Household responses to increased water rates during the California drought

This paper explores the use of fixed effects and maximum likelihood techniques to estimate household responses to water price increases during the California drought. Estimates are based on bimonthly meter readings from 599 single-family households in the Alameda County Water District over the period 1982-1992, before mid after the introduction of a steeply increasing block rate price structure. I find that household and monthly fired effects models are not successful in modeling water demand with these data. However, maximum likelihood models that explicitly consider the household's response to the rare structure result in plausible estimates of water demand. (JEL Q25).