The demand for water: Some empirical evidence on consumers' Response to a commodity uncertain in supply

Abstract

Based on a theoretical model that studies how individuals respond to a commodity uncertain in supply, this paper estimates some demand functions for water in situations where supplies are known to be stochastic. The model is applied to data consisting of a sample of Massachusetts towns, separate functions being estimated for household and in-dustriM demand. Two cross sections are estimated, one for 1962 and one for 1965, the beginning and the end, respectively, of the New England drought.As far as household demand is concerned, price, uncertainty as measured by supply variance, and particularly housing space are significant, whereas for industrial demand only the first two variables are significant. The results also indicate that firms are more responsive than households to both price and uncertainty and that their response to uncertainty lessened as a result of the drought, an implication that agrees with the theoretical model. ess [Turnovsky, 1969].Much of the work in Turnovsky [1968, 1969] is devoted to determining how the individual responds to parameter changes and in particular how he responds to uncertainty as measured by Ox?/Oa•% In this paper we endeavor to test empirically some of the implications of the theoretically derived model. (It is not necessary to give any details of the model here. The interested reader is referred to Turnovsky [1968, 1969], where they are described in detail. For our purposes all that is relevant is that they give rise to a set of demand functions of the form (1.2).) We have chosen to apply the model to the estimation of the planned demand for water, since in certain regions water supply has been extremely variable in supply during recent years. Moreover, our theory deals specifically with nonstorable commodities and, as most individuals do not have facilities to store water, water is a particularly appropriate commodity to act as a testing ground.There are, however, several stages in transforming the function (1.2) into one suitable for practical estimation and testing. First is the problem of aggregating these individual rela- 350 The Demand for Water 351 tionships to levels at which data are available. Very few data describing individual behavior are available, so that some aggregation is inevitable. In the application of this paper, data are available at the municipal level, so that aggregation to this level is implicit in our empirical work. Whether aggregate demands will have the same characteristics as individual demands is not always clear and will depend, for example, on the relative weights of the individuals in the aggregation. However, if we assume, as we shall, that all individuals respond in the same direction to all parameters, then the aggregate response will behave similarly.

Urban Water Demand for Service and Industrial Use: The Case of Zaragoza

“…More functional approach identifies water demand by end-use than by customer class (Bell and Griffin 2008). In this respect, the most usual functional forms used are the linear (García-Valiñas 2005a;Turnovsky 1969), the Cobb-Douglas (DeRooy 1974;Renzetti 1993;Ziegler and Bell 1984) and flexible functional forms such as the translogarithmic (Babin et al 1982;Dupont and Renzetti 2001;Féres and Reynaud 2005;Grebenstein and Field 1979;Renzetti 1992;Reynaud 2003).…”

“…A review of the analytical and methodological issues of these studies allows us to distinguish two basic approaches: single-equation models, either to jointly estimate the water demand of industry and services (Rees 1969;Turnovsky 1969;Ziegler and Bell 1984) or proposing separate equations for industry and services (Malla and Gopalakrishnan 1999;Williams and Suh 1986), and demand-system models to estimate the demand for different inputs implied in the productive process (Dupont and Renzetti 2001;Féres and Reynaud 2005;Renzetti 1992;Reynaud 2003). In the case of services and industries, water is commonly treated as one more input to consider in the production function.…”

“…The option chosen in most papers is to construct a proxy of the level of output. The most used are indices of industrial production (Turnovsky 1969), added values in absolute terms (Dupont and Renzetti 2001;Renzetti 1992Renzetti , 1993Renzetti and Dupont 2003;Williams and Suh 1986) or in relative terms (DeRooy 1974), employment level (De Gispert 1999) and/or geographical location (García-Valiñas 2005b). Together with the level of output, it is common to use another variable to reflect the technological characteristics of the industry (DeRooy 1974;Féres and Reynaud 2005;Ziegler and Bell 1984).…”

INDUSTRIAL WATER DEMAND WITH WATER REUSE1

“…However, IWD studies have been very limited because of the difficulty in gathering internal data and the complicated nature of industrial processes. In conventional economic studies, industrial water demand is derived from optimization behavior of firms (Turnovsky, 1969;DeRooy, 1974). The cost function is generally approximated by either a Cobb-Douglas form (Ziegler and Bell, 1984;Renzetti, 1988) or a Translog form (Grebenstein and Field, 1979;Renzetti, 1992;Reynaud, 2003).…”

“Effect of Price on the Residential Demand for Water Within an Agency”2

“…Using models similar to those developed in this District's study, these researchers estimated elasticities of 0.7 for western region outdoor use and 0.231 for domestic use (Howe and Linaweaver, 1967). Turnovsky found price elasticities of between 0.05 and 0.40 for the total residential consumption of 19 Massachusetts towns (Turnovsky, 1969). Wong, using data similar to that used in the current District study (pooled cross-sectional), found price elasticities between 0.02 to 0.28 in the Chicago area (Wong, 1972).…”