The Increasing Selectivity of Foreign Aid, 1984-2002 By David Dollar and Victoria Levin1 Abstract We examine the allocation of foreign aid by 41 donor agencies, bilateral and multilateral. Our policy selectivity index measures the extent to which a donor's assistance is targeted to countries with sound institutions and policies, controlling for per capita income and population. The poverty selectivity index analogously looks at how well a donor's assistance is targeted to poor countries, controlling for institutional and policy environment as measured by a World Bank index. Our main finding is that the same group of multilateral and bilateral aid agencies that are very policy focused are also very poverty focused. The donors that appear high up in both rankings are the World Bank's International Development Association (IDA) facility, the International Monetary Fund's Enhanced Structural Adjustment Facility (ESAF), Denmark, the United Kingdom, Norway, Ireland, and the Netherlands. As a robustness check we alternatively use institutional quality measures independent of the World Bank and find the same pattern of selectivity. We also find that policy selectivity is a new phenomenon: in the 1984-89 period, aid overall was allocated indiscriminately without any consideration to the quality of governance, whereas 10 years later there was a clear relationship between aid disbursements and institutions. This increasing selectivity of aid is good news for aid effectiveness. The bad news is that the aid agencies that we survey vary greatly in size. Some donors that are largest in absolute size, such as France and the United States, are not particularly selective. Japan comes in high on the policy selectivity index but far down on the poverty selectivity index, reflecting its pattern of giving large amounts of aid in Asia to countries that are well governed but in many cases not poor. World Bank Policy Research Working Paper 3299, May 2004 The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Policy Research Working Papers are available online at http://econ.worldbank.org. 1Development Economics, World Bank. The World Bank study Assessing Aid (1998) argues that the allocation of foreign aid would have greater impact on poverty reduction if it were targeted to the poorest countries and among them favored the ones with stronger economic institutions and policies. This argument was based on several types of evidence. First, there are specific country cases in which foreign assistance combined with sound institutions and policies to produce good results. The Marshall Plan is the most famous example. More recently, South Korea in the 1960s and 1970s, China in the 1980s, and Uganda and Vietnam in the 1990s are good examples. A second type of evidence comes from project-level studies. All kinds of development projects tend to succeed in sound institutional/policy environments, whereas projects in all sectors tend to fail in countries with weak institutions and policies [Isham and Kaufmann (1999)]. If money were not fungible, this project-level evidence would settle the debate. However, money often is fungible; it is possible that all of the successful donor-financed projects in Uganda would have been undertaken by the private sector or by the government itself if aid resources had not been available. Hence, it is useful to look for a third type of evidence at the macroeconomic level, examining whether there is a relationship between the overall volume of aid that a country receives and its achievements in terms of growth and social development. Burnside and Dollar (2000) find that there is no direct relationship between aid volumes and growth; that is, it is difficult to argue that aid works the same in all institutional/policy environments. However, they also find that growth is correlated with the interaction of aid and an index of institutions and policies. That correlation is consistent with the argument that the effectiveness of aid resources depends critically on the recipient country's own institutions and policies. This result makes a lot of sense provided that donors cannot easily manipulate the institutions and policies of developing countries. Subsequently, a number of papers have criticized this work.2 Common to many of these criticisms is a change in specification, either in terms of estimation technique, or in terms of which variables are included in the regression that explains growth. Subject to these changes, the interaction term between aid and a measure of institutions and policies sometimes loses significance. In one case, Easterly, Levine and Roodman (ELR, 2003) expand the original data set and change the specification. However, simply expanding the data set in terms of the number of countries and the time periods covered does not alter the relationship that Burnside and Dollar found. To show this, we take the extended ELR data set and regress growth, policy, and aid on the other variables (reflecting initial conditions) on the right-hand side of the growth regression, and extract the unexplained (residual) component of each variable. We then sort the data into 9 groups using the 33.3 and 66.7 percentiles of aid and policy. The average growth rate for each group appears to depend on the interaction between aid and policy (Figure 1). For 2Among the critics are Hansen and Tarp (2000, 2001), McPherson (2000), Dalgaard and Hansen (2001), Dayton-Johnson and Hoddinott (2001, 2003), Guillaumont and Chauvet (2001), Lensink and White (2001), Lu and Ram (2001), Akhand and Gupta (2002), Hudson and Mosley (2001), Roodman (2003b), and Easterly, Levine and Roodman (2003). Collier and Dollar (2002), on the other hand, find a similar result working with a different measure of institutions and policies. 2 us, this figure shows that institutions/policies matter (something that is not in dispute). But the figure also shows that the effect of aid on growth also appears to depend on the level of policy. When policy is bad, the level of aid seems to have little impact on growth; if anything, it has a slightly negative relationship. But for countries with good policies, giving sufficiently more aid seems to have a very positive impact on growth. The key statistical question is whether the impact of aid on growth is different for the poor-policy observations than for the good-policy observations. Figure 1. Growth, Aid, and Policy Growth, aid and policy 1.5 1 0.5 Growth 0 -0.5 -1 Low Medium High Aid Bad Policy MediumPolicy Good Policy To shed additional light on this question, Burnside and Dollar (2004) build a new data set focusing only on the 1990s, covering more than 100 countries, and using an overall measure of institutions/policies that is popular in the empirical growth literature. They show that in the 1990s there is simply no evidence that aid promotes growth in all institutional environments. On the other hand, they again find this correlation between growth and the interaction of aid and institutions. In fact, the interaction is more robust than institutions by themselves. If one is choosing between the hypotheses that aid works the same in all institutional environments or that the effect of aid is conditional on the institutions that directly affect growth, the evidence clearly favors the latter. 3 These various types of evidence provide the background for our analysis of the changing selectivity of aid. We develop two very intuitive selectivity indexes. The policy selectivity index measures the extent to which a donor's assistance is targeted to countries with sound institutions and policies, controlling for per capita income and population. The poverty selectivity index analogously looks at how well a donor's assistance is targeted to poor countries, controlling for institutional and policy environment as well as population. There is some similarity between our approach and the aid selectivity index developed by the Center for Global Development (CGD) [Roodman (2003a)]. However, the CGD index combines the issues of poverty and policy selectivity into a single index. Donors may legitimately have different preferences about the extent to which they want to target policy versus poverty. So we prefer to compile separate indexes of policy and poverty selectivity, each of which has a very transparent interpretation. The policy selectivity index indicates the extent to which each donor prefers better governed countries among those at a given level of per capita income. The poverty index indicates the extent to which each donor favors poorer countries among those at the same quality of governance. If one wants an overall index, we would recommend simply averaging the two, as there is no compelling evidence to suggest that one should be weighted any more than the other. In the next section we present our indexes for the past few years. The measure of institutions and policies that we use is the World Bank's Country Policy and Institutional Assessment (CPIA). As noted, in principle one might find donors with different tastes, some more policy focused and some more poverty focused. There are a couple of examples of this, notably Japan, which is fairly high up in the policy selectivity ranking but near the bottom of the poverty selectivity ranking. However, one of the interesting results that we obtain is that in general the same group of multilateral and bilateral aid agencies that are very policy focused are also very poverty focused. The donors that appear high up in both rankings are the World Bank's International Development Assistance (IDA) facility, the International Monetary Fund's Enhanced Structural Adjustment Facility (ESAF), Denmark, the United Kingdom, Norway, Ireland, and the Netherlands. In section 2 we repeat the analysis but focus only on IDA-eligible countries. We get very