User‐Specific Water Demand Elasticities

Abstract

Water demand elasticity is estimated for six user categories: residential; commercial; industrial; government; school; and total metered. Eight generalized least‐squares, linear regression models are derived for each user category. Explanatory variables tested are price; per capita income; resident population per user account; housing composition; and summer precipitation. Pooled cross‐sectional and time series annual data from the city of Columbus, Ohio, and incorporated suburbs are used. The recommended regression is a partial adjustment, generalized least‐squares model with cross‐sectional dummy variables. Price is a significant factor for all user categories except industry. The greatest price effect was for school accounts, followed by commercial, government, total metered, and residential. Governmental and school demands are most responsive to fluctuations in real income, followed by residential and total metered accounts. Commercial demand was unresponsive to real income. Both short‐run and long‐run elasticities are available from the partial adjustments model, and the time to reach 90% of the long‐run response varies from three to eight years.