Journal of Econometrics 26 (1984) 83-113.
COSTS AND BENEFITS OF PEAK-LOAD PRICING OF
A Continuous-Time Econometric Approach*
A. Ronald GALLANT
North Carolina State University, Raleigh, NC 27695, USA
Roger W. KOENKER
University of Illinois, Urbana, IL 61801, USA
We address the following question of current policy interest: Would the efficiency gains from residential time-of-use pricing for electricity exceed -the metering costs necessitated by these more complex rates?
A model of consumer preferences for daily electricity consumption is estimated based on data from the North Carolina Rate Experiment. The model is formulated in continuous time and thus is capable of evaluating demand responses and welfare consequences of quite arbitrary changes in pricing policy. A model of long-run electricity costs - viewed as a functional of the daily load cycle - is constructed based on engineering data. The models of demand and cost are combined to compute solutions to several optimal pricing problems and to estimate the potential long-run welfare gain from several alternative time-of-use pricing policies including policies incorporating so-called 'demand charges'.
We find that the best of the rate treatments used in the North Carolina experiment achieves a net welfare gain of 5¢ per day per household, or roughly half the cost of current metering equipment. Smoothly varying rates are capable of achieving nearly 180 per day per household, but would require more complex metering. Demand charges while they are quite successful in smoothing the demand cycle are not as successful as conventional pricing policies in achieving our welfare objective.