similar rankings when the analysis is constrained to low-income countries. In the years we consider, IDA always comes out near the very top of the policy selectivity ranking. Since the measure of institutions/policies that we use is a subjective one that comes from the World Bank, a natural question to ask in Section 3 is how robust this result is to the use of alternative indicators. There has recently been a proliferation of measures of institutions and policies, but because in Section 4 we want to look at the changing selectivity of aid in the past two decades, we choose for the sensitivity analysis two institutional measures that are available over time, independent of the World Bank, and widely used in empirical studies: the International Country Risk Guide (ICRG) rule of law index and the Freedom House democracy index. We argue that the two together are a good proxy for the institutions/policies that lead to growth and poverty reduction. We redo the analysis with these independent institutional measures, and obtain a ranking 4 of donor selectivity that is highly correlated with what we got with the CPIA. Using any of these measures, it is fair to say that the World Bank's IDA and the bilateral donors noted above are selective both in the sense of targeting assistance to the poorest countries, and in the sense of favoring the better governed ones among the poor countries. In Section 4 we use the ICRG and Freedom House measures to look back in time. It is possible that selectivity is a natural and obvious approach to aid allocation that has always been followed. But in fact we do not find this to be the case. In the 1984-89 period there was virtually no policy selectivity. Aid overall was allocated indiscriminately with regard to the institutional measures; the World Bank's IDA facility had virtually no relationship with the measures. Ten years later the situation had changed dramatically; IDA disbursements had a sharp positive relationship with sound institutions. Bilateral donors such as Denmark and the United Kingdom strengthened their link to sound institutions as well. We conclude from this work that a simple intuitive approach to measuring aid selectivity provides a robust methodology for making comparisons of aid allocations across donors and over time. Revealed preference suggests that most donors have bought into the aid selectivity model: today about three-fourths of the 40 or so aid agencies that we survey have a positive relationship between their allocations and a measure of sound institutions and policies, after controlling for per capita GDP and population. In general, the most policy focused agencies are also the most poverty focused ones, indicating that it is possible to target aid both to the countries where poverty is great and at the same time to countries with reasonably good governance. The historical analysis reveals that in the 1984-89 period aid in general was not allocated in favor of good governance, so that we can have some confidence that the allocation of aid has improved over time. That is the good news. The bad news is that the different aid agencies that we survey vary greatly in size. A number of the donors that are largest in absolute size are not particularly selective. Both France and the United States show up far down in both the policy and poverty selectivity indexes, indicating that much of the assistance of these big countries goes to recipients that are not particularly poor and not particularly well governed. Japan comes in high on the policy selectivity index but far down on the poverty selectivity index, reflecting its pattern of giving large amounts of aid in Asia to countries that are well governed but in many cases not poor. The European Commission was far down in the policy selectivity index in 2000 and 2001, but then made a big jump to 12th place in 2002. On poverty selectivity, the European Commission is in the bottom half of donors. Since France, the European Commission, the United States, and Japan account for the majority of foreign aid, how these large donors manage their funds has a large impact on the overall efficacy of aid. In summary, the typical aid agency has improved its aid quality in the sense of targeting more funds to poor countries with sound governance. However, there is still room for further improvement because most aid money comes from four large donors that are not particularly selective in either the policy or poverty dimension. 5 1. Overall Selectivity of Aid There is broad agreement among economists and development specialists that countries' institutions and policies are important determinants of long-term growth and poverty reduction.3 By institutions, we mean rules, norms, and behaviors, or what Hall and Jones (1999) identify as "social infrastructure." Property rights and the legal rules and behaviors that enforce them are examples of growth-enhancing institutions. A well- functioning civil service based on merit and transparency is another example. On the policy side, policies such as macroeconomic stability and a relatively open trade regime help create an environment conducive to investment and growth. At the same time there is also considerable evidence, discussed in the introduction, that these same institutions and policies are key determinants of the effectiveness of financial assistance. A government that provides a sound framework for growth and poverty reduction is also one that is more likely to use financial resources well to complement policies with necessary public investments in roads, schools, and the like. These ideas have given rise to the approach of policy selectivity, that is, channeling more aid resources to countries that have the institutional and policy framework to use the resources effectively. At the same time there is also broad agreement that concessional assistance should be targeted to the poorest countries [Russett (1978), Roodman (2003a)]. Middle- income countries with sound institutions and policies generally have good access to international capital markets. There is also less concentration of poverty in these countries. In practice quite a bit of aid is allocated to middle-income countries, so in addition to looking at policy selectivity, we think that it is also useful to look at poverty selectivity. We propose to measure the policy and poverty selectivity of donor allocations in a simple and transparent way. The Development Cooperation Directorate (DAC) of the Organization for Economic Cooperation and Development (OECD) compiles data on the allocation across developing countries of aid from more than 40 sources, bilateral and multilateral. For the purpose of our paper, we use gross ODA disbursements net of emergency aid, since this latter type of aid should not necessarily adhere to selectivity concerns of donors (Roodman [2003a]). For each source we estimate an equation across developing country recipients of the form: Log (aidij) = b0 + b1 Log (populationj) + b2 Log (per capita GDPj) + b3 Log (index of institutions/policiesj) where i indexes donors and j indexes developing country aid recipients. The data on population and per capita GDP at PPP are taken from the Penn World Tables. To observe the extent of donor selectivity, we substitute very small values (.01 million or $10,000) for recipients not receiving any aid from a donor, so that these recipients do not drop out of our regressions. 3Some of the influential recent papers on this topic are Dollar and Kraay (2002), Rodrik, Subramanian, and Trebbi (2002), Acemoglu, Johnson, and Robinson (2001), Engerman, and Sokoloff (1997), Hall and Jones (1999), and Knack and Keefer (1995). 6 Some donor agencies explicitly follow allocation rules along the lines of our estimated specification; in other cases there are complex factors behind aid allocation. We do not intend this equation as a complete model of donor behavior; we realize that political and strategic considerations do play a role in aid allocation [Alesina and Dollar (2000), Boschini and Olofsgård (2002)]. But in practice almost all aid agencies' allocations fit this kind of equation rather nicely with the following common characteristics: b1 is usually positive but less than 1 (that is, more populous countries get more aid, but less than proportionally to their population); b2 is negative (poor countries get more aid); and b3 is positive (holding constant population and per capita income, donors generally give more aid to countries with better institutions/policies). We argue that looking at the b2's across donors tells us about the degree of poverty selectivity, whereas looking at the b3's across donors tells us about their policy selectivity. We initially implement this approach by looking across all developing countries for each of the four years 1999, 2000, 2001, and 20024, using as the measure of institutions and policies the World Bank's CPIA. This measure has the advantage that it is a comprehensive assessment of institutions and policies available for all countries on a timely basis. It has the disadvantage that it is to some extent a subjective World Bank rating and that it is not as yet publicly available to other researchers. In section 3 below we check for the robustness of our rankings by repeating this approach with different institutional measures independent of the World Bank. Table 1 presents the indexes of policy selectivity for the four years. It starts with an estimation for total aid. In 2002, after controlling for population and per capita GDP, total aid has a strong positive relationship to the CPIA: a 100% increase in CPIA corresponds to 176% more aid. The relationship is much stronger for multilateral assistance (elasticity of 2.57) than for bilateral assistance (0.63). Bilateral assistance represents about two-thirds of all