Global aid allocation: Are Nordic donors different?

CSAE WPS/2004-34 Global Aid Allocation: Are Nordic Donors Different? Scott Gates Centre for the Study of Civil War, PRIO and Norwegian University of Science and Technology (NTNU) and Anke Hoeffler Centre for the Study of African Economies, University of Oxford and International Peace Research Institute, Oslo (PRIO) December 2004 Both authors gratefully acknowledge funding from the MULTI program of the Research Council of Norway for this project. We thank Morten Bøås, David Dollar, Paul Collier, Jan Dehn, Nils Petter Gleditsch, Rune Hagen, Trond Folke Lindberg, Hildegunn Nordås, Arve Ofstad, Olav H. Seim, and Astrid Suhrke for their valuable help and comments. Abstract: The Nordic development assistance programs have earned a reputation for commitment to human rights and democracy. Is the reputation deserved? We address this question by comparing how much aid donors give and to which recipient countries. Using a global panel data set, spanning the period 1980-99 and 91 recipient countries, we find that individual bilateral donors vary considerably from one another. Nordic aid distribution differs significantly from other bilateral aid donor patterns: Norway, Denmark, Sweden and Finland provide more aid to democracies but do not penalise poor trade policies. Unlike other bilateral donors the four Nordics do not provide more aid to political allies. We also find some evidence that recipients with a good human rights record receive more aid from Nordic donors. 1. Introduction International aid agencies are motivated by different objectives. For example, NORAD, the Norwegian aid agency states: ‘The purpose of Norwegian development cooperation is to contribute towards lasting improvements in economic, social and political conditions for the populations of developing countries, with particular emphasis on ensuring that development aid benefits the poorest people.’ Furthermore, one of the main goals is ‘To contribute towards promoting peace, democracy and human rights’.1 Many other international aid agencies state the achievement of the Millennium Development Goals as their main objective. Denmark, Germany, the UK, the Netherlands and other donors state explicitly that aid is aimed at reducing global poverty. However, other aid agencies seem to be at least partly motivated by self-interest. USAID’s mission statement makes this explicit: ‘U.S. foreign assistance has always had the twofold purpose of furthering America's foreign policy interests in expanding democracy and free markets while improving the lives of the citizens of the developing world.’2 In this paper we examine the patterns of aid allocation across different donor countries. The consensus is that strategic interests, colonial history, trade, and political institutions of the recipient country dominate bilateral aid (most recent studies include Alesina and Dollar (2000), Boschini and Olofsgård (2001) and Neumayer (2003)). The Nordic donors, however, tend to be regarded as exceptions. One problem in the existing literature is that the Nordic countries, Norway, Denmark, Sweden and Finland are usually aggregated in the analysis and to date no systematic examination of Nordic aid distribution patterns has been made. Using comprehensive OECD data on aid disbursements during the period 1980 to 1999, we compare bilateral aid flows, focusing in particular on the individual Nordic countries. Like previous studies we find that there are profound differences 1 http://www.norad.no 2 http://www.usaid.gov between the specific political factors that shape different countries’ aid allocation patterns. Our results indicate that the aid allocation patterns of the Nordic countries are not the same as those of other bilateral aid agencies. We find strong evidence that Nordic donors provide more aid to the poorest countries and to democracies but do not penalise countries with less open trade policies. We also find some indication that recipients with a good human rights record receive more aid from Nordic donors. The paper is organized as follows: after a brief review of previous research regarding the distribution of aid we present some descriptive data focusing on Nordic countries. In Section 3 we use a global panel data set for our regression analysis for a comparison of individual donor aid allocations. Section 4 concludes. 2. Data Description – Patterns of Aid Allocation 2.1 Global Patterns of Bilateral Aid Allocation A number of studies have examined the patterns of aid allocation (see, for example, McKinley and Little, 1978; McKinley and Little, 1979; Mosely, 1981; Maizels and Nissanke, 1984; Trumbull and Wall, 1994; Ludborg, 1998; Schraeder, et al., 1998; Alesina and Dollar, 2000; Burnside and Dollar, 2000; Boschini and Olofsgård, 2001; Collier and Dollar, 2002; Neumayer, 2003). With few exceptions most analyses of bilateral aid allocations show that political and strategic interests of donors trump concern for growth or poverty reduction. The Nordic countries, however, tend to stand out as exceptions. The problem is that for most analyses the Nordics are aggregated3 and no systematic examination of individual Nordic donors’ allocation patterns has been made. Alesina and Dollar conclude that in the aggregate Nordic donors provide the right incentives; more aid goes to poor recipients with good trade policies and open democratic regimes. This may be partly due to historical reasons, Nordic donors have no colonial legacies and since all of the four donors are small countries they do not try to foster global strategic interests. Furthermore, all Nordic donors have a tradition of a social democratic welfare state where all citizens are entitled to welfare. This principle of universality generates a shared conception of citizenship and social spending is comparatively high in Nordic countries. Moreover, this system of socio- 3 Alesina and Dollar (2000) even include the Netherlands and Canada in their aggregation. political values does not only provide the determinants for domestic but also for international justice. Foreign aid can thus be seen as an extension of the principles of the welfare state beyond the domestic borders.4 Noël and Thérien (1995) suggest in their empirical study that states with large welfare budgets are more generous in providing foreign aid. Norway, the Netherlands, Denmark and Sweden are among the welfare states with large aid budgets, however, this correlation does not hold for Finland, Switzerland and France. We now turn to a description of Nordic aid allocation before we present a detailed regression analysis of the factors that determine this allocation in section 3. 