The Determinants of Residential Water Demand. Empirical evidence for a panel of Italian municipalities

This paper presents empirical evidence on the determinants of water demand for domestic use in one Italian region, the Emilia-Romagna, by using municipal data. Two main stems in urban/domestic demand analysis can be found in the empirical literature. The first deals with the estimation of price or income demand elasticities in the short and in the long run. The price demand elasticities can be used for water demand management purposes while the income price elasticities can be useful in the forecasting process of the water requirements. The second one deals with the estimate of consumer willingness to pay for increasing in water service quality in holistic sense or concerning single characteristics of the service: safety, flavour, continuity, appearance, pollution rate and cost. The aim of the analysis is in this case the elicitation of the direct use, indirect use and non-use values associated to the water resource consumption, by means of direct or indirect techniques. In this paper we focused the analysis in the first stem of the empirical literature in which a cross section data set is required. The paper explores the topic problems of the estimating process with the analysis of the empirical literature (with particular regard to investigations that use municipal data) and with the analysis of the econometric problems related to the demand estimate. The theoretical model for the water demand analysis is also presented and discussed. Two datasets have been implemented: one with 125 municipalities and four years, the other with 40 municipalities and eleven years. Both the datasets bring together municipal water consumption and tariffs data provided by local water utilities and other municipal data (inhabitants, surface, households, income, etc.) stemming from official sources. The econometric analysis is based on both fixed effects, performing better than random effects models, and dynamic panel models. The estimated coefficient of the tariff variable arises always significant and with negative sign; the water demand price elasticity is negative with a value between -0.88 and -1.11, but not significantly different from one, considering the different specifications. The results for the income variable are quite different: in the basic model specification the estimated coefficient is positive and significant while the introduction of other socio-economic variables change the significance level and sometimes the sign. This applied study is an important starting point for the Italian environment, which lacks structured integrated datasets and consequently reliable estimates on elasticities concerning micro-economic oriented water demand studies. However, further analyses with more municipalities and more years have to be carried out to generalise and made results more robust, since the estimation of price elasticity and the investigation on the determinants of water demand is necessary information for both private and private-public management of water resources