foreign aid. In 2002, Denmark was the most selective donor, followed closely by the World Bank's IDA facility and the United Kingdom. For these donors the index number is around 4, indicating that there is sharp differentiation among recipients based on the CPIA, after holding constant population and per capita GDP. In other words, for countries with the same population and per capita GDP, but differing in CPIA, these agencies provide about 400% more financing to countries with 100% better CPIA. The reader should recognize that this is a very large difference in policy; many of the poor-performing developing countries have CPIAs around 2.5, and medium ones around 3.0. That difference is about 20% and would correspond to about 80% more financial assistance from the highly selective donors. (Twenty percent difference is about the standard deviation of the CPIA across developing countries.) Note that 29 out of 42 donor agencies have a positive coefficient on CPIA; whatever basis they have for making allocations, in practice they give more money to countries with better institutions/policies. The coefficients that are significantly different from zero at least at the 10% level are highlighted in bold. Keep in mind that small 4Since PWT data is only available up to 2000, we proxy for the 2001 and 2002 population and per capita GDP with 2000 data. 7 differences in the rankings are usually not statistically significant. What is robust, however, is that there is a group of donors that are at the top of this ranking, and the relationship of their assistance to good institutions/policies is significantly different from that of donors in the middle or at the bottom. The more selective bilateral donors include the Nordic countries, the Netherlands, the United Kingdom, and Austria. The European Commission is at #12. Some of the largest donors in absolute terms appear quite low on this scale, such as the United States (#27) and France (#30). Note also that the year-to- year change tends to be small. In ranking terms, over the four years IDA is #1, #3, #2, #2. Denmark, #4, #4, #1, #1; ESAF, #5, #1, #3, #10; France, #31, #29, #34, #30; the United States, #28, #28, #32, #27. In practice, donor allocations do not change much from year to year and recipient countries' institutions and policies do not change radically from year to year, so that the indexes are moderately stable over time. Table 2 presents the indexes of poverty elasticity. A donor could do well on the policy selectivity index simply by giving most of its assistance to the best-managed among the middle-income countries. Hence the poverty selectivity index brings in another important element of aid quality: the extent to which assistance is targeted to the poor countries. As with policy, multilateral assistance is more poverty selective than bilateral assistance. The highest rated agencies here are all multilateral ones: in 2002 IDA was most poverty focused, followed by the African Development Fund and the IMF's ESAF. The top eight in terms of poverty selectivity are all multilateral agencies. By construction, these multilateral funds are channeled to the poorest countries, so it should not really be a surprise to see them appear at the top of the ranking. Some of the bilateral donors that appeared high up in the index of policy selectivity also appear high up in the ranking based on poverty selectivity: the Netherlands (#10), Denmark (#12), Norway (#13), the United Kingdom (#16), and Sweden (#18). The large volume donors are all low down on this poverty selectivity ranking: the United States at #20, European Commission at #25, France at #29, and Japan at #35. It is important to emphasize in this work is that it is possible to do well in both poverty and policy selectivity by focusing assistance on low-income countries with relatively good institutions and policies. This appears to be the allocation strategy of the Nordic countries, the Netherlands, the United Kingdom, IDA, and ESAF. We think that the separate indexes of policy and poverty selectivity are useful, highlighting the different dimensions of aid selectivity. Nevertheless, there is always demand for a single index of aid quality. In principle there are ideal weights for poverty and policy elasticities that would lead aid to be allocated so as to have the maximum effect on poverty reduction. However, in practice we doubt that they can be estimated with great precision, and therefore we opt for the transparent approach of weighting poverty and policy equally in creating an overall index. Table 3 simply averages the policy elasticity and the (negative of) the poverty elasticity to create an overall aid selectivity index and ranking. In 2002 IDA is at the top of this ranking, followed by Denmark, ESAF, the United Kingdom, Sweden, Nordic Development Fund, the Netherlands, Norway, and the African Development Fund. At the bottom of the list one finds Portugal, France, Greece, and Spain, as well as a number of United Nations agencies and regional banks. A number of 8 the large volume donors appear in the middle to bottom of the ranking: European Commission (#16), Japan (#22), and the United States (#24). 2. Selectivity among IDA-Eligible Recipients The next step in our study is to subject these findings to different types of sensitivity analysis. Do the results depend on using the World Bank's CPIA rating rather than some other assessment of institutions and policies? This is a natural question, and in the next section we repeat this analysis using different measures of institutions and policies. Another issue is the simple linear form of the regression equation. The linear form is easy to interpret, but in practice the relationship between aid and per capita GDP is highly nonlinear, and one naturally wonders if adjusting for that nonlinearity substantially changes the assessment of aid quality. The nonlinearity, which exists for many individual donors, can be easily seen in Figure 2. This is the partial scatter of log IDA disbursements plotted against log per capita GDP (measured at PPP). The "partial" scatter means that both log IDA and log per capita GDP have been adjusted for population and CPIA. IDA disbursements basically follow a step function. Countries above a certain per capita GDP are not eligible at all (get zero), whereas for countries that are eligible there is a modest negative relationship between disbursements and per capita GDP. The linear elasticity accurately captures the fact that IDA is targeted to the poorest developing countries, but clearly one would get a better fit by recognizing that the relationship is nonlinear. Adjusting for this may or may not affect the analysis of policy selectivity. So we repeat the analysis above, but now limit the coverage to IDA-eligible countries, thereby capturing virtually all low-income countries and a number of lower middle-income recipients that still qualify for IDA. Table 4 presents the policy elasticities estimated only across this IDA-eligible world. The rankings turn out to be broadly similar to those in Table 1, estimated across all developing countries.5 In 2002, IDA was the most policy selective donor, followed by Nordic Development Fund, Global Environment Facility (GEF), and Denmark. While the rankings are broadly similar, there are a number of interesting differences. Table 5 shows the poverty elasticities estimated across the IDA-eligible aid recipients. Again, the ranking is not that different from what we found across all 5We also considered the possibility that there were differences in the allocation of grants versus concessional loans or that the results may be influenced in any one year by debt relief, which is counted as part of aid but would follow different allocation rules. In appendix table 1 we show policy selectivity indexes calculated from grants only and from aid minus debt relief, and it can be seen that these alternative approaches make almost no difference in the estimated elasticities for different donors. In appendix table 2 we calculate policy elasticities with two additional control variables: a dummy for post-conflict situation and a measure of susceptibility to economic shocks (the residual from a growth regression for the 1990s). Adding these variables addresses the issue of whether or not estimated donor policy selectivity is influenced by donor pursuit of other objectives, such as supporting post-conflict countries or targeting shock-prone environments. It can be seen in appendix table 2 that the estimated policy elasticities are unaffected by including these other considerations. 