2.2 Nordic Patterns of Bilateral Aid Allocation Throughout this paper we use the definition of aid as used by the OECD and base all of our empirical work on their data. The OECD provides online information on the aid flows from bilateral and multilateral donors to recipient countries. In our analysis we included official development assistance (ODA) to developing countries as well as official aid (OA) to the so called "Part II" countries which include more advanced developing countries as well as Central and Eastern European Countries and Newly Independent States of the former Soviet Union (CEEC and NIS).5 Data are available from 1960 and is provided in current US dollars. In this paper we analyse the total net ODA/OA flows for the period 1980-99. Total ODA/OA includes grants or loans to countries that are undertaken by the official sector in order to promote economic development and welfare. The financial terms are concessional; if the ODA/OA consists of a loan it must have a grant element of at least 25 per cent. While technical co-operation is included, grants, loans and credits for military purposes are excluded. For most years the flows are positive, however, due to the re-payments of loans the flows are negative for some years. Using the US GDP deflator we calculate aid flows in constant 1995 dollars. In this section we concentrate on the average taken over the years 1997, 1998 and 1999 in order to smooth out any unusual events. The pattern of aid allocation by donors is shown in Table I. Bilateral aid from the 22 Development Assistance Committee 4 For a detailed discussion see Stokke (1989). 5 For a detailed list please refer to the Appendix. The concepts of ODA and OA are the same -- they only differ in the type of recipient country. (DAC) member countries6 accounts for the largest share, 65.4 percent, of total global aid. About 45 percent of total aid originates from only four countries: Japan (17.3 percent), USA (9.2 percent), France (9 percent) and Germany (7.5 percent). The Nordic donors contribute considerably less in absolute terms; nevertheless, Denmark, Sweden and Norway contribute each between 2.5 and 2.9 percent of total global aid. Multilateral aid agencies give 33.6 percent. The remaining two percent of total global aid that is not from a DAC country or a multilateral organization are accounted for by bilateral Arab donors. Relative to their GDP, donors differ hugely in their generosity7. As column 4, Table I shows only four countries exceed the UN target of 0.7 percent of Gross National Income (GNI) and three of these four are Denmark, Norway and Sweden. Denmark's ODA to GNI ratio was 0.99 percent, while Norway’s was 0.87 percent, and Sweden’s was 0.74 percent. We now examine who receives this aid. In Table II we list the major recipients of Nordic Aid. For Norway topping the list are Tanzania and Mozambique, receiving 7.1 percent and 6.9 percent of Norwegian aid, respectively. Other important recipients are the Federal Republic of Yugoslavia, Bosnia, Zambia and Bangladesh. Sweden allocates the largest share to Tanzania (6.8 percent) and Mozambique (about 6 percent) followed by South Africa (about 5 percent). Much of Danish aid is also allocated to Sub-Saharan Africa. Top of the list are Tanzania (8.8 percent), Uganda (7.8 percent) and Mozambique (5.3 percent). Finland’s allocation is somewhat different, neighbouring Russia heads the list (9.3 percent), followed by China (8.2 percent) and Mozambique (6.7 percent). The Republic of Yugoslavia and Bosnia and Herzegovina also feature prominently. Estonia just makes it into the top ten, receiving 3.8 percent of Finnish aid. To summarize, much of Norwegian, Danish and Swedish aid is allocated to Sub-Saharan African countries; for each donor six out of the top ten recipients are in Sub-Saharan Africa. Finland’s aid allocation seems to be more dominated by regional concerns. Russia and Estonia are among the top ten recipients, while only two Sub-Saharan countries receive considerable amounts of aid (Mozambique and Tanzania). 6 A list of the 22 DAC countries is listed in the Appendix. 7 Hopkins (2000) provides a detailed discussion of aid volumes over time. Recently, donors have been encouraged to coordinate bi-lateral aid in the recipient countries in order to improve aid efficiency and public service delivery (World Bank, 2004). Furthermore, tied aid is now widely regarded as being more of an aid for the donor country than for the recipient countries and channelling aid through non-governmental organizations (NGOs) has been seen as preferable to giving aid to corrupt governments. In Table III we examine how much aid is channelled through multilateral organisations and through NGOs, how much is provided in grants and how much tied aid each Nordic donor provides. In each case we compare this with the average of the 22 DAC donors. The average DAC country gives about two thirds of ODA directly to the recipient countries and channels about one third through multilateral agencies. Norway and Sweden provides a relatively high proportion of ODA bilaterally, 74 and 70 percent, respectively. Thus, less than the average is channelled through multinational organizations. Denmark and Finland in contrast spend a lower than average