Charting the DSM Sales Slump

impacts, and others rely on expert judgment. Regardless of exactly how DSM impacts are estimated, a separate adjustment outside the sales forecasting model is used to measure these impacts.

Method 5 – Activity Variable: In this method, which is often empirically difficult to apply because of data limitations, utilities insert some measure of DSM programmatic activity as an explanatory variable in their sales forecasting model. The activity measure could be DSM expenditures, number of participating customers or a similar variable. If the variable yields a statistically significant and economically meaningful coefficient, it can then be used to forecast the future impact of DSM in sales.

Validating the Survey Results

In order to validate the results of our survey, we compared our results to a survey done by Hydro One Networks (“Hydro One”) in April 2011. 4

Hydro One had surveyed one hundred organizations in North America and a total of 41 organizations responded, including both Canadian and U.S. utilities. This survey had found that the most common approach to adjusting for DSM was the exogenous adjustment approach of subtracting incremental DSM savings from future impacts. The second most common approach was one where historical DSM program savings were added back in to actual historical sales and an econometric model was then estimated over this reconstituted data. The full impacts from future DSM programs were then subtracted from the gross sales forecasts in order to get a projection of electric sales that are net of DSM impacts.
 Of the 41 respondents to the Hydro One survey, 75 percent of respondents had accounted for DSM using an implicit methodology, 20 percent had accounted for DSM using an explicit methodology such as the gross-up methodology, and five percent of respondents had not reflected DSM in their load forecasts.

The results of the surveys appear quite consistent.

They each point to the fact that most utilities account for DSM impacts in their sales forecasts. Of those that account for DSM, the vast majority make exogenous adjustments to their sales forecasts to account for DSM program impacts. In addition, some other utilities employ the other approaches previously mentioned in order to estimate the impact of DSM programs on utility electric retail sales.

ABOUT THE AUTHORS: Ahmad Faruqui ( and Eric Schulz are economists with The Brattle Group based in San Francisco.


1. For a current discussion on how DSM is deployed in the United States and abroad, please consult the 23 essays in Energy Efficiency: Towards the End of Demand Growth , Fereidoon P. Sioshansi (editor), Academic Press, 2013.

2. See, for example, the application of Xcel Energy in Minnesota where the company is requesting a rate increase partly to recover fixed costs that otherwise would not be recovered because sales growth has slowed down due to its DSM programs. Especially relevant to this paper are Faruqui’s rebuttal and direct testimonies which can be found at: and

3. For a discussion of these factors, please see the paper by Ahmad Faruqui and Eric Shultz entitled, “ Demand Growth and the New

A survey of rate case methods for sales forecasting.
A survey of rate case methods for sales forecasting.
Intro Text: 
Demand side management has a growing effect on energy sales. Utilities are applying five methods to account for DSM in sales forecasts. A Brattle Group survey reveals those methods and their characteristics.
Publishing Date: 
Wednesday, May 1, 2013 (All day)