9 developing countries. The most poverty focused agencies in this ranking are the African Development Fund, Belgium, Ireland, World Food Programme, Nordic Development Fund, ESAF, and Sweden; IDA is at #12. The least poverty focused are Asian Development Bank's special funds, Japan, and the European Bank for Reconstruction and Development (EBRD). Note also that the absolute values of the elasticities are generally much lower than in table 2. Much of the poverty focus of aid overall comes from its targeting to low-income countries. Among low-income countries, all donors display much less differentiation based on per capita income. In fact, 13 of the agencies have a positive relationship between their disbursements and per capita GDP, among this IDA-eligible group. And the relationship is just about zero for aid overall and is positive for aggregate bilateral aid. This may seem odd at first glance, but keep several things in mind. Imagine two countries with the same population and per capita GDPs of $300 and $600, respectively. If a donor gives the poorer country a volume of aid that amounts to 2 percent of the recipient's GDP, the same absolute volume of aid will be 1 percent of GDP for the richer country. It is difficult for even very poverty-focused donors to channel too much of their assistance to the poorest countries because the aid relative to GDP then becomes very large and creates problems of absorptive capacity, aid dominance of the economy, and aid dependency of the government [Clemens and Radelet (2003)]. As these low-income countries develop, in general their needs for infrastructure and other public investment tend to increase, so that there is often a greater role for aid. As before, we can combine the ranking of policy selectivity and the ranking of poverty selectivity to derive an overall index of aid selectivity for IDA-eligible recipients (Table 6). The African Development Fund is at the top of this ranking in 2002 with an average elasticity of 5.39, followed by IDA at 5.32, and the Nordic Development Fund at 5.27. Up until now we have not said anything about aid quantity, but this can be introduced into the analysis as well. It is possible for a donor to be quite selective, but with a modest amount of assistance. In table 7 we adjust for that by calculating how much each donor country citizen is providing in official assistance to the IDA-eligible countries. We then divide these countries into a "strong institutions" group and a "weak institutions" group (where the latter is the bottom half of IDA-eligible countries on the CPIA ranking), and calculate how much aid each donor country citizen provided to either group. In 2002 Denmark proved the most generous to IDA-eligible countries with sound institutions, providing $79.42 per Danish citizen. Denmark is followed in the ranking by Luxembourg ($77.68), Netherlands ($53.48), Norway (45.65), Sweden ($25.92), and Ireland ($24.44). At the low end of the ranking is Portugal ($1.16 per citizen), Greece ($1.64), New Zealand ($2.08), Spain ($3.57), Italy ($3.84), Australia ($5.26), and the United States ($5.66). Among the large G-7 countries there is significant variation from Japan's $22.54 and the United Kingdom's $17.52 per citizen to the United States' $5.66 and Canada's $6.34. Aid quantities are also often measured relative to the donor country's gross national income (GNI), and these figures are shown in Table 8, again focusing on aid to IDA-eligible countries and distinguishing between the strong institutions group and the 10 weak institutions group. The rankings are similar to those based on per capita aid giving: the top providers of assistance to IDA-eligible countries with sound institutions are Denmark, Netherlands, Luxembourg, Norway, Sweden, and Ireland. 3. Selectivity Measured with Independent Indicators It should not be surprising that the World Bank allocates a lot of assistance to the countries that it ranks highly in its annual CPIA rating exercise. A natural question to turn to next is how robust this analysis is to the choice of institutional/policy measure. Recently there has been a proliferation of indexes that attempt to measure the quality of state institutions and policies. However, most of these are not available in a comparable format back in time. So, for our sensitivity analysis here we want to choose indicators that we can also use in the next section, where we turn to the issue of the changing selectivity of aid over the past two decades. For this purpose we are going to use two measures of institutional quality that are available for a large number of countries since the early 1980s: the ICRG rule of law index and the Freedom House democracy index. The rule of law index is close to the economic institutions captured in the CPIA, but narrower. The democracy measure, on the other hand, captures the extent of political rights and civil liberties, but in practice the two institutional measures together are good predictors of certain economic policies, such as macroeconomic stability and trade openness. We calculate policy elasticity indexes using first rule of law, then democracy, and finally a combination of the two. Together with the CPIA-based measure, this gives four alternatives to compare. Table 9 presents the four alternative policy selectivity indexes for the period 1995-99, during which we have both CPIA and the independent measures for large numbers of countries. If the ICRG rule of law measure is used as the indicator of institutions/policies, the four most selective donors are Denmark, Finland, International Fund for Agricultural Development (IFAD), and IDA. At the bottom of the ranking are many of the same agencies at the bottom of the ranking based on CPIA (e.g., France at #38, Spain at #40, and Portugal at #41). The correlation between the CPIA ranking and the rule of law ranking is quite high at .57. However, the magnitudes of the elasticities are not directly comparable. Log rule of law has a standard deviation of 0.34, whereas log CPIA has a standard deviation of 0.23. So, in the case of Denmark, for example, a standard deviation better on the CPIA corresponds to 112% more aid and a standard deviation better on rule of law also corresponds to 76% more aid, estimates that are not too far apart. Thus, using the independent rule of law measure yields results quite similar to using the World Bank's CPIA. Using democracy as the institutional measure produces a very similar correlation (.58). One interesting feature of this ranking is that the multilateral institutions generally drop down (IDA to #31, ESAF to #21, UNDP to #34, etc.). In general, the multilateral agencies are not supposed to take political institutions or political factors into account in allocating resources, so it turns out that the Freedom House democracy measure by itself is a poor proxy for what they are targeting. On the other hand, many of the bilateral 11 agencies have strong democracy selectivity: the Nordic countries, Netherlands, United Kingdom, Japan, Canada, are high up on the index based on the democracy measure. Finally, we create a selectivity index based on both rule of law and democracy. If we want to use both of these institutional measures in the analysis of selectivity, we have to decide how to aggregate them into a single index ­ that is, what weights to use for each measure. We use the following approach. For the period of 1995-1999, regress log CPIA on log of democracy, log of rule of law and four year dummies to get the following relationship: Log CPIA = constant - .165 log (Freedom House) + .364 log (ICRG rule of law) + year dummies Adj. R2 = .36 (5.62) (10.64) F-test = 37.05 Keep in mind that on the Freedom House index a higher number is more authoritarian. The CPIA is significantly correlated with both institutional measures (t- statistics in parentheses). If one thinks that the CPIA is the right concept (the institutional and policy package that leads to growth and poverty reduction) but that in practice it may be measured with error or politically influenced, then this equation provides a predicted CPIA that is free from such problems. We then use this predicted value to calculate a policy sensitivity index, in essence replacing the actual CPIA with a predicted CPIA based on institutional measures independent of any donor agency. Column 4 in Table 9 presents this ranking. It is correlated .65 with the ranking based on CPIA. In general the same agencies appear high up (low down) on this ranking, as in column 1. A final important point is that, with each of these institutional measures, a different poverty elasticity is calculated for each donor. It turns out that these poverty selectivity rankings are extremely highly correlated (above .90). Thus, whatever measure of institutions is used, exactly the same agencies appear as more or less poverty focused. As we did in the first section, we can combine the policy and poverty rankings into a single aid selectivity index. Table 10 repeats the original ranking of aid quality from table 3 and adds in an analogous ranking based on column 4 of Table 9 (using both rule of law and democracy together to measure institutional/policy selectivity). The rankings of aid quality are remarkably similar (and correlated .77). Both rankings share the following characteristics: at the top we find multilateral agencies such as IDA, ESAF, IFAD and African Development Fund and certain bilateral donors (the Nordic countries, the United Kingdom, Netherlands, Ireland); a number of large capital donors are in the middle of the pack (Japan, the United States, European Commission); at the bottom are donors that focus on their former colonies (France, Portugal, Spain). 4. Changing Policy Selectivity over Time In this section we go back in time and examine whether the patterns of selectivity that we observe in the 1999-2002 period represent a long-term trend or new thinking in development assistance. For this analysis we divide the data into three time periods: 1984-89, 1990-94, and 1995-99. As in the previous section we regress the aid for each recipient on its per capita GDP in PPP, its population, and our aggregate of the two 12 measures of institutional quality. We also now introduce year dummies to control for year-to-year fluctuations in aid disbursements. The CPIA is not available in a consistent format back in time, so we conduct this historical analysis using the combined rule of law and democracy measure, presented in the last section. Table 11 shows that policy selectivity has increased to a very great extent between the second half of the 1980s and the second half of the 1990s.6 For aid overall, bilateral aid, and multilateral aid, there was no significant relationship with institutional quality in the 1984-89 period. By 1995-99, each type had a significant positive relationship with institutional quality. The relationship was far stronger for multilateral aid than for bilateral assistance ­ an elasticity of 1.82 compared to 0.50. In the case of IDA, there was similarly no significant relationship with institutional quality in the early period, and then a strong one by 1995-99 (an elasticity of 3.58, putting IDA in sixth place among all the agencies). The median elasticity was 0.24 (not significantly different from zero) in 1984-89, and 1.5 in 1995-99. The historical analysis thus reveals that policy selectivity is a new phenomenon, one that is now practiced by most donor agencies. Table 12 shows the poverty elasticities calculated back in time. In general, the same agencies that we find poverty focused today, have been more poverty focused all along. Throughout this period multilateral assistance has been more poverty focused than bilateral assistance. 