
The Oil Chart Is Up, Your Fund Is Down—and It Never Owned a Barrel
A crude oil fund holds no barrel of West Texas Intermediate. It holds futures contracts stamped with expiry dates, and it has to replace them.
On a sunny spring weekend, the wholesale price of a megawatt-hour in California can slide past zero and keep falling. The California ISO has identified roughly 15,000 MW of system net load as a threshold where negative prices become likely, and when scheduled generation vaults past that floor the market clearing price turns into a payment that a generator must make to a counterparty: take this power, and we will pay you.

That is not a software glitch. It is the same auction logic that sets a positive price when supply is scarce, applied in reverse when the system is threatening to overfeed.
The mechanism is direct. In the hour-ahead market, if scheduled generation exceeds scheduled demand, the price of energy drops below zero to balance the books. At that point the direction of money reverses: the generator sending out megawatts becomes the one writing the cheque to whoever absorbs them. A 2014 analysis of CAISO’s five-minute real-time dispatch intervals found that on a specific day in 2014, energy prices were negative during a significant portion of the intervals. That day, the midday solar belly was deep enough that the market spent nearly half the afternoon telling producers to stop feeding the grid—by making continued delivery an out-of-pocket cost.
CAISO’s 2015 briefing on the duck curve flagged the same pattern. Negative energy prices indicating over-generation risk start to appear in the middle of the day, exactly when the net-load trough drops below the system’s downward-ramping capability. The price is not a valuation of the electron itself; it is a congestion signal denominated in dollars per megawatt-hour.
Why would a generator bid a negative price instead of going idle? The answer lives in the machinery.
A unit cannot be cycled below a certain minimum load without shutting off completely, and that minimum creates an operationally inflexible block of output. For a thermal plant, restarting after a shutdown is expensive enough—labour, fuel, wear and the hours lost while the boiler comes back to temperature—that a handful of negative-price hours looks cheap by comparison. The PyPSA modelling documentation puts it plainly: generators with start-up costs and minimum generation limits may find it more economical to offer electricity at a negative price than to shut down and restart later.
It gets worse for fossil-fuelled units. A paper on negative electricity prices and the Production Tax Credit, published by a research group, notes that operating fossil generation at minimum levels during negative-price periods becomes extremely expensive because the operator must pay fuel costs on top of literally paying to generate. The unit is burning money in two directions at once. Yet shutting it down would cost more still, so it stays lit. A wind or solar asset collecting a production tax credit may face an easier calculus: even after paying to deliver, the tax credit tied to each megawatt-hour can keep the spreadsheet in the black.
The timing is no accident. CAISO’s net-load shape—the duck curve—shows a deepening midday trough as solar resources erode daytime net electricity demand. The system can encounter minimum-generation thresholds that lead to oversupply, and when it does, the price heads south. Weekend spring conditions are especially prone: plentiful solar, wind and hydro press into a demand profile that is already moderate, closing in on the floor.
In 2019, a CAISO stakeholder comment observed that as renewable generation becomes the marginal resource more frequently, market prices will be zero or negative more often. That forecast is now visible in the settlement files. The belly of the duck, from late morning to mid-afternoon, is the window where over-generation risk concentrates. It is not capacity alone that triggers the negative price; it is capacity meeting a grid that has exhausted its ability to ramp down.
The evidence is public. ERCOT’s Real-Time Settlement Point Prices display publishes settlement-point values that include the Real-Time Reliability Deployment Price Adders, and those numbers can go negative. CAISO’s five-minute market data is also available; the 43-percent-negative-interval day came straight from an analysis of those feeds. Any trader, analyst or journalist can pull the settlement files and watch the price line dip below the horizontal axis. These are not simulations. They are audited settlement amounts that determine who pays whom the following week.
It is tempting to describe the root cause as “too much solar.” That misses the operational boundary. The grid does not fail at 15,000 MW of net load; it hits a point where all the dispatchable units that can turn down have already done so, and some are already running at their minimum stable levels. A unit that is cycled down to the edge of shutdown cannot go lower without tripping offline, so that block of generation must stay connected. If renewables are still pushing surplus onto the system, the ISO has two choices: order some generators to curtail, or let the price fall until it forces the least-flexible participants to pay their way out. Often both happen in sequence.
California’s weekend spring conditions, cited by CAISO, encapsulate the squeeze: abundant solar, wind and hydro output, mild demand, and a fleet of gas-fired plants that cannot be turned off fast enough because the evening ramp is only a few hours away. That arrival of inflexible minimum generation against a low net-load floor is what makes a negative price the economically rational outcome.
A negative settlement price on the wholesale market does not appear as a credit on a residential bill. Retail rates are set months ahead through regulated tariffs that blend every hour—negative and positive—into a long-term average. A single five-minute interval of minus ten dollars per megawatt-hour is a blip absorbed into a portfolio that also contains high-priced scarcity events. No utility sends the customer a line item for a sunny Tuesday afternoon.
The negative number is also not a verdict that electricity has become worthless. It is an instruction to certain generators: stay online, pay the toll, and wait for the evening ramp. As long as the duck curve remains unsmoothed, negative prices will recur at noon, and the cold-start arithmetic of thermal plants will keep some of them paying their way through the belly. For everyone else, the negative price on the screen is a reminder that the afternoon sun is now the cheapest hour on the grid—and occasionally the one that costs someone to deliver.
New material is signed by the newsroom, not by a personal byline: a name would have to come from somewhere, and there is no source for one. Corrections with a source are welcome at [email protected].
Updated

A crude oil fund holds no barrel of West Texas Intermediate. It holds futures contracts stamped with expiry dates, and it has to replace them.

The U.S. Energy Information Administration slices the natural-gas year into two neat halves: injection season runs from April 1 through October 31
Government As Tough Love An Interview with Robert Shiller Dec 16, 2025 Brandon Adams When we observe people fall from grace, we tend to get a dopamine jolt that can