Instrumental Variables Approach to Rural Water Service Demand

Ordinary least squares (OLS) demand estimates and alternative instrumental variable techniques for studying the demand for domestic rural water service sold under block rate pricing schedules are developed and compared. The data used to estimate the demand and to examine measurement error and simultaneity issues are taken from a survey of stratified, randomly sampled Illinios rural water district users. The user sample comes from 9 of the 59 Farmers Home Administration-financed Illinois rural water districts. The final sample consists of 641 monthly observations over 54 households for the year 1982. The estimated models incorporate the full information specification of demand suggested by Nordin (1976) and include a marginal price and lump-sum income effect or difference variable. The analysis suggests rejecting the OLS estimate in favor of the instrumental approach, and it does not validate the Nordin specification. However, the OLS estimate is not much different from the average estimate of the instrumental variable approach.