5. Conclusions In the past two decades foreign aid overall has become more selective in the following sense: in the second half of the 1980s, aid was allocated indiscriminately to well governed and poorly governed countries alike, at the same level of per capita income. Today there is a clear tendency to allocate more assistance to poor countries that have reasonably good economic governance. We find this relationship between aid and governance when we use the World Bank's CPIA measure; the same relationship is found if alternatively an independent assessment of institutions is used. In general, multilateral assistance is more sharply targeted both to good policies and to poverty. In 2002 the World Bank's IDA facility was #2 in policy selectivity and #1 in poverty selectivity, putting it at the top of the overall aid selectivity index that is a simple average of the two selectivity measures. The overall trend toward greater selectivity of aid can be observed for IDA as well. In the 1984-89 period there was little relationship between IDA disbursements and measures of governance; by 2002 the relationship had strengthened to the point that a standard deviation better on the CPIA measure translates into nearly 100% more assistance. 6An observant reader would notice a slight difference in the elasticities presented in column 4 of Table 9 and those for 1995-1999 period in Table 11. This difference arises because the sample in Table 9 was constrained to be the same across columns, providing a fairer comparison between different institutional measures. In Table 11 we allow all the available observations for 1995-1999 to be included in our regressions. 13 There are a number of bilateral donors that appear high up in the rankings for policy and poverty selectivity, such as Denmark, Norway, Sweden, the United Kingdom, Ireland, and the Netherlands. However, some of the large donors in absolute size, such as France and the United States, are not particularly selective in either the policy or poverty dimension. Japan is selective on policy but not on poverty. As a result, overall bilateral aid is not very selective. 14 References Acemoglu, Daron, Simon Johnson and James A. Robinson (2001). "The Colonial Origins of Comparative Development: An Empirical Investigation." American Economic Review 91:5, pp. 1369-1401. Akhand, Hafiz A. and Kanhaya L. Gupta (2002). Foreign Aid in the Twenty-First Century. Boston: Kluwer Academic Publishers. Alesina, Alberto and David Dollar (2000). "Who Gives Foreign Aid to Whom and Why?" Journal of Economic Growth 5:1 (March), pp. 33-63. Boschini, Anne and Anders Olofsgård (2002). "Foreign Aid: An Instrument for Fighting Poverty or Communism?" Working Paper. Burnside, Craig and David Dollar (2000). "Aid, Policies, and Growth." American Economic Review 90:4 (September), pp. 847­68. Burnside, Craig and David Dollar (2004). "Aid, Policies, and Growth: Revisiting the Evidence." World Bank, Policy Research Working Paper No. 2834. Clemens, Michael and Steven Radelet (2003). "The Millenium Challenge Account: How Much Is Too Much, How Long Is Long Enough?" CGD Working Paper 23, February. Collier, Paul, and David Dollar (2002). "Aid Allocation and Poverty Reduction," European Economic Review. Dalgaard, Carl-Johan and Henrik Hansen (2001). "On Aid, Growth and Good Policies." Journal of Development Studies 37:6 (August), pp. 17­41. Dayton-Johnson, Jeff and John Hoddinott (2001). "Optimal Allocation of Development Aid Across Countries." Unpublished manuscript, Dalhousie University, June. Dayton-Johnson, Jeff and John Hoddinott (2003). "Aid, Policies, and Growth, Redux." Unpublished manuscript, Dalhousie University, April. Dollar, David, and Aart Kraay (2002), "Institutions, Trade, and Growth," Journal of Monetary Economics. Easterly, William, Ross Levine and David Roodman (2003). "New Data, New Doubts: A Comment on Burnside and Dollar's `Aid, Policies, and Growth'." Forthcoming, American Economic Review. Engerman, Stanley and Kenneth Sokoloff (1997). "Factor Endowments, Institutions, and Differential Paths of Growth Among New World Economics: A View from 15 Economic Historians of the United States," in Stephen Haber, ed. "How Latin America Fell Behind: Essays on the Economic Histories of Brazil and Mexico 1800-1914." Stanford, CA: Stanford University Press. Guillaumont, Patrick and Lisa Chauvet (2001). "Aid and Performance: A Reassessment." Journal of Development Studies 37:6 (August), pp. 66­92. Hall, Robert E. and Charles I. Jones (1999). "Why Do Some Countries Produce So Much More Output per Worker than Others?" Quarterly Journal of Economics 114:1 (February), pp. 83-116. Hansen, Henrik and Finn Tarp (2000). "Aid Effectiveness Disputed." Journal of International Development 12:3 (April), pp. 375­98. Hansen, Henrik and Finn Tarp (2001). "Aid and Growth Regressions." Journal of Development Economics 64:2 (April), pp 547­70. Hudson, John and Paul Mosley (2001). "Aid Policies and Growth: In Search of the Holy Grail." Journal of International Development 13:7 (October). Isham, Jonathan and Daniel Kaufmann (1999). "The Forgotten Rationale for Policy Reform: The Productivity of Investment Projects." Quarterly Journal of Economics 114:1 (February), pp. 149-184. Knack, Steven and Philip Keefer (1995). "Institutions and Economic Performance: Cross-Country Tests Using Alternative Measures." Economics and Politics 7:3, pp. 207­27. Lensink, Robert and Howard White (2001). "Are There Negative Returns to Aid?" Journal of Development Studies 37:6 (August), pp. 42­65. Lu, Shuang and Rati Ram (2001). "Foreign Aid, Government Policies, and Economic Growth: Further Evidence from Cross-Country Panel Data for 1970­93." Economia Internazionale 54:1 (February), pp. 14­29. McPherson, Malcolm (2000). "Aid, Policies, and Growth: A Comment." Unpublished manuscript, Harvard University. Rodrik, Dani, Arvind Subramanian and Francesco Trebbi (2002). "Institutions Rule: The Primacy of Institutions over Geography and Integration in Economic Development." NBER Working Paper No. 9305. Roodman, David (2003a). "An Index of Donor Aid Performance." Center for Global Development, April. 16 Roodman, David (2003b). "The Anarchy of Numbers: Aid, Development, and Cross- Country Empirics." CGD Working Paper 32, September. Russett, Bruce (1978). "The Marginal Utility of Income Transfers to the Third World." International Organization 32:4, pp. 913-928. World Bank (1998). Assessing Aid: What Works, What Doesn't, and Why. Washington, DC: The World Bank. November. 17 Table 1. Policy Elasticity Index, 1999-2002 1999 2000 2001 2002 Donor Mean Rank Mean Rank Mean Rank Mean Rank Total Aid 1.240 1.308 1.356 1.759 Bilateral Aid 0.750 0.814 0.413 0.628 Multilateral Aid 1.989 1.732 2.132 2.568 AfDF 2.876 11 2.401 15 1.887 16 1.022 22 Arab Agencies 4.192 6 2.231 17 0.760 27 0.971 25 AsDB Spec. Funds -1.172 40 -1.325 39 -0.583 35 -0.364 33 Australia 0.378 32 1.210 24 1.364 21 1.008 23 Austria 2.014 20 2.803 11 2.311 12 3.259 5 Belgium 2.295 18 3.149 9 1.652 18 1.920 15 Canada 1.988 21 2.435 14 1.442 19 0.990 24 CarDB 1.297 25 1.076 26 1.149 24 1.404 18 Council of Europe no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Denmark 4.484 4 4.690 4 4.789 1 4.767 1 EBRD -2.254 42 -0.815 36 0.655 28 1.394 20 EC 2.515 17 1.702 21 2.204 14 2.440 12 Finland 2.528 16 2.220 18 2.267 13 2.536 11 France 0.476 31 0.206 29 -0.494 34 -0.072 30 GEF -1.417 41 -0.399 33 2.016 15 3.027 7 Germany 1.679 22 2.964 10 2.525 11 2.064 14 Greece -1.058 39 -1.264 38 0.188 31 -0.559 36 IDA 5.846 1 4.918 3 4.138 2 4.233 2 IDB Spec. Oper. Fund 0.756 29 -0.237 31 0.201 30 -0.240 32 IFAD 3.915 7 2.214 19 1.400 20 1.403 19 IMF Trust Fund no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Ireland 3.262 8 2.455 13 2.575 10 1.383 21 Italy 1.400 24 2.333 16 1.198 23 -0.120 31 Japan 4.565 2 4.208 5 2.833 8 1.901 16 Luxembourg 2.235 19 1.872 20 1.234 22 2.146 13 Netherlands 4.529 3 5.109 2 3.472 4 2.647 9 New Zealand 1.489 23 1.632 22 0.812 26 0.409 29 Nordic Development Fund 2.631 14 3.337 7 2.582 9 3.085 6 Norway 2.809 12 0.799 27 2.905 7 2.701 8 Other UN 3.157 9 0.006 30 -0.762 38 -0.504 34 Portugal 0.301 33 1.129 25 0.462 29 0.540 28 SAF & ESAF 4.449 5 5.467 1 3.924 3 2.569 10 Spain -0.290 36 -1.871 42 -3.040 42 -1.253 39 Sweden 2.651 13 3.177 8 3.321 6 3.315 4 Switzerland 1.063 27 1.596 23 1.696 17 1.673 17 UNDP 1.204 26 -0.369 32 -0.606 36 -0.531 35 UNFPA -0.663 37 -1.599 41 -2.153 41 -2.082 41 UNHCR -0.910 38 -1.560 40 -0.726 37 -1.480 40 UNICEF 0.485 30 -0.980 37 -1.726 40 -2.308 42 United Kingdom 2.924 10 4.028 6 3.407 5 3.657 3 United States 0.892 28 0.562 28 0.024 32 0.664 27 UNRWA 0.130 34 -0.811 35 -0.954 39 -0.970 38 UNTA -0.163 35 -0.603 34 -0.394 33 -0.966 37 WFP 2.587 15 2.499 12 0.834 25 0.720 26 Note: Bolded elasticities are different from zero at 10% significance level (with heteroskedasticity-consistent s.e.). 