Instrumental Variables Approach to Rural Water Service Demand

Interest in empirical studies of the demand for goods sold through discrete pricing structures has grown since Nordin suggested a modification of Taylor's demand model [24; 29]. The bulk of early empirical research centered on residential electricity demand. However, econometric studies of water demand have proliferated in recent years. Some articles have been concerned, in part, with the proper theoretical specification of demand equations incorporating block rate pricing schedules. Taylor specified price as being composed of the marginal price associated with the block in which consumption occurs, and an average price [29]. Nordin modified Taylor's specification by demonstrating utility maximizing consumers with perfect information respond to marginal price along with a lump-sum income effect associated with intramarginal prices and price schedule breakpoints [24]. There is a potential problem in the estimation of demand for goods sold through block rate pricing schemes because of the possibility of correlation between price variables and the error structure [29]. There are, then, two issues associated with the study of demand for goods sold under discrete pricing mechanisms. The first issue is whether if consumers are aware of block pricing structures and respond to marginal price and the lump-sum income change or rate premium, or are aware of only total spending and total consumption and respond to an average price [11; 5; 9; 26]. The second issue is the choice of estimating technique [4; 14; 20; 31; 25]. These two issues are not unrelated since empirical validation of consumer behavior