In the last video, we saw that increasing block tariffs are the most popular tariff structure in low income countries. The promise of this IBT structures is that it provides a way to ensure that poor households are able to obtain the water they needed a low price and that middle, and upper income households will pay more for water, and cross-subsidize the water use of poor households. In this video, we're going to look at four problems with this popular tariff structure. I will explain in more detail why it is not working as promised in developing countries. In my earlier video on ancient instincts, I showed you this figure which illustrated that increasing blocked tariffs are not targeting subsidies effectively to low income households. The IBT is thus not meeting the social objective of a water tariff. This is because most of the subsidies provided through the water utility are not being received by households in the lowest income quintile. If you'd like to see the details of the precise methodology underpinning these calculations, it is presented in our 2015 paper in the journal Utility Policy. What I want you to know is explain the four main reasons why the IBT is not helping poor households in the way it was intended to. I will also argue that this is not meeting any of the other primary objectives of tariff design, then I will explain why the IBT remains so popular despite the fact that it's not working. I'll conclude with four proposals for moving beyond the IBT structure to improve water pricing and tariff design in developing countries. The first reason that the IBT structure is not targeting subsidies effectively to poor households is that many of the poorest households in developing countries do not have private metered connections. At best then, the IBT structure is only a partial tool for ensuring that poor households have accessible potable water supplies. In fact, many poor households without a private metered connection must rely on water vendors or must purchase water from their neighbors. If vendors or neighbors selling water obtain their water supply from a private metered connection, then an IBT may actually raise prices for unconnected poor households. This is because the IBT raises prices to their providers who are using more water to sell it on. And thus, falling into the higher price blocks of the IBT. Vendors and neighbors selling water must then pass on these costs to the poor households buying from them. The second reason is that an increasing block tariff structure can only work as planned, if rich and poor households both have private metered water connections. But in many cities and low income countries, poor households obtain water from shared connections. If several households share a metered water connection and an IBT is in effect, water use by the group of households sharing a connection can quickly exceed the volume in the initial so-called lifeline block. The volume of water used by the group will quickly push the price to be paid into the highest price block. To the extent that households sharing connections are more likely to be poor than households with private connections for the exclusive use of their family members, the IBT will have precisely the opposite effect from its intent. The poor will pay higher average prices than the rich. I saw this problem for the first time in Kumasi, Ghana almost 26 years ago and published a paper about it in 1992. Next, I will show you some of the data from this first paper. In 1989, most households in Kumasi lived in one room units in multifamily dwellings. Some housing compounds were a multistory like the ones in this picture. 30 to 40 households might live in a building like this. Other multifamily buildings were single story like the ones in this picture, perhaps 8 to 12 households might live in such single story housing compounds. Many single story compounds were built around an open central courtyard. Each household would have a single room that opened into the common courtyard. There was no room for private water connection inside a household's one room unit. If such compound had a private metered water connection, it would be in the open courtyard. All the households in the compound would then obtain their water from this common tap. Some compounds with a private meter connection will actually sell water to neighbors living in other compounds without a private connection in their courtyard. Households with a shared connection receive one water bill from a water utility. They used a variety of different methods to divide the bill among themselves. Some compounds just divided the water bill equally among the households and the compound. Others use complicated systems that tried to approximate the amount of water each household had actually used. In 1989 at the time of our study, this was the increasing block tariff that was in effect in Kumasi. This IBT had three blocks. A couple things to note about this IBT structure. First, the size of the first block is quite large. It is almost 14 cubic meters. My wife and I are not poor, and have lots of water using appliances, and we use about ten cubic meters of water per month. Second, the volume metric charge and the highest price block is still very cheap. Only $0.35 US per cubic meter. Because multifamily buildings had shared meters, the more families there were in a building, the higher the average price of water for households in that building. This was because in buildings with more households, more water was used in total leading to more water sold at the volume metric rate in the highest price block. If rich households lived in buildings with many households, then they would pay the highest average cost of water. In this case, the IBT would result in poor households paying the lowest average price for water and rich households paying the highest average price for water. Do you think this was true? I know it wasn't. We didn't collect household income data in Kumasi. Instead, we counted the number of assets households owned and estimated weekly expenditures. As it turned out, based on these income proxies, richer households lived in buildings with the fewest number of other households and they paid the lowest average price for water. The poorest households actually lived in buildings in the middle range of 9 to 16 households per compound. The differences between the average price of water paid by rich and poor households was not large in Kumasi, because everyone was able to buy water cheaply, but the effect of the IBT was going in the wrong direction. Richer households paid lower average prices than poor households, because of this problem of shared connections. The third reason that IBT structures do not perform well is that they're based on the assumption that poor household use less water than rich households. Sometimes, this is true. But often, it is not. There