An Evaluation of the Effects of a Pricing Policy on the Water Consumption of Heterogeneous Households in Seattle

Abstract

In 2001, the City of Seattle implemented a new pricing policy for residential water consumption, intended to target high water users. This dissertation evaluates the impact of this policy and contributes to the water demand literature in three ways. First, this dissertation extends the use of instrumental variables to represent the effect of nonlinear price structures on heterogeneous users. Correlated random coefficient models, specified with coefficients that vary by groups such as households and neighborhoods, allow a realistic specification of heterogeneity while addressing simultaneity bias. Price elasticity estimates for the entire population range from -0.15 to -0.52, depending on the model used and the point of consumption, and are consistent with the literature. Second, this dissertation uses a hierarchical linear model to link the varying intercepts and coefficients to the underlying properties of heterogeneous individuals and groups, and to allow meaningful, policy-relevant interpretation of the estimates, as well as a sensitivity analysis of factors affecting the policy impacts. A household-level model finds an average price elasticity of approximately -0.41; the average household price elasticity also varies as expected with lot size and house value. Third, based on these predictive model results, this dissertation explores the statistical properties of welfare measures for heterogeneous groups in order to characterize the distributive impact of this pricing policy. Estimated losses in consumer surplus as a result of the pricing policy range from 5-45% of existing total bills.