18 Table 2. Poverty Elasticity Index, 1999-2002 1999 2000 2001 2002 Donor Mean Rank Mean Rank Mean Rank Mean Rank Total Aid -0.377 -0.417 -0.468 -0.490 Bilateral Aid -0.319 -0.375 -0.334 -0.381 Multilateral Aid -0.738 -0.744 -0.854 -0.833 AfDF -2.737 3 -2.461 3 -2.598 3 -2.689 2 Arab Agencies -1.942 6 -2.019 6 -1.967 6 -1.875 6 AsDB Spec. Funds -0.163 34 -0.042 35 -0.093 34 -0.133 33 Australia -0.359 26 -0.445 26 -0.343 27 -0.270 30 Austria -0.303 28 -0.337 28 -0.240 29 -0.305 28 Belgium -1.095 14 -1.165 11 -1.267 10 -1.293 9 Canada -0.604 21 -0.489 23 -0.558 21 -0.650 23 CarDB 0.076 38 0.065 36 0.042 36 0.136 37 Council of Europe no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Denmark -1.286 11 -1.164 12 -1.303 8 -1.110 12 EBRD 0.865 42 0.389 40 0.613 41 0.543 41 EC -0.254 30 -0.219 32 -0.526 23 -0.511 25 Finland -0.549 22 -0.405 27 -0.610 20 -0.756 21 France -0.389 25 -0.332 29 -0.256 28 -0.279 29 GEF 0.730 41 0.640 42 0.018 35 -0.071 34 Germany -0.352 27 -0.447 25 -0.489 24 -0.474 26 Greece 0.552 40 0.388 39 0.175 40 0.206 40 IDA -4.255 1 -4.164 1 -4.208 1 -4.200 1 IDB Spec. Oper. Fund 0.469 39 0.530 41 0.614 42 0.625 42 IFAD -2.165 5 -2.202 5 -2.340 4 -1.968 5 IMF Trust Fund no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Ireland -1.341 9 -1.367 9 -1.514 7 -1.213 11 Italy -0.608 20 -0.659 21 -0.436 25 -1.067 15 Japan -0.506 23 -0.203 34 -0.212 30 0.012 35 Luxembourg -0.432 24 -0.477 24 -0.206 33 -0.664 22 Netherlands -1.655 7 -1.654 7 -1.208 11 -1.271 10 New Zealand -0.253 31 -0.328 30 -0.207 32 -0.163 31 Nordic Development Fund -0.794 17 -1.087 13 -0.983 15 -1.093 14 Norway -1.304 10 -0.979 15 -1.174 13 -1.106 13 Other UN -0.015 35 -0.699 20 -0.698 19 -0.624 24 Portugal -0.248 32 -0.213 33 0.100 38 0.147 38 SAF & ESAF -2.750 2 -2.794 2 -2.787 2 -2.495 3 Spain 0.072 37 0.370 38 0.096 37 0.072 36 Sweden -0.810 16 -0.875 17 -0.914 17 -1.023 18 Switzerland -0.673 18 -0.712 19 -0.764 18 -0.833 19 UNDP -1.280 12 -1.288 10 -1.173 14 -1.311 8 UNFPA -1.107 13 -1.028 14 -0.980 16 -1.041 17 UNHCR -0.206 33 -0.493 22 -0.390 26 -0.158 32 UNICEF -1.458 8 -1.403 8 -1.301 9 -1.315 7 United Kingdom -0.618 19 -0.838 18 -1.184 12 -1.064 16 United States -0.864 15 -0.960 16 -0.558 22 -0.761 20 UNRWA 0.048 36 0.130 37 0.162 39 0.153 39 UNTA -0.303 29 -0.324 31 -0.211 31 -0.312 27 WFP -2.382 4 -2.260 4 -2.169 5 -2.232 4 Note: Bolded elasticities are different from zero at 10% significance level (with heteroskedasticity-consistent s.e.). 19 Table 3. Overall Aid Selectivity Index, 1999-2002 1999 2000 2001 2002 Donor Mean Rank Mean Rank Mean Rank Mean Rank Total Aid 0.809 0.862 0.912 1.124 Bilateral Aid 0.534 0.594 0.374 0.504 Multilateral Aid 1.363 1.238 1.493 1.701 AfDF 2.807 7 2.431 6 2.242 6 1.855 9 Arab Agencies 3.067 4 2.125 12 1.364 18 1.423 17 AsDB Spec. Funds -0.504 39 -0.642 40 -0.245 39 -0.115 34 Australia 0.369 31 0.828 25 0.854 23 0.639 25 Austria 1.159 22 1.570 16 1.275 19 1.782 10 Belgium 1.695 15 2.157 11 1.460 15 1.607 13 Canada 1.296 20 1.462 18 1.000 21 0.820 23 CarDB 0.611 29 0.506 28 0.553 26 0.634 26 Council of Europe no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Denmark 2.885 6 2.927 4 3.046 3 2.938 2 EBRD -1.559 42 -0.602 39 0.021 31 0.426 28 EC 1.384 18 0.961 23 1.365 17 1.476 16 Finland 1.539 17 1.312 19 1.439 16 1.646 12 France 0.433 30 0.269 31 -0.119 35 0.103 32 GEF -1.073 41 -0.519 37 0.999 22 1.549 14 Germany 1.016 23 1.706 15 1.507 13 1.269 20 Greece -0.805 40 -0.826 41 0.006 32 -0.382 36 IDA 5.050 1 4.541 1 4.173 1 4.217 1 IDB Spec. Oper. Fund 0.144 34 -0.383 35 -0.207 37 -0.432 37 IFAD 3.040 5 2.208 9 1.870 10 1.685 11 IMF Trust Fund no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Ireland 2.302 10 1.911 14 2.045 8 1.298 19 Italy 1.004 24 1.496 17 0.817 24 0.474 27 Japan 2.535 8 2.206 10 1.523 12 0.944 22 Luxembourg 1.333 19 1.174 20 0.720 25 1.405 18 Netherlands 3.092 3 3.382 3 2.340 4 1.959 7 New Zealand 0.871 27 0.980 22 0.509 27 0.286 30 Nordic Development Fund 1.713 14 2.212 8 1.783 11 2.089 6 Norway 2.057 11 0.889 24 2.039 9 1.903 8 Other UN 1.586 16 0.352 30 -0.032 33 0.060 33 Portugal 0.274 32 0.671 27 0.181 30 0.197 31 SAF & ESAF 3.599 2 4.131 2 3.356 2 2.532 3 Spain -0.181 37 -1.121 42 -1.568 42 -0.662 42 Sweden 1.730 13 2.026 13 2.117 7 2.169 5 Switzerland 0.868 28 1.154 21 1.230 20 1.253 21 UNDP 1.242 21 0.459 29 0.284 29 0.390 29 UNFPA 0.222 33 -0.285 34 -0.587 41 -0.520 39 UNHCR -0.352 38 -0.533 38 -0.168 36 -0.661 41 UNICEF 0.971 25 0.211 32 -0.213 38 -0.496 38 United Kingdom 1.771 12 2.433 5 2.296 5 2.360 4 United States 0.878 26 0.761 26 0.291 28 0.712 24 UNRWA 0.041 36 -0.470 36 -0.558 40 -0.561 40 UNTA 0.070 35 -0.139 33 -0.092 34 -0.327 35 WFP 2.485 9 2.380 7 1.501 14 1.476 15 20 Figure 2. Partial Correlation Plot for Log IDA aid and log per capita GDP, 1999 coef = -4.2545615, (robust) se = .21549824, t = -19.74 8.08969 HND NIC CPV ALB TZA UGAZMB GHA MWIYEM BGD ARM BOL MKD GEO ETH MDGMRT RWABFA MOZTCD IND MLI BEN SEN KENCIV CHN NER LSO KGZGRD GIN TGO NPLTJKGUYLKAAZE CMRPAK GMB ZWE BDI NGA KHM EGY LCA X)| GNB PHL VCT IDN lnida KNA e( GNQ BLZ SYC BWAEST TTO MUS NAM SVN GAB COM SLV JAMCRI PANLVAFJI URY LTUSWZ JOR DOMTUN BGRLBN HUN HRV SVK CHL CZE PER GTM MARCOLPNGPOLMYS ZAF KAZPRY ARG KOR COGECU THA DZA TURROMMEX VEN BLR BRA UKR RUS -7.46746 -1.90142 1.75517 e( lnrgdpch | X ) 21 Table 4. Policy Elasticity Index for Aid to IDA-eligible Countries, 1999-2002 1999 2000 2001 2002 Donor Mean Rank Mean Rank Mean Rank Mean Rank Total Aid 1.625 3.071 2.506 2.389 Bilateral Aid 1.225 3.226 2.297 2.230 Multilateral Aid 3.057 3.376 2.939 3.070 AfDF 7.266 6 10.446 2 8.878 3 6.610 6 Arab Agencies 7.394 5 7.214 9 5.892 12 6.336 7 AsDB Spec. Funds -4.046 40 -4.168 41 -1.247 38 0.675 32 Australia -1.182 39 2.597 26 3.905 18 2.960 22 Austria 1.578 27 2.836 25 1.091 28 3.483 20 Belgium 1.325 29 4.490 18 0.545 31 -0.086 37 Canada 3.470 18 3.908 20 3.586 19 2.200 25 CarDB 2.330 24 2.139 27 2.812 20 2.817 23 Council of Europe no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Denmark 8.055 2 9.562 5 9.733 1 7.808 4 EBRD -1.040 37 -1.591 39 -2.146 40 -2.706 41 EC 5.032 10 3.481 21 2.069 26 2.050 29 Finland 3.465 19 4.484 19 5.391 13 4.150 19 France 1.068 30 1.840 29 0.287 34 0.477 33 GEF no aid to IDA countr. n/a 1.829 31 6.453 8 8.603 3 Germany 1.328 28 3.009 24 2.532 23 1.575 30 Greece -0.359 35 -0.125 37 1.829 27 -0.748 38 IDA 8.979 1 6.795 13 6.211 10 9.758 1 IDB Spec. Oper. Fund 1.892 25 2.108 28 0.592 30 2.067 28 IFAD 5.330 9 6.398 14 4.653 15 7.533 5 IMF Trust Fund no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Ireland 4.893 11 7.063 11 7.444 7 5.146 16 Italy 1.637 26 6.824 12 2.375 25 4.718 18 Japan 5.925 7 8.797 6 6.288 9 5.655 12 Luxembourg 3.167 21 3.415 22 2.628 22 5.968 10 Netherlands 7.593 3 9.931 3 4.445 17 5.447 13 New Zealand -0.276 33 1.829 30 2.658 21 2.218 24 Nordic Development Fund 4.253 15 8.764 7 7.767 5 9.287 2 Norway 4.297 14 4.713 17 7.477 6 5.246 15 Other UN 3.807 16 0.561 33 -2.468 41 2.154 27 Portugal 0.256 31 0.560 34 -0.311 36 0.418 35 SAF & ESAF 7.540 4 11.729 1 9.155 2 4.832 17 Spain 3.588 17 5.251 15 1.028 29 6.178 8 Sweden 4.708 12 7.094 10 5.909 11 5.796 11 Switzerland 3.191 20 3.320 23 2.474 24 3.373 21 UNDP 2.591 23 0.067 35 -2.069 39 -1.095 40 UNFPA 0.001 32 1.363 32 0.278 35 0.924 31 UNHCR -1.060 38 -2.060 40 0.356 33 0.462 34 UNICEF -0.715 36 -0.623 38 -1.033 37 -0.823 39 United Kingdom 5.363 8 9.701 4 8.739 4 6.169 9 United States 4.585 13 4.756 16 4.775 14 2.193 26 UNRWA no aid to IDA countr. n/a no aid to IDA countr. n/a no aid to IDA countr. n/a no aid to IDA countr. n/a UNTA -0.285 34 -0.054 36 0.477 32 0.061 36 WFP 2.724 22 7.571 8 4.446 16 5.444 14 Note: Bolded elasticities are different from zero at 10% significance level (with heteroskedasticity-consistent s.e.). 22 Table 5. Poverty Elasticity Index for Aid to IDA-eligible Countries, 1999-2002 1999 2000 2001 2002 Donor Mean Rank Mean Rank Mean Rank Mean Rank Total Aid 0.046 -0.172 -0.120 -0.042 Bilateral 0.107 -0.048 -0.010 0.083 Multilateral -0.342 -0.420 -0.374 -0.267 AfDF -4.156 1 -3.653 1 -4.050 1 -4.177 1 Arab Agencies -1.655 3 -1.600 6 -1.003 9 -0.968 8 AsDB Spec. Funds 2.068 40 2.043 41 1.879 41 2.188 41 Australia 1.042 39 0.563 37 0.597 37 0.705 37 Austria 0.092 31 0.222 33 0.308 34 -0.461 22 Belgium -1.583 5 -1.643 3 -1.562 4 -1.439 2 Canada -0.413 23 -0.015 30 0.012 29 0.098 29 CarDB 0.656 37 0.705 38 0.677 38 0.772 38 Council of Europe no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Denmark -1.712 2 -1.608 5 -1.151 6 -0.533 20 EBRD 0.653 36 0.389 35 0.890 40 0.914 39 EC -0.714 15 -0.332 23 -0.670 14 -0.785 14 Finland -0.240 27 -0.222 27 -0.603 19 -0.481 21 France -0.828 14 -0.585 17 -0.550 21 -0.583 19 GEF no aid to IDA countr. n/a 0.794 39 0.349 36 0.217 32 Germany -0.486 19 -0.373 21 -0.324 23 -0.156 26 Greece 0.584 35 0.536 36 0.195 31 0.546 35 IDA -1.174 8 -1.188 8 -1.060 8 -0.885 12 IDB Spec. Oper. Fund 0.371 33 0.179 32 0.283 33 0.178 30 IFAD 0.088 30 -0.385 20 -0.571 20 -0.334 24 IMF Trust Fund no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Ireland -1.555 6 -1.517 7 -1.956 3 -1.341 3 Italy -0.887 11 -0.888 12 -0.667 15 -0.779 15 Japan 0.680 38 1.154 40 0.868 39 1.307 40 Luxembourg -0.850 12 -0.735 16 -0.719 12 -0.897 10 Netherlands -0.655 16 -0.925 10 -0.613 18 -0.621 18 New Zealand 0.478 34 0.324 34 0.203 32 0.325 34 Nordic Development Fund 0.019 29 -0.946 9 -1.129 7 -1.257 5 Norway -0.644 17 -0.555 18 -0.633 17 -0.632 17 Other UN -0.080 28 -0.219 28 -0.321 24 -0.415 23 Portugal -0.448 22 -0.292 24 -0.106 27 -0.115 27 SAF & ESAF -1.641 4 -3.438 2 -2.014 2 -1.212 6 Spain -0.353 24 -0.394 19 -0.440 22 -0.935 9 Sweden -1.148 9 -0.913 11 -0.760 11 -1.087 7 Switzerland -0.454 20 -0.236 25 0.050 30 0.205 31 UNDP -1.026 10 -0.871 14 -0.686 13 -0.743 16 UNFPA -0.593 18 -0.795 15 -0.641 16 -0.849 13 UNHCR -0.451 21 -0.226 26 -0.139 26 0.272 33 UNICEF -0.835 13 -0.881 13 -0.832 10 -0.897 11 United Kingdom 0.114 32 -0.025 29 -0.170 25 0.632 36 United States -0.280 26 0.109 31 0.324 35 -0.110 28 UNRWA no aid to IDA countr. n/a no aid to IDA countr. n/a no aid to IDA countr. n/a no aid to IDA countr. n/a UNTA -0.315 25 -0.364 22 -0.064 28 -0.328 25 WFP -1.294 7 -1.636 4 -1.298 5 -1.258 4 Note: Bolded elasticities are different from zero at 10% significance level (with heteroskedasticity-consistent s.e.). 