One of the policy interventions that one might use to improve water and sanitation conditions in developing countries is to change the planning approach. I'm going to present to you a simple topology of four different types of planning protocols that are used around the world. I want you to imagine that you are responsible for designing a rural water supply program. I want you think about which one of the four types that you would be most likely to use. Here's my simple typology. This 2 x 2 table shows four cases A, B, C, and D. The columns describe whether or not the state uses a planning protocol that needs information on household preferences or demand. The rows describe whether important planning decisions largely rest with high level government, or with lower level government or community organizations. If decisions largely rest at lower levels of government, we often use the term decentralization, or sometimes, devolution. If instead, decisions rest within formal community organizations or groups of households, we use the term participatory. Let's discuss the advantages and disadvantages of each of these four cases. Like all topologies, this one is overly simplistic. But hopefully, it helps us to think more critically about the differences between planning protocols that can be deployed in the WASH sector. Let's start with case A. In case A, WASH planners do not collect information on households' preferences. Donors or national government used a top-down planning approach. All the important decisions on matters, such as site selection, technology, pricing, institutional arrangements, rest with top-level government. Centralized top-down planning approaches often work best when there's a clear mission, like putting a man on the moon or eliminating smallpox. For many people working in the WASH sector, their mission feels just as clear. They feel that we are already know what needs to be done. We need to provide 100% of the world's population with clean, safe drinking water and hygienic sanitation facilities. This means we need to build new water and sanitation infrastructure everywhere it doesn't already exist. In 2010, the United States Congress adopted legislation that was broadly consistent with this case A planning approach. The Senate Foreign Relations Committee approved legislation launching a government effort to provide 100 million people in the world's poorest countries with first-time access to safe drinking water and modern sanitation. The bill established an Office of Water within the United Stated Agency for International Development to implement country-specific strategies. They decided this measure would establish programs in poor nations to build water wells, treatment facilities, and latrines. It would also promote the use of low-cost sanitation technologies such as hand-washing stations. So the United States Agency for International Development would decide, on its own, WASH strategies for countries and had seemingly already decided on the specific technologies to deploy. Proponents of case A cite several advantages of this planning protocol. First, if you already know what needs to be done, this planning approach enables you to stay focused on your objective. There's no need to waste time or money on studies or to consult with communities. Second, a centralized approach prevents the effort from being "diluted" or "derailed" by local institutions or local politics. Now let's think for a moment about the disadvantages of a case A planning approach. Do you recognize this man? When I showed this photo to my students here at the University of Manchester, they didn't. This is a photo of Friedrich Hayek. Hayek was an economics professor at the London School of Economics in England, and later at the University of Chicago in the United States. Hayek won the Nobel Prize in economics in 1974 for his work on the information challenges faced by central planners trying to manage socialist economies. He referred to the economic calculation problem. It was created by dispersed knowledge. By dispersed knowledge, he meant the information that households held, but that central planners really needed. The central planners had no feasible way of obtaining all the dispersed knowledge they required to do the economic calculations necessary to run the economy effectively. So here's a question for us in the WASH sector. When you think about the most appropriate planning protocol to use, is improving water and sanitation conditions in low income countries like managing a socialist economy? Or, is it like putting a man on the moon? In other words, is it feasible for a donor or national government to obtain all the information needed to make all the important decisions on site selection, of new facilities, on type of technology, on pricing, and on institutional arrangements? One of the disadvantages of a top-down planning approach is that it may not be obvious what should be done or what people want. Moreover, the information needed for planning may not be available to a central planner. Another disadvantage of the centralized, top-down planning approach is that the project may fail altogether without local support or buy-in. Now let's move on to case B in our topology. In case B, project staff do collect information on households' preferences in order to inform their decisions. But the national government or donor still uses a top-down planning approach. All the important decisions about site selection, technology, pricing, and institutional arrangements still remain with the state. Cases A and B are closely related. What distinguishes them is the type and extent of demand information that the state believes it needs for planning purposes. For example, the state's planning protocol might simply need to know household water and sanitation coverage to estimate what needs to be done. Or, its planning protocol might require coverage and household information data. Or, much more information might be needed about things like households' knowledge, perceptions, and beliefs about health and hygiene. But for both case A and case B, once the state has the information it needs, it decides what to do. There are three main advantages of a case B planning protocol. First, additional information can be incorporated into the planning process, compared to case A. Second, like case A, case B avoids time-consuming community involvement. Third, the state has the opportunity to adjust or change household information it deems unreliable or inaccurate. But there are several disadvantages of case B. First, the information requirements are even more onerous than case A. And the planning process becomes more time-consuming and expensive. Second, if people are uninformed, short-sighted, or simply don't know what they need, there's little point in collecting information about their preferences. Now let's look at case C. In case C, the state does not collect information on households' preferences. Instead, the national government or a donor uses a participatory, decentralized planning approach. All the important decisions now rest with the local community or with households. There are two main advantages of a