Bad Energy Forecasts: Not Just a Federal Problem

Failing to adapt to changing trends isn’t just a federal problem. It also affects regional forecasts of energy demand that have major implications for environmental and energy policy.

This post was written by Jeff Inglis and Tony Dutzik, with research assistance from Gideon Weissman

Michael Grunwald’s June 24 Politico article about the startlingly low U.S. government forecasts for growth in renewable energy threw the spotlight light yet again on the ways in which outdated assumptions can contribute to misguided public policy.

The biggest problem with the Energy Information Administration (EIA) forecasts Grunwald lampoons is not their failure to predict the distant future, which is always difficult. Rather, it is their failure to properly account for events happening right now. For example, Grunwald writes:

The agency’s latest reference case suggests solar capacity will double from 2014 levels by 2026. But … an industry analysis based on actual projects in the pipeline has projected that solar capacity will double by next year—and so far it’s on track to do just that.

Solar energy prices have been dropping rapidly in recent years, with the price of residential and commercial systems falling by 12 to 15 percent annually in 2012 and 2013 [PDF], with similar declines in 2014. Yet, according to the Union of Concerned Scientists, the cost assumptions used in EIA’s modeling do not fully reflect the recent drop in costs for renewable energy.

Failing to adapt to changing circumstances isn’t just a federal problem, however. It also affects regional forecasts of energy demand that have major implications for environmental and energy policy.

Here in New England, our regional grid operator, ISO-New England, makes annual demand forecasts that shape key decisions surrounding the management of the region’s electricity system. Like the EIA, ISO-New England has been behind the curve in adapting to changes in energy demand. The following chart compares actual data on the amount of electricity needed to serve demands of the region’s grid with ISO-NE forecasts since 2005. (See Figure 1.) By 2014, New England was using about 14 percent less power from the grid than had been forecast just a decade earlier.

Figure 1. Actual ISO-NE Net Energy for Load, Compared with Recent Forecasts [from ISO-New England annual Capacity, Energy, Load and Transmission (CELT) reports]


Unlike the EIA, ISO-NE has been aggressively working to adapt its forecasting methods to changing conditions. In recent years, ISO-NE has published an annual energy efficiency forecast, which takes into account the role of the region’s top-ranked energy efficiency programs in reducing energy demand, as well as a distributed generation forecast that accounts for rapid growth of solar power in the region.

But even these efforts have been flawed. ISO-NE’s energy efficiency forecast assumes that energy-saving measures will rapidly become more costly and less effective, based on the notion that the “low-hanging fruit” of efficiency savings will be picked over time, leaving only more costly savings opportunities. ISO-NE’s own data, however, show that energy efficiency costs haven’t increased appreciably in recent years, despite growing energy savings from those programs. (See Figure 2 below.)  In Massachusetts, for example, each unit of electricity saved through efficiency programs in 2013 cost 20 percent less than ISO-NE had forecast in a document published just a year earlier.  

Figure 2: ISO-New England Energy Efficiency Cost Projections, Massachusetts (from ISO-NE Energy Efficiency Forecasts)

Meanwhile, ISO-NE’s solar forecast (PDF) assumes that the growth in solar energy in the region will slow in the years to come, based in part on anticipated policy changes and in part on solar cost forecasts from – where else – EIA’s Annual Energy Outlook. [1]

All of this would be merely an interesting academic topic if there weren’t real stakes involved. Currently, New England policy-makers are considering spending billions of ratepayer dollars to expand natural gas pipeline capacity in order to reduce the risk of future shortages.  Doing so would increase the region’s dependence on natural gas and undermine efforts to reduce greenhouse gas emissions. If we are going to make those and other decisions wisely, we need to incorporate the latest data into forecasts of our energy future, and also acknowledge sources of uncertainty that could make today’s forecasts look like tomorrow’s foolish delusions.


For more examples of failure to account for changing technologies in energy forecasting, see this excellent piece by Eric Gimon and Sonia Aggarwal of America’s Power Plan in GreenTech Media.


And for more on the ins and outs of New England’s battle over natural gas pipelines, see this excellent three-part series by analysts at Acadia Center.





[1] Based on ICF International study (PDF) commissioned by ISO-NE to inform the 2015 PV forecast.