proportion of their ODA on bilateral aid, 59 and 58 percent, respectively. The average DAC country provides about 90 percent of its aid as grants, for all Nordic countries this is considerably higher, just under 100 percent. About one quarter of all DAC aid is provided as tied aid. Among the Nordic donors, Denmark provides a relatively high proportion of ODA as tied aid (29 percent), followed by Finland (15 percent). The other two Nordic donors tie very little of their bilateral aid, Norway only provides one percent as tied aid, while Sweden provides about two percent. Interestingly, Nordic donors, with the exception of Sweden, channel negligible amounts of aid through NGOs. In order to complete our brief overview of Nordic aid we examine the allocation by major purpose in Table IV. Traditionally much of aid has been provided as project aid, often in an attempt to address the issues of fungibility and to provide show cases for donor involvement.8 However, programme aid can help to strengthen ownership of reform in selected countries and reduce the cost of collaboration between donor and recipient (World Bank, 2004). Comparing the major purposes of aid commitments we see that all of the Nordic donors provide more aid for social and administrative infrastructure, but less aid for economic infrastructure than the average DAC donor. Within the provision of social infrastructure all Nordic donors put a 8 Mosley and Eeckhout (2000) discuss project and programme aid. particular emphasis on the provision of health care. All Nordics provide free access to health care in their countries and may thus have a particular motivation and expertise in providing basic health care in recipient countries. All of the Nordic donors provide less programme assistance than the average DAC donor (7 percent); Norway only provides 0.5 percent as programme assistance. Our brief overview provides us with a number of stylised facts regarding contemporary aid distribution. Multilateral agencies provide one third of the total global aid budget and four large donors provide about 54 percent of total global aid. Nordic countries may be comparatively small donors in total aid terms but are very generous in providing aid. Denmark, Sweden and Norway are among the few donors who fulfil the UN target of giving more than 0.7 percent of their GNI as aid. Much of Nordic aid goes to Sub-Saharan Africa, with the notable exception of Finland. Much of Finnish aid is allocated to the neighbouring Russia and Estonia. Nordic countries provide a larger than average share of aid as grants. With the exception of Denmark very little aid is tied. Apart from Sweden hardly any aid is channelled through NGOs. All Nordic donors tend to concentrate on the provision of social infrastructure rather than economic infrastructure, a relatively large proportion is provided for health care. Nordic donors prefer project aid to programme aid. In order to examine these differences in aid allocation in more detail, we now turn to regression analysis in Section 3. 3. Data Analysis We analyse the global allocation of aid during 1980-1999 by using averages for five year sub periods, 1980-84, 1985-89, 1990-94 and 1995-99. We are able to estimate the aid allocation to 91 recipient countries. However, we do not have information for all countries for all years, making this global panel unbalanced. We estimate our model by pooled OLS and report robust standard errors. The dependent variable is the logarithm of aid given by a particular donor or a group of donors.9 As explanatory 9 Following the standard practice in this literature we added a small amount of aid to all observations, thus avoiding missing values when taking the logarithm of the zero observations. As an alternative treatment of the zero values we used Tobit estimation and found the results to be qualitatively similar. Results presented in Table VII. variables we use a range of socio-economic characteristics of recipient countries, as well as their history and geo-strategic importance. The results are shown in Table V. In the first two columns we investigate the donor behaviour of the multinational versus bilateral donors. Here we summed the total aid given by all multi-lateral agencies and for the bilateral aid we summed the total aid given by the 22 DAC countries. The results suggest that the aid allocation of multi- and bilateral donors is relatively similar. The effect of income is negative, donors give more aid to poorer countries.10 Countries with large populations also receive more aid. We use the parallel market premium as an indicator of trade policy. Repressive trade regimes are characterized by a higher premium, but it can also be argued that generally poor macro-economic management causes a higher premium. We think that this proxy of trade policies is a suitable one because it allows us to analyse macroeconomic policies in general. In any case it is preferable to a dummy categorizing open or closed trade regime because this continuous variable allows us to consider different degrees of openness. For multi- as well as bi-lateral donors the coefficient on trade policy is insignificant. We measure democracy using the Polity IV data set which measures openness of political institutions on a scale of 1 to 10, with higher values indicating more democratic regimes (Jaggers and Gurr, 1995). Democracy is significant at the ten percent level, more democratic countries receive more aid. We also include a measure of human rights violations which is based on information published by Amnesty International. The index (as compiled by Cornett and Gibney, 2004) runs from 1 to 5, with higher values indicating worse human rights. Our democracy and human rights variables are not highly correlated; the correlation coefficient is only -0.06. Thus, democracy and the human rights variable seem to measure different aspects of political rights. The human rights index is not significant, (i.e. neither multilateral nor bilateral donors seem to take human rights into account when allocating aid). With respect to regional allocation multi- and bilateral donors are somewhat different. While both give more to Egypt, the multilateral donors give significantly less to Israel while the bilateral donors give significantly more to Israel. 