that tests market designs and balance sheets alike.
Understanding which stage you’re in is the key analytical skill. A stage-two spike is routine — the system working as designed. A stage-three event signals genuine systemic stress and is where the real economic and human costs accumulate.
The mirror image: negative prices
Volatility cuts both ways. When inflexible supply — nuclear plants that can’t easily cycle, renewables earning per-MWh production credits — exceeds demand in an export-constrained area, prices fall below zero. Generators pay to keep running rather than incur shutdown and startup costs.
Negative prices are the accurate signal of oversupply, just as spikes signal shortage. They incentivize the structural responses: storage to absorb surplus, transmission to export it, flexible demand (hydrogen electrolyzers, managed EV charging, data centers with flexible load) to consume it. A market that regularly prints negative midday prices and high evening prices is practically begging for batteries — and developers are listening.
Who wins, who loses, who hedges
Spikes redistribute wealth violently and quickly. Winners: generators that are running during the spike, especially low-cost baseload units earning enormous inframarginal rents; owners of demand response and storage that can sell into the peak. Losers: load-serving entities that are short — that bought less forward cover than their actual load — and must buy the difference at spike prices; in the worst cases, thinly capitalized retailers fail outright, and their customers are dumped onto providers of last resort at the worst possible moment.
This is why hedging discipline is existential in electricity retailing. The standard practice — hedging the large majority of expected load forward and keeping only a small residual exposed to spot — exists because of exactly this risk. Every retail electricity bankruptcy in market history is, at bottom, a story about being short during a spike.
Volatility and the energy transition
Two structural trends are reshaping volatility. Renewables suppress average prices (zero marginal cost shifts the supply stack down) but increase volatility around the average: steeper ramps, sharper evening peaks as solar fades, and weather-driven supply swings. The result is a market with lower middles and more extreme tails — challenging for traditional generators, attractive for flexibility providers.
Storage and flexible demand dampen volatility from the other side. Batteries buy low and sell high by construction, compressing spreads as they scale. Flexible large loads — data centers that can shift compute, industrial processes that can pause — act as shock absorbers. The long-run equilibrium is a market where volatility still exists (physics demands it) but where a growing share of the spike revenue accrues to the flexible resources that tame it.
Demand response: the missing elastic
The volatility story has a dog that doesn’t bark: demand response. In most markets, end users barely participate in real-time balancing — they don’t see the price, and their equipment can’t respond automatically. This absence is what makes demand so inelastic and volatility so extreme. Every increment of demand flexibility added to the system disproportionately dampens spikes, because it shaves the vertical top of the supply curve where prices explode.
The technology for this is arriving: smart thermostats that pre-cool homes before peak hours, water heaters and EV chargers that soak up midday solar, industrial processes paid to pause. RTOs have built the market mechanisms — emergency demand response programs, price-responsive load participation in energy markets — but enrollment remains a fraction of potential. The gap between technical potential and actual participation is one of the largest untapped efficiency gains in the power sector. A grid where 10–15% of peak load could flex within minutes would have structurally lower and less frequent price spikes — and the markets to pay for that flexibility already exist.
Hedging instruments: how professionals survive volatility
Sophisticated participants don’t just endure volatility — they trade it. Beyond standard forward contracts, the toolkit includes heat-rate-linked deals (tying power prices to gas prices), weather derivatives (payouts triggered by extreme temperature indices), and options on energy (caps and floors that limit exposure while preserving upside). Retailers layer these into structured hedges: for illustration, a retailer might hedge 90% of expected load with forwards, buy call options covering extreme-weather scenarios for the next 5%, and leave a small residual exposed to spot.
The cost of this insurance is embedded in retail margins — it’s one reason fixed-rate plans carry a premium over expected wholesale costs. And the discipline of hedging is what separates survivors from casualties: every major retail failure in market history traces back to inadequate hedging against exactly the volatility this piece describes. Volatility isn’t just a market phenomenon; it’s the central risk around which the entire retail electricity business is organized.
Reading a spike like an analyst
When prices spike, ask four questions. What was the reserve margin? Tight reserves plus high prices is the market working; comfortable reserves plus high prices suggests market power or a design flaw. What tripped? Forced outages are the usual trigger — the identity and size of the outage explain the magnitude. What was the marginal unit? The fuel and location of the price-setting resource tells you whether this was a local congestion event or system-wide scarcity. How long did it last? Duration distinguishes routine balancing from structural stress.
Volatility is not noise in electricity markets — it is the signal. The spikes contain the information about where the system is fragile, which investments pay, and whether the market design is doing its job. Learn to read them, and you can read the grid.