are rich households that use little water and their are poor household that use a lot of water. Why do you think this is the case? One reason is that household water uses strongly affected by the number of people in the household. So, a rich household with two members might use a lot less water than a poor household with six members. It is also important to understand the effective targeting of subsidies is not about the average water use of rich and poor households. On average, rich households may use some more water, but it is the variance of water use across rich and poor households that drives the effectiveness of the targeting of the subsidies. When I talked to you earlier about ancient instincts, I showed you this table. It presents the correlation between household water use and income for four data sets from locations in developing countries. The correlations are positive, but quite low ranging from 0.08 to 0.23. You might be thinking that the correlation between household income and water use would be higher in high income countries, but this is not the case. This next table shows the correlation between household income and water use in data sets from eight OECD countries. It's shown that correlations are even lower than in the four data sets from developing countries shown in the previous table. The correlations range from -0.18 in Korea to plus 0.24 in Norway and three of the eight countries their correlation is not statistically significant. This data taken from both developing and industrialized country show that the basic rationale for IBT is simply incorrect. The correlation between household water use and household income is much lower than people think. The consequences that if you give subsidies to households with low water used. A lot of these households will be middle and higher income ones, and this is not what is intended at all. The fourth reason that IBT structures do not perform well is that they implicitly assume that rich households will cross subsidize poor households, because the volume metric charge for water in the upper blocks will be above the average cost. But in developing countries, this is not true. The volume metric charge for water in all the blocks is below, usually far below the average cost of water service. This means that all households are being subsidized. It also means that the more water a household uses, the more subsidies it receives. So even if there is a high correlation between household and common water use, the subsidies will still not be well targeted to poor households. These four problems interacts. So in any given city, determining the overall performance of IBT structure is complicated. The resulting distribution of subsidies as often hard for water utilities to estimate empirically, but it is clear that IBT is not effectively targeting subsidies to poor households in developing countries. It is also clear that IBT's are counter productive in terms of the financial cost recovery objective. This is because too much water is being sold to the volume metric rate in the lifeline block far below the cost of supply. Moreover, IBTs cannot possibly send the correct economic signal to households to use water wisely. Because by definition, IBT send multiple signals. Finally, IBTs are among the most complex and less transparent of all different types of tariff structures. After all I've said, you may be thinking to yourself, so just why our IBTs so popular? I think they are popular for three reasons. First, the perception of fairness in water pricing is so important that it turns reality. Second, the IBT subsidizes middle and upper income households. In fact, they receive more subsidies than poor connected households, so they are not going to complain. Some politicians may consider this attribute than IBT advantage. It appears to help the poor, but actually it helps the politicians middle income constituency even more. Third, I think water utility professionals and regulators actually enjoy designing complex water tariff structures. In our personal lives, most of us are price takers in the market. In other words, market forces present us with a take it or leave it price. We never had the opportunity to design complex price structures. Market forces don't allow us to do so. When water professionals get the chance to actually set prices, lots of ideas are proposed and debated and they enjoy such control over prices. The challenge, however is that water pricing and tariff design is becoming more important in an era of increasing urbanization, water scarcity and climate change. It is increasingly harmful to be using a water tariff structure that send the wrong economic signal to users about the economical value of water services. Does not protect the poor and leaves us with bankrupt water utilities throughout the Global South. We have to do better. Here, I do not have time to go into the detailed methodological approaches needed to estimate tariffs structures and the volume metric prices required for immiscible water utility. Well, let me share with you four thoughts on what is needed to improve water pricing and tariff design. First, in almost all cities in developing countries, more households need to be receiving the correct economic signal about the scarcity value of water and the marginal cost of supplying an additional unit to a household. This suggests the need for simple pair of structures like the uniform volume metric rate. Second, IBTs had been used in some cases, because water utilities have argued that they're simply unable to administer a main tested subsidy to poor households. With the big data revolution, this is becoming an increasingly untenable argument. In industrialized countries, almost any private company that wants to can know whether a household is rich or poor and they can find out by combining electronic data records from different sources and it doesn't cost that much to do so. The big data revelation is coming to developing country soon. So means testing should become more feasible for water utilities in these country, as well. Third, currently very few water utilities in developing countries use seasonal pricing. In places where there is a pronounce seasonal peak in system why water use, say during a hot dry season. It makes a lot of sense to use seasonal pricing to dampen peak demand. This will slow down the need for expensive capacity expansion. Fourth, in those cities where marginal costs are above the average cost of water and wastewater services, sending the correct economic signal through the volume metric rate will result in revenues above the amount water utilities need to cover their costs. In such cases, the use of a negative fixed charge should be explored. Negative fixed charges can be used to reduce households's waters bills while at the same time, maintaining the direct economic signal to households.