10 Like Alesina and Dollar (2000) we also ran specifications with a squared income term. Although we find both the level and the squared term to be significant we decided not to report these results because the inflexion point was outside the sample range. Thus, we conclude that the often hypothesized curvilinear relationship is not relevant for any of the recipient countries included in our analysis. Neither donor group gives more aid to Latin American nor to Sub-Saharan African countries; these dummies are insignificant. Religion does not seem to be a very important determinant either. Countries with higher proportions of Catholics or Muslim populations are no different than other countries; the coefficients are insignificant. However, countries with a higher proportion of Protestants receive less aid. To summarize, when looking at aggregates multi- and bilateral donors allocate aid according to income, population and democracy. Egypt receives more aid than other countries, ceteris paribus. Human rights and trade policies do not seem to be important factors in the global allocation of aid. Our model is provides a better explanation for multilateral aid (R2=0.59) than for bilateral aid (R2=0.42). A comparison of the aid allocation patterns of Nordic donors to the previous bilateral donors suggests that Nordic donors follow a somewhat different pattern. In column 3 we consider the total aid given by Denmark, Finland, Norway and Sweden. The income effect is negative and larger countries receive more aid. This conforms to the general pattern of bilateral aid allocation although the descriptive statistics indicating that Nordic donors allocate a lot of aid to smaller countries. The coefficient on the parallel market premium is positive, i.e. countries with closed trade policies receive more aid. Countries with higher democracy scores receive more aid and countries with worse human rights records receive less aid. Egypt receives more aid, ceteris paribus. The Israel dummy is negative at the ten percent level. Our descriptive statistics in section 2 seemed to suggest that Nordic donors give a lot of aid to SubSaharan African countries; however, the region as a whole does not receive preferential treatment when we account for a number of country specific characteristics. The Latin America dummy is also insignificant. Countries with higher proportions of Protestants, Catholics and Muslims receive less aid. Our model provides a reasonable fit (R2 =0.52). In columns 4-7 we investigate the four Nordic donors separately. Like in the previous models the coefficient on the income term is negative and highly significant. Nordic donors allocate more aid to poorer countries. Like the other DAC donors all Nordic donors give more to large countries. All Nordics give more aid to countries with poor trade policies but also to more democratic regimes. Only Denmark gives less aid to countries with poor human rights, this variable is not significant for the other three donors. Egypt receives significantly more aid from Denmark, Sweden and Finland. Israel receives significantly less aid from Norway, Denmark and Finland. Sub-Saharan Africa does not receive more aid but Denmark and Finland allocate significantly less to Latin American countries. Norway, Denmark and Finland give less aid to countries with high proportions of Protestants, Catholics and Muslims. For Sweden, none of the religious variables are significant. We also introduce an index of political allegiance between the donor country and the recipients. Using UN voting patterns, Gartzke and Jo (2000) derive an index from -1 to 1 with higher values indicating more similar voting patterns. With respect to this UN variable none of the Nordic donors give more aid to political allies. To summarize, unlike the average bilateral donor, Nordic donors allocate aid according to democracy and human rights records but not to political allies. Countries with poor trade policies receive more aid, ceteris paribus. In general, the model seems more suited to explain the Norwegian and Danish allocation than the Swedish and Finnish one. Our model also provides a good explanation of the aid allocation by the big five bilateral donors (USA, UK, France, Japan and Germany) and for the Netherlands and Canada. We present these results in columns 9 to 15. All donors give more aid to larger countries. The parallel market premium is significantly negative for the US and Japan, i.e. countries with less open policies receive less aid. In contrast, France, the Netherlands and Canada give more aid to poor trade policy countries. The US, the UK, Germany and the Netherlands allocate more aid to democracies; France, Japan and Canada do not. However, Japan and Canada give less aid to countries with poor human rights; this variable is insignificant for all the other donors. All donors without exception allocate more aid to recipients who are their political allies with respect to UN voting patterns. Recipients also tend to receive more aid if they are former colonies. Obviously, geo-strategic interests vary across donors but Egypt receives more aid from all donors apart from the UK. Israel receives more aid from the USA, Germany and the Netherlands but less from the UK and Japan. Sub-Saharan African countries receive less aid from Japan and Latin American countries receive more aid from the Netherlands and Canada but less from the UK. Catholic countries receive more aid from France and less aid from Japan and Canada. Protestant and Muslim countries tend to receive less aid across the donors. The model explains the aid allocation of France (R2=0.72) and the UK (R2=0.66) particularly well and least well for the US (R2=0.46). When we compare the aid allocation for individual donors we can thus observe distinct patterns. All of the donors provide more aid to poor countries. Two donors (US and Japan) give less aid to recipients with poor trade policies. Some donors allocate more aid to democracies and Japan as well as Canada give less aid to regimes with poor human rights records. However, all of these donors give more aid to former colonies and to their UN-friends. The patterns for the Nordic countries (Norway, Denmark, Sweden and Finland) are distinct. All Nordic donors give more aid to democracies and in the aggregate they give less aid to recipients with poor human rights. UN voting behaviour seems to be insignificant in their allocation decisions. However, poor trade policies are not penalised, countries with poor trade policies receive more aid. A recent debate in the World Bank regards the LICUS, Low Income Countries under Stress.11 These countries are very poor, have epidemic diseases, organized crime, some are experiencing civil war or a war has just recently ended. The term LICUS is currently applied to a large variety of countries, some are resource rich but policy poor (Angola), some have exceptionally weak government capacity (Haiti) and some are post-conflict countries (Sierra Leone). The Bank estimates that about 500 million people live in such countries. As can be seen in Table VI, six out of the current 13 LICUS countries receive aid from Nordic donors. Sweden and Norway provide aid to five countries each. In contrast, Denmark gives to only one of these countries. Finland does not provide aid to any LICUS country. Sweden and Norway in this way differ significantly from the other two Nordic donors. On a parenthetical note, two important donors, Canada and the Netherlands, are often seen as different to the big five donors and more like the Nordic donors. Our regressions suggest that they are indeed different. Although they both give more aid to UN- friends, they seem to reward democracies (Netherlands) and penalise human rights violations (Canada). However, comparing the results for individual bilateral donors in Table IV, we cannot conclude that the Netherlands and Canada are like Nordic donors. Their patterns of giving vary considerably from the Nordics, particularly in comparison to Norway and Sweden. They should therefore not be treated as if they follow the same behavioural patterns as for example Alesina and Dollar (2000) assume. 11 World Bank (2003). Robustness Checks Cornett and Gibney (2004) provide two measures of human rights. One is based on information published by Amnesty International and the other on information by the US State Department. In the appendix, Table Va, we examine whether the results are sensitive to the use of the different human rights variables. The two variables are highly correlated (ρ=0.87) and the results are qualitatively similar. The main difference is that the human rights variable is not significant at the conventional levels for the Nordic donors (column 3). All other results remain unchanged. We also examined whether the results are sensitive to the selection of a particular time period. Due to a number of reasons, for example the end of the Cold War and a shift in donor behaviour from conditionality towards selectivity, one could hypothesize that aid allocations in the 1980s and those in the 1990s were motivated by different determinants. The study by Dollar and Levin (2004) suggests that donors have become more selective in the sense that during the 1980s aid was allocated indiscriminately to recipients with poor or good governance, but that donors have since shifted to assisting countries with good governance. In order to test this hypothesis we run our model on data for the 1980s and the 1990s separately. We could not confirm the hypothesis that donors have become more selective with respect to recipients’ economic policy. With respect to democracy we could not find any change in donor behaviour either. However, the coefficient on human rights was insignificant for most donors in the 1980s but was significant for a number of donors in the 1990s. 12 Thus, we cannot confirm that donors’ behaviour changed significantly over the past 20 years, possibly with the exception of being more sensitive to the recipients’ human rights record. One further econometric issue is that the aid allocation regressions are not independent of each other. It may be the case that aid allocations are positively or negatively correlated. Donors may follow other donors and support certain recipients or perhaps decide to give aid to recipients who do not receive aid from other donors. Aid allocations may be complements or substitutes and thus not independent of each other. To investigate this possibility we used the method of Seemingly Unrelated 12 Multi, DAC and Nordic donors, Denmark, France (p=0.104), France, Japan and Canada. For Germany the human rights variable is significant in both periods, but only at the ten percent level in the 1980s. In the 1990s the variable is significant at the one percent level. Regressions (SURE). We present these results in Table VIII. In the first block we allow the error terms of the multi- and bi-lateral aid allocations to be correlated. The correlation coefficient is 0.79 and we can reject the hypothesis that this correlation is zero. However, the regression coefficients are similar to our OLS estimates and our results are qualitatively unchanged. We arrive at the same conclusion when we allow for correlation between the error terms of the multilateral allocation, the DAC bilateral allocation (excluding Nordic donors) and the Nordic donors. Again, we can reject the hypothesis that the error terms are uncorrelated. The main results remain unchanged. 