5 Dec 2008

Water and Sanitation in Developing Countries

The Copenhagen Consensus Centre (CCC) is working to prioritize the international community's effort to solve the world's biggest challenges in the most cost-efficient manner.

In 2004, experts comissioned by the CCC ranked 25 different programs to address problems (HIV, global warming, water sanitation, etc.) based on greatest bang for the buck. HIV came out at #1, water sanitation projects took places 6-8 and climate change projects were placed last on the list. (Many people dislike Bjorn Lomborg, director of the CCC for his skepticism of climate change.)

Anyway, I am not here to debate that question or even the ranking, which I think is mostly right. (I'd divide the list into projects the developed world should address -- climate change -- and the developing world should address -- water sanitation.)

I am here to discuss the background papers prepared for CC 2004 and CC 2008. I think they make good reading, and -- a rarity -- they reveal the tension among researchers making different assumptions and using different models.

In the lead CC 2004 paper [pdf], Rijsberman says that the main problem is not a shortage of water, but the distribution of water and who bears the costs of distribution. He suggests that water should be managed at the community level, on a small scale. He concludes with:
Providing community-managed access to low-cost water supply and sanitation is a major opportunity to increase global welfare. This requires an integrated approach to water supply, sanitation and hygiene education. The focus should be on service delivery rather than on infrastructure – and sanitation should receive at least as much attention, and more funding, than water supply. Social marketing and micro-credit programs are key ingredients of the package. The benefits to society range from improved health, particularly a reduction in diarrhoeal diseases, and reduced loss of time engaged in gaining access to water and sanitation but particularly reduction of time lost to disease. The total net benefits of making a significant progress towards increasing access, defined as MDG target 10, or halving the people without access by 2015, are estimated to be US$300-400 billion.
In a comment [PDF] on Rijsberman's paper, Vaux points out the importance of institutions for reallocating water, addressing/preventing pollution, and preventing overdrafting of groundwater in areas that are already vulnerable. Vaux basically says that Rijsberman's cost-benefit analysis means nothing if available money cannot be used effectively.

In a second comment [pdf], Boland discusses the complexity of water and sanitation (commodity? public good?) and notes the importance of understanding local culture and institutions and the need to have a pragmatic plan for phased-in adoption. He concludes by saying that:
particular attention should be given to the two issues for which there is little satisfactory experience: (1) the means of adapting interventions to local conditions and (2) the strategies for achieving financially sustainable delivery of services. In the case of the second opportunity, efforts to promote the adoption of low-cost technologies must begin with a clear and detailed understanding of why those technologies will not be adopted in the absence of intervention.
In 2008, the CCC asked experts to rank a new set of 30 projects to address global development. Their top choice (in terms of benefits resulting from expenditure) was micronutrient supplements for children. Water projects came in at 15, 16, 20 and 24.

As part of the process, the CCC commissioned a new set of papers to discuss water and sanitation.

The lead paper [PDF] [updated DOC] by Whittington et al. (including my adviser, Michael Hanemann) makes the important point that history matters, i.e., existing water and sewer systems in the developed world resulted from a need to fight fires, deal with institutionalized pollution (discharge of untreated black water), etc. Their main point is that these systems cannot be implemented in the developing world without consideration of local conditions. Most important -- in terms of cost-benefit analysis -- is the fact that many people in developing countries already have some access to water. When these people are given the choice between improved (cleaner) water and initial electricity, they choose electricity. Their lower willingness-to-pay makes it harder to justify expensive projects, even when funding comes from elsewhere.

They also have an interesting discussion of costs:
In industrialized and developing countries alike, most people are unaware of the magnitude of the true economic costs of municipal water and sanitation network services. There are several reasons why these economic costs are so poorly understood.

First, the capital costs are heavily subsidized by higher levels of government, (and, in developing countries, by international donors), so that households with services do not see the true capital costs reflected in the volumetric prices they pay. Second, in many cities tariff structures are designed so that industrial water users subsidize residential users; households thus do not even see the full operation and maintenance costs in the prices they pay. Third, because many water utilities run financial deficits (in effect running down the value of their capital stock), water users in aggregate do not even see the full costs of supply. Fourth, most cities do not pay for their raw water supplies: typically the water is simply expropriated from any existing water sources (and their users) in outlying rural areas. Fifth, wastewater externalities are typically imposed on others (downstream) without compensation.

Sixth, the subsidies provided to consumers of water and sanitation services are not only huge, but also regressive. It is often not politically desirable for the majority of people to understand that middle- and upper-income households, who generally use more water, are thus actually receiving the most benefit from subsidies. Tariff designs may in fact be made overly complicated in order to offset this reality and appear to be helping poorer households (Komives et al., 2005). Most fundamentally, poor households are often not connected to the water and sanitation network at all and hence cannot receive the subsidized services. Even if they do have connections, the poor use less water than richer households, thus receiving lower absolute amounts of subsidy.
Last, I want to highlight the authors' excellent method of estimating project benefits. Because parameter value and interactions are uncertain, the authors assign distributions to each value and then calculate probability distributions for outcomes, showing the percentage likelihood that a given project will produce benefits in excess of costs. This method is far more complicated to use than simple point estimates, but it also provides a far more realistic perspective for evaluation.

This paper also had two comment papers. In the first [PDF], Davis points out that Whittington et al.'s conclusion that many water projects fail to achieve meaningful benefit/cost ratios may be the result of paying attention to full infrastructure projects (ignoring simpler improvements such as latrines and hand pumps) and ignoring important costs of inaction (higher mortality from dirty water). In the second [PDF], Rijsberman (2004 main paper author) and Zwane make much the same point. [The papers are not the same, but I can't find bigger differences.]

Bottom Line: Read these papers if you are interested in water and sanitation in developing countries.