How Dynamic Electricity Pricing Can Improve Market Efficiency

New research co-authored by Wharton’s Arthur van Benthem demonstrates how consumers could benefit from aligning electricity prices with the cost of producing and distributing that power.

Energy economics has long grappled with approaches to peak load pricing or time-of-use electricity pricing. Most people are used to paying a flat rate for every unit of the electricity they consume, barely aware that their distribution company procures that power at rates that vary by the time of the day and the source, be it fossil fuel or renewable energy. That disconnect has persisted without efficient and real-time electricity pricing because of politically unpalatable consumer resistance and technological obstacles.

A new research paper by experts at Wharton and elsewhere demonstrates the potential gains for consumers from aligning electricity prices with the cost of producing and distributing that power. The paper, titled “The Efficiency of Dynamic Electricity Prices,” used statistical analysis and a machine-learning exercise to model different electricity-pricing scenarios.

The study has three main findings: First, time-of-use rates and critical-peak pricing — the two commonly used time-varying rate plans — each correct about 10% of mispricing. Second, complex time-of-use rate structures based on historical prices often backfire. Third, real-time pricing with price ceilings can capture most potential efficiency gains, but that opportunity is likely some time away. When implemented in combination, TOU and CPP pricing policies deliver 17-20% of the efficiency gain that is possible with real-time pricing.

“Power producers currently benefit from extreme peak prices in the wholesale market. Real-time pricing on the end customer side would soften those peak prices, as demand shifts out of peak hours,” said Wharton professor of business economics and public policy Arthur van Benthem, who co-authored the paper with Andrew J. Hinchberger, a graduate student at Northwestern University, Mark R. Jacobsen, economics professor at the University of California, San Diego, Christopher R. Knittel, professor of applied economics at MIT’s Sloan School of Management, and James M. Sallee, professor of economic analysis and policy at Haas School of Business, UC Berkeley.

The authors modeled their time-of-use pricing scenarios based on historical data on hourly wholesale prices from 2000 to 2020 across major electricity markets that covered two-thirds of the U.S. population. The dataset covered seven independent U.S. system operators: PJM (East Coast), ISO-NE (New England), NYISO (New York), ERCOT (Texas), MISO (Midwest), SPP (South Central), and CAISO (California).

The paper highlighted the contrast between the extreme volatility in the wholesale electricity markets and the prices retail consumers pay. Mean electricity prices are usually in the $20-40 per megawatt hour range, but extreme pricing occurs either negative or in the thousands of dollars per megawatt hour. In one instance involving Ercot, the wholesale auction price soared to more than $9,000 per megawatt-hour. The cost of electricity generation varies substantially across time, as the paper explained. Low levels of demand are typically met by solar, wind, and inexpensive baseload power plants, while higher levels of demand are served by costlier “peaker” plants.

In contrast to the volatility in the wholesale electricity markets, retail consumers have price certainty: nationally, 90-95% of end-users face a flat price schedule, the paper noted, citing data from the U.S. Energy Information Administration.

The paper also explained the pricing differences between wholesale and retail electricity markets: A wave of deregulation created wholesale electricity markets that yield prices which achieve a real-time balance between supply and demand. But those prices “are rarely passed through to final consumers,” the paper noted.

Instead, end-users face prices that represent costs averaged over an extended time period, often a year or more, it added. “This creates an inefficiency because end users purchase power at prices that are well below the cost of provision in some hours and well above that cost in others.” The paper estimated the inefficiencies from mispricing working out to $2 billion annually.

The process of determining retail electricity rates allows utilities to pass much of their costs to consumers. “Utilities agree on flat or time-of-use rates during so-called rate cases with state public utilities commissions,” van Benthem said. “These flat prices account for high peak-hour pricing in the wholesale market. So indirectly, peak pricing is reflected in the flat rates that customers pay. This does not insulate utilities from all cost fluctuations, but they have other tools (e.g., long-term contracts) to fill in that role.”