4. Conclusions The Nordic development assistance programs have earned a reputation for their generosity as well as their commitment to human rights and democracy. In our paper we examine whether this reputation is deserved. First, we find that Nordic donors, with the exception of Finland, are relatively generous in the provision of aid. Norway, Denmark and Sweden are among the few countries to fulfil the UN target of giving at least 0.7 percent of their GNI as aid. Nordic donors give to poor countries, many of which are Sub-Saharan African. Relatively little of their aid is tied (Denmark is an exception) and they concentrate on social infrastructure provision, mainly in the health sector. Relatively little of their aid is channelled through multilateral agencies and the amount channelled through NGOs is negligible. We then address the question whether Nordic donors differ in their aid allocation patterns by comparing how much aid donors give and to which recipient countries. Using a global panel data set we find that individual bilateral donors vary considerably from one another. Nordic aid distribution differs significantly from other bilateral aid donor patterns: Norway, Denmark, Sweden and Finland provide more aid to democracies but do not penalise poor trade policies. We also find some evidence that recipients with a good human rights record receive more aid from Nordic donors. Unlike any of the other DAC donors, Nordic donors do not give more aid to political allies. Nordic aid allocation seems remarkably free from self-interest and, indeed, more orientated towards their stated objectives of poverty alleviation, the promotion of democracy and human rights. Norway and Sweden serve as leaders in these regards. One slight exception may be Finland, which more recently has provided more aid to recipients in its region (mainly Russia and Estonia) rather than to very poor developing countries. Bilateral donors have multiple objectives and in this study we confirm that bilateral donors’ aid allocations differ significantly. Our emphasis is largely on describing who gives aid to whom but we do not prescribe who should receive aid.13 Aid is most effective in recipient countries with good governance and strong institutions. In the past donors tried to buy policy reform by attaching conditions to aid packages; however, in general conditionality did not achieve the desired objectives. Conditionality did not bring about lasting reform if there was no strong domestic movement for change. Weak domestic ownership and an unwillingness of donors to withdraw assistance in cases of non-compliance are the most often cited causes behind the failure of conditionality.14 The work by Burnside and Dollar (2000) suggested selectivity as a new strategy; donors should give aid to recipients with good policies because only in good policy environments will aid be growth enhancing. The study by Burnside and Dollar received considerable attention and criticism. Hansen and Tarp (2001) and Dalgaard, Hansen and Tarp (2004) mainly base their criticism on the econometric shortcomings. A bigger criticism of the Burnside and Dollar article is a policy critique. If aid is mainly allocated to select countries with good policies, this implies that donors should disengage from countries with bad governance and poor institutions. Yet, if donors want to make progress on the Millenium Development Goals and combat the adverse regional and global consequences generated by LICUS, disengagement is not an option. Donor engagement in these countries must be different to the average poor country and should be tailored to the specific situation. Increasing funding is not enough because government delivery mechanisms are poor. Instead the emphasis should be on an indirect, but catalytic role in building the ownership of reforms in societies. This may include supporting civil society groups, independent think tanks and distribution of information. A focused reform agenda should be developed and agreed upon, the success of early reforms improves the chances of a larger future reform agenda. A further step would consist of strategic capacity building which 13 McGillivray (2004) provides a detailed discussion of descriptive and prescriptive approaches. 14 For a further discussion see Collier (1997) and World Bank (1998). could include secondments from IFIs and bilateral donors as well as the mobilization of expertise in the diaspora community. In addition the operational challenge of delivering services to poor people needs to be tackled. Since these countries do not provide strong authorizing environments for incurring debt liabilities the World Bank should not be the major development agency in LICUS. Bilateral donors, the EU and UNDP are grant making agencies and are therefore most likely to provide assistance to LICUS countries. These recent changes in the development agenda could enable Nordic donors to use their deservedly excellent reputation to initiate and help to design reforms, assist with capacity building and service delivery in LICUS. Norway and Sweden, in particular, already give aid to about half of these cases. These two donors are presented with an exciting opportunity to take the (joint) lead in some of these countries and help to develop the LICUS programme. 5. References Alesina, A. and D. Dollar. 2000. “Who Gives Foreign Aid to Whom and Why?” Journal of Economic Growth 5: 33-63. Andersen, O. M. 2000. “Sector Programme Assistance”, in F. Tarp (Ed.), Foreign Aid and Development: Lessons Learnt and Directions for the Future. Routledge, London and New York, pp. 178-194. Boone, P. 1996. “Politics and the Effectiveness of Foreign Aid”, European Economic Review, 40, 289-329. Boschini, A. and A. Olofsgård. 2001. “Foreign Aid: an Instrument for Fighting Poverty or Communism?” Mimeo. http://www.worldbank.org/wbi/B-SPAN/sub_instrumental.htm Burnside, C. and D. Dollar. 