Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

The Millennium Development Goals reflect a collective desire for actions that reverse the problems associated with access to clean water in much of the developing world: high rates of diarrhea and insufficient water supply and sanitation services (WSS). Public policies for WSS must address the usual efficiency criteria, but also significant equity concerns – health, time, and energy costs fall disproportionately on the poor, women and children. Like many developing countries, Sri Lanka has considered the possibility of privatizing WSS services. However, there are concerns that privatization will exacerbate the existing inequalities because of higher tariffs and no service for remote poor communities. Without a detailed understanding of why and how people desire WSS services, it is difficult to increase coverage of these services and address accessibility and affordability concerns of poor households. This paper describes the results of a conjoint survey administered in Sri Lanka to assess demand for drinking water service in terms of price, microbial safety, reliability, and quantity.

Conjoint surveys represent a unique form of a field experiment in which the researcher manipulates the features of a contingent commodity, in this case WSS service, to create options that may not currently exist. Evidence suggests that consumers place value on multiple service attributes, particularly health, and not just costs. People in developing countries use a wide variety of methods to obtain water and rigorous testing of their preferences for WSS service is rare.

Following extensive preparations (focus groups, stakeholder discussions, D-optimal experimental design), the conjoint survey design presented four WSS delivery alternatives defined by 5 attributes. Eighteen hundred respondents from three coastal towns each answered 3 conjoint questions for a total of 5,404 choices.

Conditional and mixed logit analyses confirm our expectations: consumption charges decrease utility, whereas volume, safety, and hours of supply increase utility (the impact is nonlinear for hours of supply). The models also show that private taps and mini-grids are preferred to public taps, but that many preferred to continue with the current status quo. Mixed logit models account for unobserved heterogeneity in tastes across subjects. The mixed logit model suggests that while overall safer water is preferred, respondents displayed a diversity of preferences (i.e., large standard deviations) across the safest and second safest quality water. In addition, we find diverse preferences among the households in terms of quantity and service options, but not supply hours. The interaction terms show that the poor are no different in terms of preferences for volume, microbial safety, and hours of supply. However, consumption charges are a source of greater disutility for the poor. This finding has serious equity implications for allocating scarce public services and achieving universal water access. Overall, we find that demand for alternative water services is low; most households prefer their current option to the service alternatives presented. Thus the feasibility of using the private sector to supply safe water that will attract large numbers of paying customers is questionable.