According to the paper, time-of-use (TOU) rates and critical-peak pricing (CPP), or dynamic pricing, is “relatively simple and feasible.” However, more complicated time-of-use pricing schemes often backfire. Real-time pricing complicates consumer decision-making because it is too complex, confusing, or risky to impose on residential customers, it noted. Instead, utilities have been moving in the direction of time-varying rates with more and more complex time-of-use prices, it added. The study found that this does not deliver any benefits: “The vast majority of the efficiency gains come from a time-of-use plan having just two rates,” the paper stated.

Simplicity, Predictability and Efficiency

The study compared the pros and cons of pricing schemes along three dimensions, as set out in an earlier version published by the Kleinman Center for Energy Policy at Penn.

Simplicity: This refers to the ease with which retail consumers can understand their electric bill. Instead of tracking different prices based on their electricity usage over time, the simplest is a standard, single-price scheme. But simpler tariffs are less able to capture variation in the wholesale market, causing pricing inefficiency.

Predictability: This is about the ability of retail customers to know the price of electricity in a given hour sufficiently in advance of their decision-making. Fixed prices shield consumers from unpleasant “surprise bills.” But wholesale electricity prices are difficult to predict far ahead of time. Consequently, predictable pricing will be hard to align with market conditions.

Efficiency: This visualizes settings where customers pay the true variable cost of power, with the right incentives to conserve electricity when it is expensive. A fixed, single retail price causes over-use of electricity in peak periods and under-use in off-peak periods. Utilities and their customers could be made better off by shifting some consumption away from peak hours. This could have favorable environmental consequences, as it causes drastic shifts in load throughout the day. Electricity generation using peak-load generation technologies have higher variable costs and also often emit more pollution than base-load generators.

Why Is More Efficient Pricing Workable Now?

According to the authors, the present time is opportune to consider ways to better align power costs with retail prices for two reasons. First, more complex prices are now feasible with the large-scale adoption of smart meters which enable utilities to receive real-time information on consumer demand and to send real-time price signals to consumers. Computerized “smart” electricity meters that allow high-frequency measurement at the customer level are already a reality in most parts of the U.S., the paper noted.

Second, changing market conditions make it possible to achieve gains from better pricing. They include efforts to transition to clean energy and expand electrification of end uses. Power loads have surged with the electrification of transportation and buildings, as well as explosive growth in data centers, but those are also the sectors that are likely to respond to time-varying prices. “Failing to set efficient prices for these growing sources of demand would be a critical missed opportunity,” the paper stated.

“The power sector is suddenly faced with enormous increases in demand from vehicle electrification and data centers,” van Benthem noted. “At the same time, there’s more and more renewable energy that enters the grid, but that is difficult to store. Shifting electricity demand between different hours of the day is crucial to keep the grid stable and clean, and providing efficient price signals will give the right incentives for this to happen.”

Even as those gains beckon, utilities and regulators have resisted real-time electricity pricing because they fear that consumers will complain about price surges and unpredictable bills, the paper stated. It cited the case of skyrocketing wholesale-market prices in Texas in the winter of 2021, when customers who had opted for real-time pricing faced huge electricity bills.

Baby Steps to Real-time Electricity Pricing

The time-of-use and critical-peak pricing plans that the paper suggests capture only a modest fraction of the mispricing, each at 10% of the efficiency gains that real-time pricing could achieve. There is good reason for that conservative approach, at least for now.

In the case of CPP plans, it is best if the utility can set CPP rates close to events where usage peaks. In practice, CPP rates are typically set in advance during rate hearings and are often set for multiple years; thus, they are not tailored to specific peak events, and they fail to meaningfully improve efficiency, the paper noted. “Fully realizing the gains from real-time pricing may only be possible in the future,” it concluded.

Utilities could also consider real-time pricing paired with strict price caps. “This type of pricing can recover the vast majority of the current efficiency gap while still protecting customers from extreme price fluctuations,” the paper stated. But that path has hurdles: While the estimates of efficiency gains from real-time pricing assume demand sensitivity to varying price schedules, consumers may feel overwhelmed and ignore much of it, it noted. Real-time pricing with caps would work better with programmable, automated loads, which may be realized in the future and not immediately, the paper added.