Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Estimation of energy savings from voluntary conservation programs

Thesis/Dissertation ·
OSTI ID:5482815
Self-selection is one of the major econometric problems that emerges in the evaluation of voluntary energy conservation programs. Two approaches can be used to control sample selection bias. The first approach is the sequential or joint estimation of both a discrete-choice program participation model and a multivariate regression model of energy savings. The second approach is to use first-difference models in conjunction with pooled time-series/cross-section data to adjust the systematic difference between program participants and non-participants. This study attempts to measure the bias of the classical ordinary least-squares estimators, and to compare the efficiency of a menu of alternative estimators for controlling selectivity bias through Monte Carlo experiments. In addition, these estimators are also applied to a data set from a demand-side-management residential energy-conservation program. The study results show that the classical ordinary least-squares estimators are biased, and the first-difference estimators are the best compared to other estimators.
Research Organization:
Colorado Univ., Boulder, CO (United States)
OSTI ID:
5482815
Country of Publication:
United States
Language:
English