Instrumental Price Estimates and Residential Water Demand

Instrumental estimates of two price specifications, one motivated by the consumer decision problem given full information about rates and charges and the other an average price formulation, are developed to correct for measurement error when residential water sales are made at a schedule of rates, rather than at uniform prices. Annual water purchases of single-family residences are regressed on these instrumental price estimates, family income, and household size by ordinary least squares, based on a sample of 326 observations from metropolitan Denver, Colorado, for 1976. The resulting demand estimates are robust to the price concept specified, given proportional variation in all rates and charges, and are consistent with findings in the literature. The overall price elasticity estimates range between −0.14 in a linear model to −0.44 in a log-log model, while the estimated income elasticity varies between 0.40 and 0.55.