23 Table 6. Overall Aid Selectivity Index for Aid to IDA-eligible Countries, 1999-2002 1999 2000 2001 2002 Donor Mean Rank Mean Rank Mean Rank Mean Rank Total Aid 0.790 1.622 1.313 1.215 Bilateral Aid 0.559 1.637 1.153 1.073 Multilateral Aid 1.700 1.898 1.657 1.669 AfDF 5.711 1 7.049 2 6.464 1 5.393 1 Arab Agencies 4.524 5 4.407 8 3.448 9 3.652 7 AsDB Spec. Funds -3.057 40 -3.106 41 -1.563 41 -0.757 40 Australia -1.112 39 1.017 28 1.654 20 1.128 25 Austria 0.743 30 1.307 25 0.391 32 1.972 20 Belgium 1.454 24 3.067 15 1.054 27 0.677 32 Canada 1.942 20 1.962 21 1.787 18 1.051 26 CarDB 0.837 28 0.717 31 1.068 26 1.022 27 Council of Europe no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Denmark 4.883 3 5.585 3 5.442 3 4.171 5 EBRD -0.847 38 -0.990 40 -1.518 40 -1.810 41 EC 2.873 9 1.907 22 1.369 23 1.417 22 Finland 1.853 21 2.353 18 2.997 12 2.315 18 France 0.948 26 1.212 26 0.419 31 0.530 33 GEF no aid to IDA countr. n/a 0.518 32 3.052 11 4.193 4 Germany 0.907 27 1.691 24 1.428 22 0.866 31 Greece -0.472 37 -0.331 38 0.817 28 -0.647 39 IDA 5.076 2 3.991 11 3.636 8 5.322 2 IDB Spec. Oper. Fund 0.761 29 0.965 29 0.154 35 0.945 29 IFAD 2.621 12 3.392 14 2.612 15 3.934 6 IMF Trust Fund no aid reported n/a no aid reported n/a no aid reported n/a no aid reported n/a Ireland 3.224 7 4.290 9 4.700 4 3.244 12 Italy 1.262 25 3.856 12 1.521 21 2.748 17 Japan 2.623 11 3.821 13 2.710 14 2.174 19 Luxembourg 2.008 17 2.075 20 1.674 19 3.432 10 Netherlands 4.124 6 5.428 4 2.529 16 3.034 13 New Zealand -0.377 36 0.753 30 1.227 24 0.946 28 Nordic Development Fund 2.117 15 4.855 6 4.448 6 5.272 3 Norway 2.471 13 2.634 17 4.055 7 2.939 15 Other UN 1.944 19 0.390 35 -1.073 39 1.284 23 Portugal 0.352 31 0.426 34 -0.103 37 0.266 34 SAF & ESAF 4.591 4 7.583 1 5.585 2 3.022 14 Spain 1.971 18 2.822 16 0.734 29 3.556 8 Sweden 2.928 8 4.004 10 3.335 10 3.442 9 Switzerland 1.822 22 1.778 23 1.212 25 1.584 21 UNDP 1.808 23 0.469 33 -0.692 38 -0.176 38 UNFPA 0.297 32 1.079 27 0.459 30 0.886 30 UNHCR -0.304 35 -0.917 39 0.248 34 0.095 36 UNICEF 0.060 33 0.129 37 -0.101 36 0.037 37 United Kingdom 2.625 10 4.863 5 4.455 5 2.769 16 United States 2.432 14 2.324 19 2.225 17 1.151 24 UNRWA no aid to IDA countr. n/a no aid to IDA countr. n/a no aid to IDA countr. n/a no aid to IDA countr. n/a UNTA 0.015 34 0.155 36 0.270 33 0.194 35 WFP 2.009 16 4.603 7 2.872 13 3.351 11 24 Table 7. Aid per Bilateral Donor Country's Capita to IDA-eligible Countries, 1999-2002 1999 2000 2001 2002 Strong Weak Strong Weak Strong Weak Strong Weak Institutions Institutions Institutions Institutions Institutions Institutions Institutions Institutions Donor $/capita Rank $/capita Rank $/capita Rank $/capita Rank $/capita Rank $/capita Rank $/capita Rank $/capita Rank Australia 1.84 21 12.76 8 2.40 20 4.07 13 2.07 19 3.41 14 5.26 17 0.62 20 Austria 8.25 14 23.36 3 6.31 15 12.96 5 12.78 11 12.30 4 7.51 14 5.26 13 Belgium 6.46 15 8.20 11 7.40 14 7.67 9 7.75 15 9.11 9 16.62 10 11.41 7 Canada 6.40 16 4.08 16 5.05 16 3.28 14 5.15 17 2.77 16 6.34 15 7.49 12 Denmark 78.27 1 26.93 2 74.93 1 17.74 3 78.97 1 23.00 3 79.42 1 20.04 3 Finland 12.01 12 7.84 12 8.33 13 2.73 16 8.94 12 2.93 15 7.65 13 4.68 15 France 14.02 10 14.02 6 11.21 11 12.77 6 7.94 14 11.19 7 9.97 12 28.04 1 Germany 10.47 13 7.66 13 8.81 12 5.55 11 8.36 13 5.71 11 13.59 11 7.79 11 Greece 2.22 20 0.15 22 1.47 21 0.16 22 1.44 20 0.08 22 1.64 21 0.33 22 Ireland 18.16 7 2.07 18 23.18 6 3.01 15 27.57 6 2.03 17 24.44 6 19.34 4 Italy 3.23 19 1.86 19 3.94 18 0.98 21 1.37 21 0.54 21 3.84 18 9.76 10 Japan 18.18 6 23.18 4 17.41 9 15.05 4 17.58 8 12.25 5 22.54 7 4.29 16 Luxembourg 36.95 3 38.15 1 62.32 2 40.53 1 33.87 4 35.91 1 77.68 2 16.27 5 Netherlands 24.41 5 14.01 7 31.15 5 17.91 2 47.59 3 24.32 2 53.48 3 15.03 6 New Zealand 1.03 22 4.16 15 1.09 22 1.16 20 1.02 22 1.27 20 2.08 20 0.39 21 Norway 53.54 2 16.62 5 43.67 3 10.94 7 48.27 2 11.93 6 45.65 4 23.42 2 Portugal 15.03 9 1.21 21 19.81 7 1.55 19 5.76 16 1.37 19 1.16 22 3.30 18 Spain 4.00 17 1.47 20 3.85 19 1.93 18 12.92 10 1.80 18 3.57 19 2.45 19 Sweden 24.73 4 9.36 10 33.63 4 6.29 10 29.50 5 6.89 10 25.92 5 10.68 9 Switzerland 17.43 8 11.96 9 15.77 10 10.51 8 15.20 9 10.22 8 21.82 8 10.80 8 United Kingdom 13.28 11 4.41 14 18.02 8 4.56 12 19.91 7 4.27 13 17.52 9 4.94 14 United States 3.57 18 2.25 17 4.53 17 2.40 17 4.34 18 4.71 12 5.66 16 3.67 17 25 Table 8. Aid to IDA-eligible Countries per Donor Country's GNI, percent (1999-2002) 1999 2000 2001 2002 Strong Weak Strong Weak Strong Weak Strong Weak Institutions Institutions Institutions Institutions Institutions Institutions Institutions Institutions Donor %GNI Rank %GNI Rank %GNI Rank %GNI Rank %GNI Rank %GNI Rank %GNI Rank %GNI Rank Australia 0.009 21 0.061 5 0.012 21 0.021 12 0.011 20 0.018 12 0.025 16 0.003 20 Austria 0.032 14 0.092 1 0.028 16 0.056 5 0.056 9 0.054 4 0.030 14 0.021 14 Belgium 0.026 17 0.033 10 0.032 14 0.034 7 0.034 16 0.040 6 0.067 8 0.046 7 Canada 0.031 15 0.020 15 0.022 17 0.015 15 0.023 17 0.013 18 0.028 15 0.033 10 Denmark 0.241 1 0.083 3 0.254 1 0.060 3 0.266 1 0.077 3 0.247 1 0.062 3 Finland 0.049 11 0.032 11 0.036 13 0.012 18 0.039 13 0.013 16 0.031 13 0.019 16 France 0.058 8 0.058 6 0.052 9 0.059 4 0.036 15 0.051 5 0.043 12 0.120 1 Germany 0.041 13 0.030 14 0.039 12 0.025 11 0.037 14 0.026 11 0.057 10 0.033 11 Greece 0.019 18 0.001 22 0.014 19 0.001 22 0.013 18 0.001 22 0.013 21 0.003 22 Ireland 0.083 7 0.009 19 0.108 7 0.014 16 0.121 5 0.009 20 0.091 6 0.072 2 Italy 0.016 19 0.009 20 0.021 18 0.005 21 0.007 22 0.003 21 0.019 18 0.048 6 Japan 0.051 10 0.065 4 0.046 10 0.040 6 0.053 11 0.037 7 0.071 7 0.013 18 Luxembourg 0.085 6 0.087 2 0.162 3 0.105 1 0.089 7 0.095 2 0.189 3 0.040 9 Netherlands 0.097 4 0.055 7 0.133 4 0.076 2 0.198 2 0.101 1 0.204 2 0.057 4 New Zealand 0.008 22 0.030 13 0.008 22 0.009 19 0.008 21 0.010 19 0.014 20 0.003 21 Norway 0.157 2 0.049 8 0.122 6 0.031 8 0.132 3 0.033 8 0.109 4 0.056 5 Portugal 0.133 3 0.011 17 0.191 2 0.015 14 0.054 10 0.013 15 0.010 22 0.028 12 Spain 0.027 16 0.010 18 0.028 15 0.014 17 0.090 6 0.013 17 0.022 17 0.015 17 Sweden 0.091 5 0.035 9 0.131 5 0.025 10 0.127 4 0.030 9 0.101 5 0.042 8 Switzerland 0.045 12 0.031 12 0.043 11 0.029 9 0.040 12 0.027 10 0.053 11 0.026 13 United Kingdom 0.054 9 0.018 16 0.075 8 0.019 13 0.083 8 0.018 13 0.067 9 0.019 15 United States 0.011 20 0.007 21 0.013 20 0.007 20 0.012 19 0.013 14 0.015 19 0.010 19 26 Table 9. Comparison of Policy Elasticities -- CPIA and Other Indicators, 1995-1999 Democracy & CPIA Rule of Law Democracy Rule of Law Donor Mean Rank Mean Rank Mean Rank Mean Rank Total Aid 1.444 0.398 -0.273 0.832 Bilateral Aid 1.262 0.146 -0.223 0.240 Multilateral Aid 1.740 0.852 -0.359 1.861 AfDF 2.133 15 1.955 5 0.081 32 3.256 9 Arab Agencies 3.575 6 1.130 11 -0.683 14 1.954 18 AsDB Spec. Funds 0.888 28 0.008 31 0.204 33 0.321 27 Australia 1.637 21 1.790 6 -0.344 23 4.382 4 Austria 1.586 22 1.283 10 -0.140 28 2.371 11 Belgium 1.304 25 -0.306 36 -0.747 11 -0.274 32 Canada 2.570 11 0.183 27 -0.947 8 0.656 24 CarDB 0.111 36 0.002 32 -0.479 20 0.313 28 Council of Europe n/a n/a n/a n/a n/a n/a n/a n/a Denmark 4.889 2 2.237 1 -2.175 1 6.493 1 EBRD -1.064 41 0.916 15 0.278 35 1.911 19 EC 1.524 23 0.916 14 -0.488 19 2.082 16 Finland 2.451 13 2.033 2 -0.842 9 4.595 3 France 0.383 30 -0.487 38 0.548 42 -2.466 41 GEF -1.220 42 0.826 16 -0.437 22 2.227 14 Germany 1.930 18 0.793 17 -0.268 25 1.612 21 Greece -0.445 40 0.536 21 0.476 40 0.246 30 IDA 5.384 1 1.971 4 -0.007 31 3.601 8 IDB Spec. Oper. Fund 0.185 34 -2.099 42 -1.457 2 -2.199 39 IFAD 3.154 8 2.010 3 -0.321 24 3.689 7 IMF Trust Fund n/a n/a n/a n/a n/a n/a n/a n/a Ireland 1.932 17 1.481 9 -0.722 12 3.718 6 Italy 1.070 27 -0.433 37 0.415 37 -1.876 38 Japan 3.674 4 0.575 20 -1.123 6 2.203 15 Luxembourg 1.222 26 0.632 19 -0.835 10 1.995 17 Netherlands 3.659 5 0.081 29 -1.200 3 1.179 23 New Zealand 1.332 24 0.414 22 -0.565 17 1.772 20 Nordic Development Fund 2.210 14 0.939 13 -0.678 15 2.271 13 Norway 3.219 7 1.483 8 -1.062 7 4.036 5 Other UN 1.729 20 0.129 28 -0.508 18 0.521 25 Portugal -0.118 38 -1.635 41 -0.115 29 -3.444 42 SAF & ESAF 4.347 3 0.985 12 -0.455 21 2.378 10 Spain -0.143 39 -1.119 40 -0.217 27 -2.305 40 Sweden 2.543 12 0.670 18 -1.141 5 2.306 12 Switzerland 1.988 16 0.063 30 -0.684 13 0.515 26 UNDP 0.359 32 0.287 24 0.271 34 0.307 29 UNFPA 0.462 29 -0.252 35 0.300 36 -0.946 36 UNHCR -0.108 37 0.228 26 0.420 38 -0.649 35 UNICEF 0.381 31 -0.703 39 0.527 41 -1.837 37 United Kingdom 2.950 9 1.774 7 -1.168 4 4.950 2 United States 1.892 19 0.289 23 -0.647 16 1.571 22 UNRWA 0.176 35 0.242 25 0.442 39 -0.342 34 UNTA 0.289 33 -0.021 34 -0.045 30 -0.078 31 WFP 2.785 10 -0.007 33 -0.245 26 -0.281 33 Note: Bolded elasticities are different from zero at 10% significance level (with heteroskedasticity-consistent s.e.). 