case C planning protocol. First, the state does not invest in projects people do not want. Households and local communities decide on their own priorities. Sometimes this is referred to as a "demand filter". This means that investments must pass through a filter, or test, to demonstrate that households really want the projects. One type of demand filter might be a monetary payment by households or a copayment by a local community. Or, the demand filter might be a community labor contribution to the construction of a project. If the community didn't want the WASH project, it would not make its contribution, and then it wouldn't receive the project. The second advantage of a case C planning protocol is that it promises to overcome Hayek's dispersed knowledge problem. This is because the state does not need to collect information on household demand. Instead, households will reveal this information as part of the planning process. Let me give you one concrete example of a demand filter. Donors have given away millions of impregnated mosquito nets to households and countries in Africa to prevent malaria. But some households may not want to sleep under mosquito nets, perhaps because they prevent air circulation and leave people feeling too hot. In this case, there would be low demand. But since the nets are given away for free, people might as well take them, right? In some cases, these free nets end up being used for fishing. If households had been involved in the donors' decision to provide free mosquito nets, they might have suggested that the funds involved be used for other priorities instead. Sadly, it turns out that mosquito net fishing is wreaking havoc on many fisheries in Africa. This is an article in The New York Times, published on January 24, 2015, that describes the problems that arise. The mesh of the mosquito nets is very small in order to keep out mosquitos. But when mosquito nets are used for fishing, they capture very small fish and disrupt the breeding cycle of fish, so much that they can destroy the fishery. The case C planning approach has been tried by donors. Case C type projects are often called social funds. The World Bank has financed many of them. The basic idea is that funds are made available to communities to deploy on projects that the communities themselves want most. Another case C example is how the United States federal government gives block grants to states or cities. These block grants can be spent on local priorities. Here in Manchester, England, the city recently received a large grant from the UK central government. Manchester can then decide its own priorities for municipal investments. Of course, it is always possible that local communities will not select projects that donors or higher level government would recommend. This figure shows the infrastructure investments selected by 100 community groups in Nepal. Health clinics are not on the list. If a donor were advocating for public health projects, probably, it would disappointed with the outcome of this process. There are many strong opponents of case C and the use of demand filters. They cite several problems. First, before the intervention, households may not understand the benefits of WASH interventions. So they may make poor, short-sighted decisions. Second, piped water and sewer infrastructure is expensive, and higher level government must be involved in the financing of such capital investments. Third, some planning functions cannot be done efficiently at the local level. For example, there are economies of scale in contracting for the drilling of large numbers of boreholes. There are also information asymmetries between contractors and small communities. For example, contractors may know much more about the likelihood that a drilled well will be successful and hit ground water in a region than a community would know. Higher level government is also likely to be better able to negotiate with large contractors for fair prices than local communities could. Fourth, when regulation is needed, higher level government cannot easily delegate its responsibilities. Local communities cannot be expected to have the expertise needed to regulate private sector water suppliers, for example. Behavioral economists and others have been especially critical of the idea of a demand filter. For example, Professor Michael Kremer at Harvard University has argued, categorically, many economists believe those who most need a project are more likely to pay for it. They are wrong. Professor Kremer goes on to say, "Studies of demand for non-acute [health] care as a function of price show nothing to suggest that the act of paying for something makes a person more likely to use it. Nor is it the case that those who most need a product are more likely to pay for it: those who purchase mosquito nets are no more likely to be sick at the time of purchase; families with small children, who are most likely to die from diarrhea, are no more likely to buy chlorine". Now let's move to discuss case D. In case D, the state does collect information on households' demand. But it uses this preference information for a different purpose than in case B. In case D, donors and national government use a participatory, decentralized planning approach. In case D, important decisions on site selection, technology, pricing, infrastructure arrangements are made by both the state and local communities and households. Case D is sometimes called the demand-driven approach. The main advantage of a case D planning approach is that the state and local communities are each responsible for what they do best. The state does not over-reach and try to make all the important decisions. Instead, the state uses the information it collects on household preferences to design the legal, regulatory, and financial framework for WASH investments. Opponents of case C are likely to have the same objections to case D. They still don't like demand filters. They still believe household preferences are not an appropriate basis for WASH investment decisions. And they think that poor households are likely to suffer the most from a demand-driven planning approach because they will be excluded from the program. I have now described to you a simple topology of four stylized planning approaches that you could use in WASH sector. In the 1950s, 60s, and 70s, cases A and B were the most widely used planning approaches in the rural water sector. In the 1980s many WASH sector professionals began questioning whether cases A and B were working. They began experimenting with cases C and D, planning the protocols. In the 1990s and the early 21st century, different variations of the case D demand-driven planning approach were tried. However, in practice, the demand filters deployed were very modest. Donors and higher level government continued to pay almost all of the capital cost of WASH infrastructure. Today, behavioral economists and others are arguing for a return to the case A and case B protocols. Are they right? What do you think? If you were responsible for the design and implementation of a rural water supply program, which of the four planning protocols would you use?