Alternate Price Specifications for Estimating Residential Water Demand with Fixed Fees

Using a new model formulation and data from a sample of Colorado utilities, we investigated the price specification controversy (marginal price versus average revenue) when estimating residential water demand. The improved statistical fit using average revenue as the price variable was shown to be an artifact of the unitary elastic identity created when monthly rate schedules contain a fixed fee. When the fixed fee was purged from the data, average price was not significant, but marginal price remained significant. In the preferred double-log marginal price model, estimated price elasticity was –0.3, and conservation programs had no significant effect on water use.