Brattle economists have co-authored an article published in the February 2017 issue of Electricity Policy’s Electricity Daily that addresses the issue of econometric models over-forecasting utility sales, resulting in revenue shortfall.

The article points out a variety of factors that contribute to the slowdown in sales growth, including a slowdown in economic growth, rising energy prices, the changing behavior of consumers, utility energy efficiency programs, governmental codes and standards, the rise of distributed generation, and fuel switching. Such factors pose the challenge of quantifying the effects of each and producing a more accurate forecast, which will affect rate design and resource planning for the utility.

To address this challenge, the authors constructed an econometric model to capture the influence of factors with a sufficiently long history, and integrated it with an end-use model to capture the influence of utility energy efficiency programs and governmental codes and standards with shorter histories. This yielded a hybrid forecasting system that combined the best features of the two types of models.

By using Public Service Company of New Mexico as a case study, the authors were able to summarize the structure of the hybrid forecasting system, describe the data that were used to estimate and calibrate the models embodied within the system, and present the results of the sales forecast.

The article, “Overcoming the Over-Forecasting Bias of Pure Econometric Models: A Utility Case Study,” was authored by Brattle Principal Ahmad Faruqui, Associate Josephine Duh, and Ingrid Rohmund, an economist at Applied Energy Group.

Experts Involved
Other Contributing Authors