2000. “Aid, Policies and Growth”, American Economic Review, 90, 847-868. Collier, P. 1997. The Failure of Conditionality, in C. Gwin and J. Nelson (eds) Perspectives on Aid and Development. Washington D.C.: Overseas Development Council. Collier, P. and J. Dehn. 2001. “Aid, Shocks, and Growth”. World Bank Policy Research Working Paper No 2688. Washington DC. Collier, Paul and David Dollar. 2002. “Aid Allocation and Poverty Reduction”. European Economic Review 45: 1470-1500. Cornett, Linda and Mark Gibney. 2004. “Political Repression and Human Rights Abuse”. Mimeo. University of North Carolina, Asheville. Dalgaard, C.-J, H. Hansen, F. Tarp. 2004. ”On the empirics of foreign aid and growth”. EconomicJournal 114: F191- F216 DFID, 2001. Statistics on International Development. DFID, Statistics Department, East Kilbride. Dollar, D. and Levin, V. 2004. “The Increasing Selectivity of Foreign Aid, 19842002”. World Bank Policy Research Working Paper 3299, Washington D.C. Frey, B. S. and F. Schneider, 1986. “Competing Models of International Lending Activity”. Journal of Development Economics 20: 225-245. Gartzke, Erik and Dong-Joon Jo. 2002. The Affinity of Nations Index, 1946-1996. Version 3.0. http://www.columbia.edu/~eg589/datasets.html. Hansen, E and F. Tarp. 2001. “Aid and growth regressions”. Journal of Development Economics 64: 547-570. Hopkins, R. F. 2000. “Political Economy of Foreign Aid”, in F. Tarp (Ed.), Foreign Aid and Development: Lessons Learnt and Directions for the Future. Routledge, London and New York, pp. 423-449. Ludborg, P. 1998. “Foreign Aid and International Support as a Gift Exchange”. Economics and Politics 10: 127-141. McGillivray, M. 2004. “Descriptive and Prescriptive Analyses of Aid Allocation: Approaches, Issues, and Consequences”. International Review of Economics and Finance, 13, 275-292. McKinley, R.D. and R. Little. 1978. “The French Aid Relationship”. Development and Change XX: 459-478. McKinley, R.D. and R. Little. 1979. “The US Aid Relationship: A Test of the Recipient Need and the Donor Interest Models”, Political Studies, 29, 236-250. Maizels, A. and M. K. Nissanke. 1984. “Motivations for Aid to Developing Countries”, World Development, 12, 879-900. Morrissey, O. 2002. “British Aid Policy since 1997: Is DFID the Standard Bearer for Donors?”, CREDIT Research Paper, Centre for Research in Economic Development and International Trade, University of Nottingham. Mosely, P. 1981. “Models of Aid Allocation Process: A Comment on McKinley and Little” Political Studies 29: 245-253. Mosley, P. and Eeckhout, M. J. 2000. “From project aid to programme assistance”, in F. Tarp (Ed.), Foreign Aid and Development: Lessons Learnt and Directions for the Future. Routledge; London and New York, pp. 131-153. Neumayer, E. 2003. Neumayer, Eric. The Pattern of Aid Giving - The Impact of Good Governance on Development Assistance. Routledge: London and New York. Noël, A. and J.-P. Thérien. 1995. “From Domestic to International Justice: The Welfare State and Foreign Aid”. International Organization, 49(3): 523-553. OECD. 2001. International Development Statistics. Data CD Rom. Schraeder, P.J, S.W. Hook and B. Taylor, 1998. “Clarifying the Aid Puzzle: A Comparison of American, Japanese, French, and Swedish Aid Flows”. World Politics 50: 294-320. Stokke O. 1989. “The Determinants of Aid Policies: Some Propositions Emerging from a Comparative Analysis” in: Stokke O. (ed.) Western Middle Powers and Global Poverty: The Determinants of the Aid Policies of Canada, Denmark, the Netherlands, Norway and Sweden. Uppsala: The Scandinavian Institute of African Studies, 275322. Trumbull, W.N. and H.J. Wall, 1994. “Estimating Aid-allocation Criteria with Panel Data”, Economic Journal 104: 876-882. World Bank, 1998. Assessing Aid: What Works, What Doesn’t, and Why. Oxford University Press: Oxford. World Bank. 2003. Assessing Aid. What Works, What Doesn’t and Why. Policy Research Report. Oxford University Press: Oxford. World Bank. 2004. World Development Report. Oxford University Press: Oxford. Table I. Who Gives Aid? – A Donor Profile Total DAC Multilateral Nordics Japan USA France Germany UK Netherlands Denmark Sweden Canada Norway Australia Spain Italy Switzerland Austria Belgium Finland Portugal Ireland Luxembourg Greece New Zealand ODA (const US$) 39943.8 26140.3 13406.4 2313.1 6891.2 3691.7 3577.4 2999.3 1481.8 1347.8 765.8 718.6 717.5 652.8 571.8 567.0 453.4 435.5 381.6 297.2 175.9 132.8 96.7 68.1 59.8 56.5 % of Total % of Bilateral 100 65.4 33.6 5.8 17.3 9.2 9.0 7.5 3.7 3.4 1.9 1.8 1.8 1.6 1.4 1.4 1.1 1.1 1.0 0.7 0.4 0.3 0.2 0.2 0.1 0.1 na 100 na 8.8 26.4 14.1 13.7 11.5 5.7 5.2 2.9 2.7 2.7 2.5 2.2 2.2 1.7 1.7 1.5 1.1 0.7 0.5 0.4 0.3 0.2 0.2 ODA % GNI ODA per Capita 0.25 0.1 0.41 0.27 0.26 0.8 0.99 0.74 0.31 0.87 0.27 0.24 0.15 0.31 0.23 0.32 0.32 0.25 0.31 0.62 0.15 0.27 55 13 61 37 25 86 144 81 24 147 30 14 8 61 47 29 34 13 26 160 6 15 Note: We used the US GDP deflator to convert the current net aid flows into 1995 constant US dollars. Aid figures are three year averages (1997-99). The difference between the total aid and the sum of multilateral and DAC Bilateral Flows is mainly accounted for by Arab donors. Table III. Who Receives Aid – Recipient Profile Norway's Top Ten Recipients Sweden's Top Ten Recipients % of % of US$ Norway's US$ Sweden's (Millions) ODA (Millions) ODA Tanzania 46.05 7.05 Tanzania 48.91 6.81 Mozambique 44.75 6.85 Mozambique 42.91 5.97 Yugoslavia, Fed. Rep. 35.14 5.38 South Africa 35.84 4.99 Bosnia - Herzegovina 31.84 4.88 Viet Nam 32.50 4.52 Zambia 30.61 4.69 BosniaHerzegovina 27.31 3.80 Bangladesh 30.56 4.68 Ethiopia 27.27 3.79 Uganda 26.82 4.11 Bangladesh 24.84 3.46 Russia 25.39 3.89 Nicaragua 23.60 3.28 Ethiopia 25.09 3.84 Angola 21.65 3.01 Angola 21.38 3.28 Zimbabwe 19.59 2.73 Denmark's Top Ten Recipients Finland's Top Ten Recipients % of US$ Denmark's US$ % of country (Millions) ODA (Millions) Finland's ODA Tanzania 67.91 8.87 Russia 16.43 9.34 Uganda 59.62 7.79 China 14.31 8.14 Mozambique 40.88 5.34 Mozambique 11.72 6.67 Bangladesh 40.28 5.26 Tanzania 10.31 5.86 Viet Nam 36.31 4.74 Yugoslavia, Fed 8.27 4.70 Ghana 35.60 4.65 BosniaHerzegovina 8.22 4.67 