Letters to the Editor
May 2004
To the Editor:
The article "NERC's Cloudy Crystal Ball" (March 2004, p. 61) contends that the North American Electric Reliability Council (NERC) has consistently underestimated the growth in U.S. electricity demand. The only evidence offered for this conclusion is that observed data did not encircle the 45-degree line in a graph of actual vs. forecast percentage growth rates. Conjectures such as this are invalid for numerous reasons.
First, average peaking temperatures are used to determine projections of electricity demand while actual climatic conditions influence reported peak demand. Because the author relied only on percentage changes between actual and projected demand, misalignment of temperature assumptions is likely to lead to imprecision.
Second, electricity demand forecasts are reduced by interruptible and curtailable loads. Large industrial customers, for example, will often agree to shed load when capacity is constricted in exchange for lower tariffs. When sufficient capacity exists, however, these loads are served because it is economic to do so. Any direct comparison between actual and projected load must take this circumstance into account.
Third, electricity demand projections are long-range in nature due to the extended asset lives of generating, transmission, and distribution capacity. Accordingly, demand projections disregard short-term business cycles that are transitory in character. This approach to electricity demand forecasting is appropriate because: (1) electrical consumption generally follows overall economic conditions that cycle around a longer-term trend (see Figure 1); and (2) capacity reserves are held to meet unexpected but temporary demand.
Finally, like all forecasting protocols, NERC members update their annual projections using the most recent observable data. As long-range projections are updated using recent history, they will tend to embody the actual economic conditions existing at the time of the forecast. Consequently, annual projections made over an extended period will themselves cycle around a longer-term trend. Figure 2 shows the aggregated NERC projections over the period 1998-2001 tend to mean revert (80 percent statistical confidence bands also are shown as a gauge of projection uncertainty).
The U.S. economic expansion from March 1991 through November 2001 was the longest on record during the post-World War II era. Figure 2 illustrates that electricity demand forecasts made at the end of the economic expansion reverted to the long-run trend. Accordingly, the cycling pattern of short-term actual demand also explains the cycling pattern of long-term projected demand.
In the first sentence of the article, the author encourages the use of reasonable, sound and sensible approaches to properly estimate the most appropriate magnitude of long-run peak demand. I couldn't agree more.
John Harris, Ph.D.
Chair, NERC Load Forecasting Working Group
The Author Responds:
Thank you for your feedback. I carefully ensured that the NERC projections used in my analysis were not reduced by interruptible or curtailable loads. Perhaps some of the historical peaks had been reduced through curtailment, but that would only mean NERC under-forecasted the total demand for megawatts by an even larger margin than my analysis suggests.
While you are right to point out my dataset includes the largest economic expansion in U.S. history, please also note that it contains two recessions: 1991 and 2001. I accept that the economic cycle and/or unusually hot or cold summers lead to forecast inaccuracies. However, my article focuses on demonstrating statistical bias in NERC projections rather than commenting on their accuracy. History has revealed that regardless of boom, bust, hot, or cold, NERC consistently forecasted lower U.S. peaks than actually occurred 5 and 10 years out. -Tom Replogle, ICF Consulting
Articles found on this page are available to subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.
Public
Utilities Fortnightly & Spark