27 Table 10. Comparison of Overall Aid Selectivity Indexes, 1995-1999 CPIA Democracy & Rule of Law Donor Mean Rank Mean Rank Total Aid 0.931 0.600 Bilateral Aid 0.807 0.255 Multilateral Aid 1.233 1.307 AfDF 2.442 7 3.072 3 Arab Agencies 2.612 4 1.807 11 AsDB Spec. Funds 0.544 31 0.199 29 Australia 1.008 21 2.579 8 Austria 0.855 25 1.305 14 Belgium 1.109 20 0.193 30 Canada 1.554 13 0.485 26 CarDB -0.026 37 0.152 31 Council of Europe n/a n/a n/a n/a Denmark 3.050 3 3.929 1 EBRD -0.950 42 0.706 25 EC 0.893 24 1.202 18 Finland 1.390 17 2.631 7 France 0.400 32 -1.187 39 GEF -0.926 41 0.907 22 Germany 1.163 19 0.925 21 Greece -0.375 40 0.051 33 IDA 4.719 1 3.788 2 IDB Spec. Oper. Fund -0.238 39 -1.622 41 IFAD 2.573 5 2.918 5 IMF Trust Fund n/a n/a n/a n/a Ireland 1.484 14 2.524 9 Italy 0.913 23 -0.704 38 Japan 2.031 10 1.258 16 Luxembourg 0.757 29 1.150 19 Netherlands 2.558 6 1.275 15 New Zealand 0.746 30 0.989 20 Nordic Development Fund 1.397 16 1.476 13 Norway 2.198 9 2.682 6 Other UN 0.935 22 0.229 28 Portugal 0.055 36 -1.737 42 SAF & ESAF 3.159 2 2.125 10 Spain -0.123 38 -1.320 40 Sweden 1.690 11 1.552 12 Switzerland 1.334 18 0.443 27 UNDP 0.816 27 0.775 23 UNFPA 0.791 28 -0.005 34 UNHCR 0.105 34 -0.136 35 UNICEF 0.852 26 -0.368 37 United Kingdom 1.656 12 2.967 4 United States 1.438 15 1.257 17 UNRWA 0.068 35 -0.228 36 UNTA 0.298 33 0.075 32 WFP 2.410 8 0.725 24 28 Table 11. Policy Elasticity Index, 1984-1999 1984-1989 1990-1994 1995-1999 Donor Mean Rank Mean Rank Mean Rank Total Aid 0.057 0.297 1.046 Bilateral Aid 0.152 0.128 0.500 Multilateral Aid -0.020 0.947 1.816 AfDF 3.631 1 2.821 2 3.384 8 Arab Agencies -0.585 29 0.614 19 1.968 19 AsDB Spec. Funds -1.027 34 -0.901 35 0.357 27 Australia 1.935 5 1.756 8 4.565 3 Austria -1.187 35 1.762 7 2.597 10 Belgium -0.599 30 0.311 21 -0.268 32 Canada 0.060 24 -0.470 34 0.606 25 CarDB 0.128 22 -0.204 32 0.235 29 Council of Europe -0.009 26 -0.009 30 n/a n/a Denmark 2.389 2 4.450 1 6.302 1 EBRD n/a n/a 2.650 3 1.970 17 EC 0.462 17 1.134 11 2.325 12 Finland 0.578 14 1.048 12 4.444 4 France 0.545 16 0.078 26 -2.360 39 GEF n/a n/a n/a n/a 2.060 15 Germany 0.868 9 1.236 10 1.789 20 Greece n/a n/a n/a n/a 0.318 28 IDA 0.694 13 0.941 14 3.582 6 IDB Spec. Oper. Fund -2.100 38 -1.346 37 -2.908 41 IFAD -0.173 27 0.685 17 3.051 9 IMF Trust Fund 0.028 25 n/a n/a n/a n/a Ireland 0.707 12 1.004 13 3.521 7 Italy 1.187 7 -0.061 31 -1.664 37 Japan 0.407 18 2.182 5 2.029 16 Luxembourg n/a n/a -0.314 33 1.969 18 Netherlands 0.561 15 0.844 15 1.369 23 New Zealand 0.806 10 0.812 16 1.421 22 Nordic Development Fund n/a n/a 2.139 6 2.352 11 Norway 2.194 3 1.687 9 4.016 5 Other UN 1.477 6 0.522 20 0.179 30 Portugal -0.678 32 -2.137 39 -3.217 42 SAF & ESAF 0.187 21 0.170 24 2.153 14 Spain -0.202 28 -2.720 41 -2.417 40 Sweden 0.775 11 0.653 18 2.258 13 Switzerland -0.678 31 0.251 23 0.855 24 UNDP 0.290 19 0.271 22 0.576 26 UNFPA 0.086 23 0.025 28 -1.240 36 UNHCR 0.986 8 0.042 27 -0.489 34 UNICEF -0.772 33 -1.515 38 -2.283 38 United Kingdom 2.175 4 2.489 4 4.964 2 United States -1.596 36 -1.100 36 1.580 21 UNRWA n/a n/a -0.003 29 -0.367 33 UNTA 0.196 20 0.152 25 -0.225 31 WFP -1.767 37 -2.438 40 -0.836 35 Note: Bolded elasticities are different from zero at 10% significance level (with heteroskedasticity-consistent s.e.). 29 Table 12. Poverty Elasticity Index, 1984-1999 1984-1989 1990-1994 1995-1999 Donor Mean Rank Mean Rank Mean Rank Total Aid -0.948 -0.535 -0.507 Bilateral Aid -0.869 -0.396 -0.372 Multilateral Aid -1.432 -1.208 -1.063 AfDF -2.836 2 -2.805 2 -2.678 2 Arab Agencies -1.048 20 -1.378 10 -1.547 9 AsDB Spec. Funds 0.041 34 0.068 38 -0.155 34 Australia -0.878 24 -0.654 26 -0.857 18 Austria 0.047 35 -0.224 32 -0.335 30 Belgium -0.867 26 -0.818 21 -0.828 20 Canada -1.451 14 -0.783 22 -0.725 21 CarDB 0.019 33 0.065 36 -0.014 35 Council of Europe 0.140 37 0.045 35 n/a n/a Denmark -2.351 3 -2.220 3 -1.678 6 EBRD n/a n/a 0.129 39 0.309 40 EC -2.182 4 -1.176 13 -0.868 17 Finland -1.064 19 -0.630 27 -0.857 19 France -1.016 21 -0.727 23 -0.230 32 GEF n/a n/a n/a n/a 0.330 41 Germany -0.877 25 -0.450 30 -0.389 28 Greece n/a n/a n/a n/a 0.132 39 IDA -3.752 1 -3.797 1 -3.928 1 IDB Spec. Oper. Fund 1.028 38 0.589 41 0.670 42 IFAD -1.542 12 -1.320 11 -2.053 3 IMF Trust Fund -0.040 31 n/a n/a n/a n/a Ireland -0.613 29 -0.883 19 -1.268 13 Italy -1.701 8 -0.974 17 -0.718 23 Japan -1.014 22 -0.715 25 -0.509 26 Luxembourg n/a n/a -0.375 31 -0.473 27 Netherlands -1.674 9 -1.420 9 -1.551 8 New Zealand 0.007 32 -0.083 34 -0.255 31 Nordic Development Fund n/a n/a -0.553 29 -0.713 24 Norway -2.019 5 -1.494 6 -1.578 7 Other UN -1.481 13 -1.009 16 -0.203 33 Portugal -0.358 30 -0.090 33 -0.006 36 SAF & ESAF -1.596 11 -1.439 7 -1.853 4 Spain 0.118 36 0.492 40 0.041 37 Sweden -1.177 18 -0.836 20 -1.141 15 Switzerland -1.335 15 -1.162 14 -0.720 22 UNDP -1.243 17 -1.525 5 -1.433 10 UNFPA -0.972 23 -1.123 15 -1.118 16 UNHCR -0.804 27 -0.726 24 -0.515 25 UNICEF -1.747 7 -1.438 8 -1.301 12 United Kingdom -1.864 6 -1.230 12 -1.326 11 United States -1.329 16 -0.906 18 -1.162 14 UNRWA n/a n/a 0.066 37 0.090 38 UNTA -0.760 28 -0.575 28 -0.372 29 WFP -1.635 10 -1.713 4 -1.691 5 Note: Bolded elasticities are different from zero at 10% significance level (with heteroskedasticity-consistent s.e.). 30 Table 13. Overall Aid Selectivity Index, 1984-1999 1984-1989 1990-1994 1995-1999 Donor Mean Rank Mean Rank Mean Rank Total Aid 0.503 0.416 0.776 Bilateral Aid 0.511 0.262 0.436 Multilateral Aid 0.706 1.078 1.439 AfDF 3.233 1 2.813 2 3.031 4 Arab Agencies 0.231 27 0.996 13 1.758 11 AsDB Spec. Funds -0.534 36 -0.484 38 0.256 29 Australia 1.407 8 1.205 9 2.711 6 Austria -0.617 37 0.993 14 1.466 15 Belgium 0.134 28 0.565 24 0.280 28 Canada 0.756 18 0.157 30 0.665 26 CarDB 0.055 29 -0.134 36 0.125 31 Council of Europe -0.075 32 -0.027 32 n/a n/a Denmark 2.370 2 3.335 1 3.990 1 EBRD n/a n/a 1.260 8 0.831 24 EC 1.322 9 1.155 10 1.596 13 Finland 0.821 15 0.839 18 2.650 7 France 0.780 16 0.403 27 -1.065 39 GEF n/a n/a n/a n/a 0.865 22 Germany 0.873 14 0.843 17 1.089 20 Greece n/a n/a n/a n/a 0.093 32 IDA 2.223 3 2.369 3 3.755 2 IDB Spec. Oper. Fund -1.564 38 -0.967 39 -1.789 42 IFAD 0.685 20 1.002 12 2.552 8 IMF Trust Fund 0.034 30 n/a n/a n/a n/a Ireland 0.660 21 0.943 15 2.394 9 Italy 1.444 7 0.457 25 -0.473 37 Japan 0.710 19 1.449 6 1.269 18 Luxembourg n/a n/a 0.031 31 1.221 19 Netherlands 1.117 10 1.132 11 1.460 16 New Zealand 0.399 25 0.448 26 0.838 23 Nordic Development Fund n/a n/a 1.346 7 1.533 14 Norway 2.106 4 1.590 5 2.797 5 Other UN 1.479 6 0.765 20 0.191 30 Portugal -0.160 35 -1.023 40 -1.606 41 SAF & ESAF 0.892 13 0.805 19 2.003 10 Spain -0.160 34 -1.606 41 -1.229 40 Sweden 0.976 11 0.744 21 1.699 12 Switzerland 0.328 26 0.706 22 0.787 25 UNDP 0.767 17 0.898 16 1.004 21 UNFPA 0.529 22 0.574 23 -0.061 35 UNHCR 0.895 12 0.384 28 0.013 34 UNICEF 0.487 23 -0.039 34 -0.491 38 United Kingdom 2.020 5 1.859 4 3.145 3 United States -0.133 33 -0.097 35 1.371 17 UNRWA n/a n/a -0.034 33 -0.228 36 UNTA 0.478 24 0.364 29 0.074 33 WFP -0.066 31 -0.362 37 0.428 27 31 Appendix Table 1. Sensitivity Tests for Policy Elasticity Index with Different Aid Flows (2002) Gross ODA (see Table 1) Grants Only Gross ODA Net of Debt Relief7 Donor Mean Rank Mean Rank Mean Rank Total Aid 1.759 1.199 1.888 Bilateral Aid 0.628 0.662 0.630 Multilateral Aid 2.568 1.768 2.663 AfDF 1.022 22 1.296 18 1.022 23 Arab Agencies 0.971 25 -0.270 30 0.971 24 AsDB Spec. Funds -0.364 33 0.118 26 -0.364 32 Australia 1.008 23 1.008 19 1.198 22 Austria 3.259 5 3.259 5 3.144 5 Belgium 1.92 15 1.849 14 2.081 15 Canada 0.99 24 0.990 20 0.960 25 CarDB 1.404 18 0.547 23 1.404 18 Council of Europe no aid reported n/a no grants reported n/a no aid reported n/a Denmark 4.767 1 4.911 1 5.000 1 EBRD 1.394 20 1.394 16 1.394 20 EC 2.44 12 2.013 13 2.440 12 Finland 2.536 11 2.556 8 2.536 10 France -0.072 30 -0.264 29 0.005 30 GEF 3.027 7 3.027 6 3.027 7 Germany 2.064 14 2.151 11 2.193 13 Greece -0.559 36 -0.559 33 -0.559 35 IDA 4.233 2 4.108 2 4.163 2 IDB Spec. Oper. Fund -0.24 32 -0.255 28 -0.240 31 IFAD 1.403 19 no grants reported n/a 1.403 19 IMF Trust Fund no aid reported n/a no grants reported n/a no aid reported n/a Ireland 1.383 21 1.383 17 1.383 21 Italy -0.12 31 0.079 27 -0.625 36 Japan 1.901 16 2.162 10 1.817 16 Luxembourg 2.146 13 2.146 12 2.146 14 Netherlands 2.647 9 2.649 7 2.512 11 New Zealand 0.409 29 0.409 25 0.409 29 Nordic Development Fund 3.085 6 no grants reported n/a 3.085 6 Norway 2.701 8 2.514 9 2.701 8 Other UN -0.504 34 -0.504 31 -0.504 33 Portugal 0.54 28 0.540 24 0.531 28 SAF & ESAF 2.569 10 no grants reported n/a 2.569 9 Spain -1.253 39 -1.187 36 -1.233 39 Sweden 3.315 4 3.298 4 3.315 4 Switzerland 1.673 17 1.672 15 1.673 17 UNDP -0.531 35 -0.531 32 -0.531 34 UNFPA -2.082 41 -2.082 38 -2.082 41 UNHCR -1.48 40 -1.480 37 -1.480 40 UNICEF -2.308 42 -2.308 39 -2.308 42 United Kingdom 3.657 3 3.443 3 3.444 3 United States 0.664 27 0.633 22 0.612 27 UNRWA -0.97 38 -0.970 35 -0.970 38 UNTA -0.966 37 -0.966 34 -0.966 37 WFP 0.72 26 0.720 21 0.720 26 7Debt relief defined as DAC's reported debt forgiveness grants and reorganized debt. 32 Appendix Table 2. Policy Elasticity Index with Shocks and Post-Conflict Dummy in the Regression (2002) Original (Table 1) With Shocks and Post-Conflict Donor Mean Rank Mean Rank Total Aid 1.759 1.638 Bilateral Aid 0.628 0.637 Multilateral Aid 2.568 2.238 AfDF 1.022 22 2.074 14 Arab Agencies 0.971 25 1.047 26 AsDB Spec. Funds -0.364 33 0.021 31 Australia 1.008 23 1.334 22 Austria 3.259 5 2.691 9 Belgium 1.92 15 1.653 20 Canada 0.99 24 0.462 29 CarDB 1.404 18 1.079 24 Council of Europe no aid reported n/a no aid reported n/a Denmark 4.767 1 4.260 2 EBRD 1.394 20 1.147 23 EC 2.44 12 1.845 17 Finland 2.536 11 2.563 11 France -0.072 30 -0.602 35 GEF 3.027 7 3.320 4 Germany 2.064 14 1.933 15 Greece -0.559 36 -0.818 36 IDA 4.233 2 4.274 1 IDB Spec. Oper. Fund -0.24 32 -0.263 32 IFAD 1.403 19 1.789 19 IMF Trust Fund no aid reported n/a no aid reported n/a Ireland 1.383 21 1.592 21 Italy -0.12 31 1.064 25 Japan 1.901 16 2.276 13 Luxembourg 2.146 13 2.308 12 Netherlands 2.647 9 2.575 10 New Zealand 0.409 29 0.650 28 Nordic Development Fund 3.085 6 3.498 3 Norway 2.701 8 2.782 8 Other UN -0.504 34 -0.282 33 Portugal 0.54 28 0.874 27 SAF & ESAF 2.569 10 3.183 6 Spain -1.253 39 -0.904 38 Sweden 3.315 4 3.169 7 Switzerland 1.673 17 1.808 18 UNDP -0.531 35 -0.352 34 UNFPA -2.082 41 -1.788 42 UNHCR -1.48 40 -1.263 40 UNICEF -2.308 42 -1.450 41 United Kingdom 3.657 3 3.302 5 United States 0.664 27 0.344 30 UNRWA -0.97 38 -0.907 39 UNTA -0.966 37 -0.877 37 WFP 0.72 26 1.909 16 33