Residential Water Demand: A Quantile Regression Approach

Quantile regression has been introduced into residential water demand. Differences between quantile regression and conditional mean regression are found on extreme quantiles. Monotonically increasing price elasticities are found by quantile regression. The hypothesis of a "U-shaped" elasticity pattern can only be partially established for high quantiles. Natural experiments are used to compare price elasticity among comparable cross sectional sub samples, which empirically indicates: 1) lowering break-over thresholds for block rate makes affected consumers more price elastic; 2) high quantile households are less price elastic in the summer than the winter; and 3) low quantile households and those with smaller yards and pools are more price elastic.

MsThesis_WangXR_FINAL.pdf (2.39 MB)

Author(s)

Publication Date