Which factors determine water saving behaviour? evidence from spanish households

Abstract

In this paper we propose to answer the following questions: What does the adoption of water saving measures by Spanish households depend on? Which factors have the most influence on water saving behaviour? To this end we have conducted an empirical study in which a binary model (probit/logit) with interaction effects between city size, educational level and climatic area, along with other quantitative and qualitative variables, is estimated using household level data. The results of this estimation, together with the evaluation of probability changes that occur when any of the qualitative factors change, suggest that the level of income, the size of the city where the dwelling is, the educational level of the reference person and the climatic location of the city are factors that influence the attitude towards saving water.

Keywords:

• water saving attitudes
• water demand management
• binary models
• interaction effects

Funding

This work was supported by the Science and Education Ministry (Spain) under Grant ECO2009-07936.

Notes

^1. Savenije and van der Zaag (2002, pp. 99–100) define water demand management as ‘the development and implementation of strategies aimed at influencing demand, so as to achieve efficient and sustainable use of a scarce resource’.

^2. For more information about methodological issues see INE (2009).

^3. This decision could be modified. Nevertheless, we think that it satisfies two important requirements: i) this variable is adequate in order to approach household attitudes towards water saving; ii) we have a well-balanced sample of ones and zeros for y_i. This second point is very important to obtain robust results from our analysis.

^4. We estimated a first model:

In this regression the parameters of bedu and lcitsize were not significant at 5%, so we concluded that there were not significant differences between basic and medium educational levels. Analogously there were not significant differences between large- and medium-sized cities. These results allowed us to group the categories where there are not significant differences.

^5. A preliminary estimation found that the income factor does not interact with other factors, so it is considered as independent.