Customers Pay When Big Utilities Make Big Errors In Electricity Forecasts

Date: 25 Jul 2019 | posted in: Energy, Energy Self Reliant States | 0 Facebooktwitterredditmail

In over 30 states, government agencies count on monopoly electric utilities to accurately forecast how many power plants to build based on how much energy customers will demand in the future. With no competition, these utilities have the sole responsibility for ensuring supply matches demand. On these forecasts hinge billions of dollars in capital investment and costs to consumers, and millions in shareholder profits.

Unfortunately, despite (or because of) decades of experience, utility forecasts often miss the mark by a wide margin.

In an investigation of electric utility forecasts from seven of the largest ten U.S. electric utilities in the past decade, the Institute for Local Self-Reliance found that the average forecast overestimated demand by 17% over three years. Over longer time periods, the forecasts were even less accurate. The errors, all overestimating use, reinforce an environment in which utility companies––with no market competition––can win permission from public regulators to build unnecessary power plants. The costs can be passed to consumers while the rewards accrue to the utility’s shareholders.

In 33 states, monopoly utilities are required to submit integrated resource plans (IRPs) to state public utilities commissions for approval. These plans include intentions to build or buy new power generation resources (like the Mankato gas plant Xcel intends to buy in Minnesota). The plans also detail the utility’s net energy forecast, the total amount of power that the utility expects to sell. State regulators rely on these utility forecasts to determine whether proposed utility actions are in the public interest.

Unfortunately, the last decade has revealed that major utilities are poor predictors of future energy demand. ILSR examined net energy forecast data from seven of the largest U.S. utilities over the past decade. The forecast data was compared to actual energy sales data for those same utilities. ILSR’s findings are summarized in the graphs below. The blue lines show utility forecasts filed with regulators, over three to five successive years. The orange lines illustrate actual energy sales in millions of megawatt-hours.

Of all the utilities examined, Florida Power & Light made the biggest changes to its forecast based on actual sales data. Between 2007 and 2008, the utility sharply reduced its forecast for the years 2010-2016 to more closely align with sales. However, even its revised forecast overestimated use by 5 percent just six years later. Other utilities missed the mark by far more.

These graphs starkly demonstrate that utilities have tended to wildly overestimate future energy sales. In fact, while every utility forecast examined predicted an increase in energy sales, actual sales stayed relatively constant or even decreased over time. The average difference between utility forecasts and actual sales from the oldest data ILSR examined (2006 to 2008, depending on the utility) was 12 million megawatt hours. That is equivalent to the annual output of two or three standard, 500-megawatt nuclear reactors and enough to power 1.4 million homes.

How might a utility stand to benefit from an erroneous forecast? Utilities aren’t building new nuclear plants, but they have been on a building spree of gas power plants in recent years. Based on the average annual run-time of a U.S. combined cycle gas plant, a utility would need almost five 500-megawatt gas plants to meet their average forecast overage of 12 million megawatt-hours per year. A gas plant that size likely costs $650 million upfront and another $3 billion in fuel costs over its lifetime. For five plants, that’s $3.2 billion upfront. Customers would be responsible for the entire $3.2 billion (and likely all the $15 billion in fuel costs), while the utilities making these forecasts earn close to a 10% profit on each dollar spent. In other words, planning to build $3.2 billion in gas plants to meet a bad forecast could mean $320 million in additional profits.

ILSR surveyed just the seven of the 10 largest U.S. utilities, but Rocky Mountain Institute did a similar survey of peak demand forecasts across many more utilities with similar findings. Over a 10-year timeframe, utilities averaged a 12 percent over-forecast in peak energy use, reinforcing a business model motivation to build more capacity.

Chart from Rocky Mountain Institute

Utilities have exchanged protection from competition for the oversight of public regulators, with the agreement that regulators protect the public interest. The data on utility forecasts suggests a need for strong scrutiny by Public Utilities Commissions, otherwise electric customers from California to Georgia to Minnesota could be padding utility shareholder pocketbooks for unnecessary power plants.

This article originally posted at For timely updates, follow John Farrell on Twitter or get the Energy Democracy weekly update.

Featured photo credit: David via Flickr (CC BY 2.0)

Avatar photo
Follow Alden Aaberg:
Alden Aaberg

Alden Aaberg worked as intern with the Energy Democracy Initiative in Summer 2019. In this role, he contributed to research and podcast post content.

Avatar photo
Follow John Farrell:
John Farrell

John Farrell directs the Energy Democracy initiative at the Institute for Local Self-Reliance and he develops tools that allow communities to take charge of their energy future, and pursue the maximum economic benefits of the transition to 100% renewable power.