In this video, we will examine some evidence on how well the Chilean and Colombian subsidy schemes actually worked in practice. The analyses come from the 2003 World Bank economic review paper that I mentioned in the last video. Let's start with Chile. In 1998, the program gave subsidies to about 450,000 households. That was about 13% of the households in the country at the time. The average household subsidy was about $10 per month. Which was about 8% of income for households in the poorest income decile. 8% of income is a substantial increase in a poor person's income. So this result looks good. Also, the total cost of the subsidy was much less than providing a universal subsidy. So this result also looks good. This table shows that the number in value of water subsidies varied widely by region. But this is to be expected because both population and income vary by region. The percentage of households in each region that received the subsidies is shown in the middle column. This also varied by region. The right most column shows the size of the subsidy by region. In regions one and two, the average size of the subsidy was quite large, over $15 per month. However, if we consider the distribution of the subsidies by income decile, instead of by region, the problems with the implementation of the Chilean subsidy scheme become more apparent. The poorest 10% of households only received about 18% of the total subsidies expenditure. More than 60% of the subsidies went to households that were above the third decile of the income distribution. So targeting errors were still quite large. For the poorest 10%, the subsidy had a big impact on their household income, but for most beneficiaries, the impact was actually quite modest, only about 2% of income or less. Now let's look at Colombia. Let's start by looking at a table that shows how well actual housing classifications correlated with household income. Look at the first row. This shows the housing groups where the households in the poorest 10% of the income distribution live. As you can see, only about 19% of the poorest households live in the lowest household classification, low-low. Very few of the poorest households live in the top three housing classification, so that's good. But 31% actually live in the medium-low classification. Now let's look at the last row for the richest households in the top 10% of the income distribution. You might expect that most of them would live in the best housing category, high. But this is not the case. These well-to-do households are actually concentrated in the medium-low, medium, and medium-high housing classifications. So it seems that the Colombian housing classification is not a very accurate indicator of a household's income. Moreover, the implementation of the housing classification protocol resulted in the vast majority of households that have water connections falling into housing classifications, low and medium-low. Nearly 83% of households were classified into the three poorest groups and were eligible to receive a subsidy. The result was that the customer base left to finance cross subsidies was too small. Let's next compare the Chilean and Columbian subsidy schemes. To do this we will look at relative concentration curves. Let me explain how these works. Relative concentration curves show the percentage of total monetary transfers, or subsidies, that accrue to households that are at, or below, a certain range of the income distribution. The horizontal axis measures centiles of income distribution from poorest to richest. The vertical axis measures the accumulated percentage of total transfer. If you've taken a course in economics you will recognize the relative concentration curve is similar to a Lorenz Curve. Before I show you the relative concentration curves for the Chilean and Colombian water subsidies schemes, I suggest you pause this video for a few minutes and think about what shape of curve you would like to see if the schemes we're working well. In other words, if most of the subsidies were being received by the poorest households, what would the curves look like? These are the relative concentration curves for Chile and Colombia, the higher and more concave the curve, the better the targeting property of the subsidy. Unfortunately, this is not what we see. Up the the 30th percentile of the income distribution, the distribution of water subsidies is almost identical for the two schemes. For both schemes, the poorest 30% of the households receive only about 40% of the total subsidies distributed. Above the 30th percentile of the income distribution, the Chilean system performs better, transferring more money to middle income households than to upper income households. Next, I will show you another relative concentration curve. Here the vertical axis is not the accumulated percentage of total subsidies. Instead, it is the accumulated percentage of total beneficiaries, the number of people receiving the subsidy. Again, what we'd like to see are the relative concentration curves that are higher and more concave. Which would indicate better targeting of the subsidy to poor households. In this figure, that shows the accumulated percentage of beneficiaries on the vertical axis, the relative concentration curve for the Chilean scheme, is in fact, higher and more concave than the curve for the Columbian scheme. The curve for the Columbian scheme is virtually a straight line. With the Colombian scheme, the poorest 50% of households include only about 45% of the total beneficiaries. With the Chilean scheme, the poorest 50% of households include about 70% of the total beneficiaries. Not great, but clearly the Chilean scheme here is better than the Colombian scheme. In summary, the Columbian scheme provide subsidies to almost all poor households with private connections. But it also gives some benefits to almost all non-poor households in the country. So it has large areas of inclusion. Quite similar the fact of those that result from IBTs. The Chilean scheme is much smaller and targets a narrower group of households. But it misses many deserving households, and delivers subsidies to many non-poor households. Errors of both exclusion and inclusion. The extent of the errors of inclusion in both countries is striking. The legislation establishing the Chilean subsidy scheme defines beneficiary households as those in the first two income deciles. Yet close to 70% of the scheme's beneficiaries were not in this group. The Columbian approach, based on geographic targeting, has an additional problem. It is unlikely to be dynamically efficient. This is because some of the benefits of the subsidy scheme are probably capitalized into housing prices. If owners move after the establishment of the subsidy, a part of the benefit of the subsidies would go to the owner of the dwelling at the time the scheme was introduced rather than to the current occupant. Finally, I want to emphasize, again, that these results are now somewhat dated. Both subsidy schemes may have been improved. However, I've been unable to locate better information in English. Given that the evidence for both the Chilean means tested scheme and the Columbian geographic targeting scheme are not very impressive, what can we do to assist poor households? I'll address this question in the next video, the last of our four videos on the design of water subsidies.
