A trading desk with no power stations and no customers can sell electricity in a day-ahead auction, buy the same quantity back the following morning in real time, and never deliver a megawatt. The position clears, settles financially, and disappears. In several liberalised markets this activity — virtual bidding — regularly accounts for a meaningful share of day-ahead supply and demand, and it influences the prices that generators, retailers and large consumers actually pay.
The practice matters most when markets are tight and the gap between day-ahead and real-time prices widens. That gap is where virtual traders earn their return, and also where physical participants absorb basis risk they may not be able to hedge. Knowing who is trading, what they are permitted to do, and how their activity interacts with the physical schedule is part of reading a market’s price signals correctly.
What a Virtual Bid Actually Does
A virtual bid is an offer or bid submitted into a day-ahead energy market by a participant that has no intention of physically producing or consuming the electricity. The auction algorithm treats it like any other order, and the resulting position is settled against the real-time price. The gross payoff is the spread between the two prices multiplied by the quantity cleared, before transaction fees, collateral costs and other settlement adjustments are applied. There is no delivery, no schedule and no plant behind the transaction.
The two directions are simple. A virtual supply position sells in the day-ahead market and buys the same quantity back in real time, earning a return when day-ahead clears above real-time. A virtual demand position does the reverse, buying day-ahead and selling real time, and earns a return when real-time is higher. Operators label the two directions differently, so the tariff definitions matter more than a summary of them. Neither position carries a physical obligation.
Virtual transactions are submitted at pricing nodes in nodal markets, not at a single system price. That means the position carries locational exposure as well as a time spread — a virtual trader at a constrained node is implicitly taking a view on congestion between the day-ahead and real-time network models, not only on the system-wide balance of supply and demand.
Why Market Operators Permit It
The core argument is convergence. Day-ahead and real-time markets clear the same physical system at different times and with different information. If day-ahead prices sit persistently above or below real-time outcomes, forward prices stop being a usable hedge and suppliers carry basis risk they cannot manage. A participant with no physical position can arbitrage that gap, and the act of arbitraging it pushes the two prices toward each other.
The second argument concerns conduct. Physical generators can, in principle, bid in a way that lifts the day-ahead price, since they own the capacity the auction needs. A virtual supplier has no plant to protect and will sell whenever day-ahead looks expensive relative to its expectation of real-time. That adds supply to the day-ahead auction independent of who owns the generating fleet, which disciplines unilateral behaviour — though it does nothing about structural scarcity or coordinated problems.
Adoption is uneven. Most large US regional transmission organisations, along with CAISO, run some form of day-ahead virtual bidding, with PJM introducing the mechanism in the early 2000s, MISO following later in the decade and CAISO adding its version, convergence bidding, in the early 2010s. European day-ahead markets under the single price-coupling algorithm generally do not offer an equivalent product, with financial participants in most of those markets taking positions through derivatives exchanges rather than through the physical auction. The availability and design of such products varies by jurisdiction, so the comparison is with the prevailing arrangement rather than a uniform rule.
Why Convergence Is Never Complete
The two markets do not solve the same problem. A day-ahead auction is typically a security-constrained unit commitment run on hourly granularity, using a forecast of demand and a network model built from known outages. Real-time dispatch resolves the same system in short intervals against actual conditions, actual outages and forecast error. Some of the resulting price difference reflects information that arrived after the day-ahead gate closed, and no financial position can trade that ex ante.
Price caps and floors also limit what convergence can do. Where a market administers an offer cap or a floor, a virtual position cannot push prices past it, and a spread that is being held open by the cap stays open. The same applies to non-convexities and mitigation rules that keep day-ahead prices below true marginal scarcity. Convergence works within the administratively defined price range, and the boundary of that range is where virtual activity stops being useful.
The most important constraint is procedural. Virtual positions clear inside the auction that also commits physical resources, so a large virtual supply bid can displace generation from the day-ahead schedule. Some designs then run a separate reliability commitment that ignores virtual positions and re-commits against physical expectations. The gap between a schedule influenced by financial participants and a dispatch driven by physics is covered in real time, and the cost of covering it lands on imbalance prices.
Where the Design Has Been Contested
Two strands of criticism recur. The first concerns persistent one-directional flow. When a structural spread keeps virtual supply profitable for months, volumes accumulate on one side and day-ahead prices are pressed toward the real-time level. One reading is that the day-ahead market had been clearing too high and the arbitrage is doing its job. The other is that the resulting day-ahead price understates the value of firm capacity and weakens the revenue signal that supports investment.
The second strand concerns the more complex products built alongside simple virtual bids. Up-to-congestion transactions let a participant take a position between two nodes, so the payoff depends on the congestion component as well as the energy spread, and a portfolio of such positions can generate revenue that is not simply the difference between two prices. Operators have revised these products over time as they learned how portfolios behaved, and enforcement attention has followed the same trail.
There is also a resource-adequacy dimension that is easy to miss. In markets where capacity resources carry must-offer obligations into the day-ahead auction, a virtual supplier competes directly with generation that is contractually required to be there. The virtual bid lowers the day-ahead price the obligated unit receives, while the obligation remains. Whether that is a healthy discipline on capacity payments or an unintended transfer away from the resources the market is paying to keep available is a live question in several jurisdictions.
What to Watch
Storage changes the picture. A battery arbitraging day-ahead against real-time is, economically, a virtual bidder that also has a state of charge, a round-trip efficiency loss and a duration limit. As storage capacity grows in a market, it competes for the same spreads that virtual traders harvest, and it can do something they cannot: hold the position physically when the price moves against it. The spread available to purely financial participants narrows as physical flexibility expands.
Most operators publish virtual transaction volumes and settlement outcomes by node. Those datasets are one of the few public windows into where the market disagrees with its own forward prices, and they reward watching: a node where virtual volumes are persistently one-sided is a node where the day-ahead model and the real-time system are telling different stories. Reading that signal is more useful than treating virtual bidding either as an unqualified efficiency gain or as a speculative add-on to the physical market.
References
- PJM — Manual 11: Energy & Ancillary Services Market Operations, virtual transaction and up-to-congestion provisions
- California ISO — Convergence Bidding Business Practice Manual
- ISO New England — Market Operations Manual, virtual bidding provisions
- MISO — Business Practice Manual for Energy Markets, virtual trading rules
- Regulation (EU) 2015/1222 establishing a guideline on capacity allocation and congestion management (CACM), on single day-ahead and intraday coupling
