Post-2015 Consensus: Water Resource Management Perspective, Whittington

Perspective Paper

Summary

Both the UN Panel of Eminent Persons and the Open Working Group for Sustainable Development Goals struggled with how to frame goals for the broader water resources management challenge. It is easy to agree that water resources should be used more efficiently, but that does not necessarily mean minimizing water use. Most water problems are essentially local, and tailored local solutions must be developed. These must also recognize that water is a renewable resource; after use by humans, each molecule of water reenters the natural hydrological cycle.

All this means that it does not make economic sense to present an average benefit-cost ratio for investment in water infrastructure such as dams. Some projects make good economic sense, while others do not. One approach to providing a figure for comparison with other options under consideration in this project would be to use a baseline of some of the most economically attractive dams in the world and assume that the BCR for others would be lower. Using this approach, we find that even the large multipurpose dam in the Blue Nile gorge in Ethiopia – one of the most attractive places in the world for hydropower generation – has a much lower benefit-cost ratio than smaller-scale WASH interventions.

Health policy interventions (e.g., vaccination, malaria, HIV-AIDS, hospital care) consistently dominate the Copenhagen Consensus rankings, whereas these are generally at the bottom of the list of development priorities in low-income countries. Similarly, they are not at the top of the priority list for many households in developing countries. I suggest there are five systemic reasons why the Copenhagen Consensus process overlooks the economic importance of large-scale infrastructure projects.

First, analysts underestimate the costs of targeted health and social interventions, in part because they implicitly assume that infrastructure is in place to enable the efficient delivery of health services. Ex-post cost-benefit analysis of infrastructucture projects often reveals cost overruns, and analysts responsible for the appraisal of large water resources projects have long been aware of this issue. More ex post cost-benefit analysis of targeted health and social policy interventions is needed to temper overly optimistic cost estimates.

Second, analysts overestimate the benefits of targeted health interventions, in part because arbitrary assumptions are made about the value of mortality risk reductions. In practice, poor households place a surprisingly low value on reduced mortality risk. Since the value of lives saved often dominate the benefits, this can inflate the apparent attractiveness of health projects, whereas this is a less important part of the benefits of infrastructure projects.

The third reason is the difference in the planning horizon, which is much longer for infrastructure projects and hence makes proper comparison difficult. Fourth, projects such as dams can often get private sector funding and may fall outside the scope of international aid and, finally, the Nobel Panel appears to demand very high-quality evidence, often only available from randomised trials where strong causal links have been demonstrated.

There are also a number of additional challenges in the analysis of large-scale water resource projects. Since the use of water is so all-pervasive, it is rarely possible to include all the economic benefits of improved water service. It is also difficult to account for both the positive and negative impacts of reduced hydrological variability, particularly as water resources cross administrative and political boundaries. Finally, the causal links between investments in water infrastructure and economic growth run in both directions.

This discussion raises the question of what can be done to improve the Copenhagen Consensus Project’s approach. I have two suggestions. First, sector specialists should be asked to provide more evidence from ex-post cost-benefit analyses to support the findings from their ex-ante studies. The second is to divide sector policy proposals into short and long-range interventions, using some necessarily arbitrary threshold (e.g. 10 years) that would separate large-scale infrastructure projects from the “simple” delivery of health and social services. The Nobel Panel could rank proposals in each category separately, making a judgment on the fraction of total funding to be allocated to each.