Egypt 32.65 4.26 Nepal 8.06 4.58 India 30.84 4.03 Nicaragua 7.81 4.44 Burkina Faso 28.78 3.76 Viet Nam 7.80 4.44 South Africa 26.99 3.52 Estonia 6.67 3.79 Note: We used the US GDP deflator to convert the current net aid flows into 1995 constant US dollars. Aid figures are three year averages (1997-99). Table III: Loans and Grants bilateral % of total multilateral % of total grants % of bilateral loans % of bilateral Tied aid % of bilateral Technical co-operation % of bilateral Developmental food aid % of bilateral Emergency and distress relief % of bilateral Contributions to NGOs% of bilateral Administrative costs% of bilateral Denmark Finland Norway Sweden 59.20 40.85 99.71 0.29 29.2 8.11 0.00 8.50 0.68 8.41 57.93 42.31 118.67 -18.67 15.3 25.17 0.00 19.23 1.75 6.29 73.50 26.50 98.61 1.39 0.9 13.49 0.00 24.57 0.00 6.75 70.31 29.69 99.74 0.26 1.9 4.11 0.00 23.71 8.92 7.87 DAC Countries 67.16 32.84 89.56 13.08 26.34 38.43 3.08 12.87 3.39 8.99 Notes: Tied aid are based on commitments (excluding technical co-operation and administrative costs). All figures are 1999 values. Data Source: OECD (2001). Table IV: Aid by Major Purposes Commitments % of bilateral Social and Administrative infrastructure Education a) of which: Basic education Health of which: Basic health Population b) Water supply and sanitation Government and civil society Other social infrastructure/service Economic infrastructure Transport and communications Energy Other Production Agriculture "Industry, mining and construction" Trade and tourism Multisector Programme assistance Debt relief c) Emergency aid Administrative expenses Unspecified TOTAL Denmark Finland Norway Sweden 45.6 1.2 0.1 14 10.4 0.1 20.1 8.4 1.9 9.4 6.6 2.5 0.2 13.2 13.2 0 0 10 1.7 0 0 10.2 9.9 100 34.2 7.4 0.2 6.4 3.5 0.5 4.6 8.2 7.2 5.3 1.1 3.6 0.6 4.9 3.9 0.9 0.1 8.3 1.4 14.1 18.3 6.7 6.8 100 43.9 10.4 6.5 5.8 2.3 2.1 3 12.8 9.8 6.5 1.7 3.1 1.7 6.7 5.1 1.4 0.2 10 0.5 2.2 21.5 5.5 3.1 100 33.8 5.8 2.9 4.2 1.2 2.7 3.6 10.5 7 9 3.4 2.6 3 3.8 3.1 0.1 0.6 6.3 1.8 2.8 23.6 7.9 11 100 TOTAL DAC 29.9 10.7 1.2 4.2 2 1.8 4.1 4.2 4.9 17.2 8.7 4.6 3.9 8.1 5.5 2.2 0.4 7.4 6.9 7.4 11.1 5.9 6.1 100 Notes: All figures are for 1999 and they provide percentages of bilateral aid commitments. a) Including students and trainees. b) Population and reproductive health. c) Including forgiveness of non-ODA debt. Source: OECD (2001) Table V: Aid Allocation by Donor Ln GDP Ln Population (1) (2) (3) (4) (5) (6) (7) Multi DAC Nordic Norway Denmark Sweden Finland -1.525 -1.054 -1.388 -1.149 -1.087 -0.759 -0.409 (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.003)*** 0.265 0.351 0.415 0.366 0.402 0.374 0.370 (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Trade Policy 0.002 -0.000 0.042 0.032 0.037 0.050 0.042 (BMP) (0.515) (0.954) (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Democracy 0.047 0.060 0.129 0.139 0.075 0.133 0.058 (0.062)* (0.056)* (0.000)*** (0.000)*** (0.024)** (0.000)*** (0.015)** Human Rights -0.041 -0.106 -0.240 -0.017 -0.448 -0.041 0.137 (AI) (0.707) (0.370) (0.048)** (0.879) (0.001)*** (0.760) (0.270) -1.667 0.610 -0.625 0.547 UN-friend Egypt Israel SS-Africa Latin America Catholic Protestant Muslim (0.123) (0.590) (0.657) (0.585) 0.847 2.306 2.127 0.371 2.944 0.626 2.755 (0.000)*** (0.000)*** (0.000)*** (0.184) (0.000)*** (0.064)* (0.000)*** -1.396 3.593 -1.115 -1.902 -1.631 -1.123 -1.189 (0.069)* (0.000)*** (0.076)* (0.002)*** (0.001)*** (0.223) (0.074)* -0.304 -0.249 0.093 0.057 0.141 -0.124 0.059 (0.199) (0.413) (0.799) (0.866) (0.716) (0.769) (0.837) 0.292 0.240 -0.477 -0.389 -0.809 -0.562 -0.803 (0.405) (0.580) (0.157) (0.228) (0.010)** (0.116) (0.001)*** -0.580 -0.351 -0.899 -1.134 -1.388 -0.405 -0.729 (0.183) (0.506) (0.039)** (0.005)*** (0.001)*** (0.425) (0.060)* -1.079 -1.630 -2.340 -1.605 -2.806 -0.626 -1.263 (0.097)* (0.059)* (0.007)*** (0.065)* (0.000)*** (0.537) (0.056)* 0.260 0.046 -0.932 -1.391 -1.572 -0.781 -1.150 (0.270) (0.861) (0.016)** (0.000)*** (0.000)*** (0.105) (0.001)*** Observations 296 296 301 290 270 290 288 R-squared 0.594 0.422 0.519 0.488 0.467 0.323 0.397 Note: OLS regressions with White corrected standard errors. P-values in parentheses, ***, ** and * denote significance at the one, five and ten percent level, respectively. Table V continued Ln GDP Ln Population (1) (2) (3) (4) (5) (6) (7) USA UK France Japan Germany Netherlands Canada -1.393 -0.820 -0.287 -0.681 -0.593 -0.992 -1.120 (0.000)*** (0.000)*** (0.010)** (0.000)*** (0.000)*** (0.000)*** (0.000)*** 0.060 0.526 0.609 0.674 0.520 0.454 0.541 (0.603) (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Trade Policy -0.045 -0.001 0.011 -0.021 0.004 0.021 0.010 (BMP) (0.000)*** (0.754) (0.022)** (0.000)*** (0.249) (0.000)*** (0.089)* Democracy 0.124 0.058 0.011 -0.000 0.077 0.123 0.022 (0.003)*** (0.024)** (0.613) (0.993) (0.001)*** (0.000)*** (0.440) Human Rights 0.131 -0.033 -0.117 -0.193 -0.031 0.059 -0.346 (AI) (0.456) (0.780) (0.176) (0.091)* (0.726) (0.601) (0.004)*** UN-friend Own colony Other colony Egypt Israel SS-Africa Latin America Catholic Protestant Muslim 2.621 2.399 3.706 6.133 5.073 1.668 2.929 (0.003)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.053)* (0.002)*** 0.598 0.574 0.777 0.000 0.137 0.812 (0.000)*** (0.000)*** (0.000)*** (.) (0.146) (0.000)*** 0.082 -0.029 0.136 0.165 0.038 0.229 (0.441) (0.691) (0.014)** (0.022)** (0.481) (0.000)*** 4.735 0.163 2.559 0.650 1.937 1.983 1.464 (0.000)*** (0.670) (0.000)*** (0.008)*** (0.000)*** (0.000)*** (0.000)*** 3.282 -4.812 0.218 -1.470 2.147 0.811 -1.057 (0.006)*** (0.000)*** (0.700) (0.025)** (0.001)*** (0.090)* (0.143) -0.267 -0.398 0.342 -1.513 0.265 0.143 -0.473 (0.595) (0.159) (0.141) (0.000)*** (0.310) (0.656) (0.132) 0.827 -0.582 -0.150 0.251 0.019 0.751 0.603 (0.160) (0.052)* (0.599) (0.404) (0.952) (0.022)** (0.059)* 0.403 -0.550 1.045 -1.059 0.012 -0.237 -1.022 (0.566) (0.133) (0.004)*** (0.005)*** (0.971) (0.553) (0.011)** -0.307 -0.890 -1.716 -1.716 -1.180 -1.916 -0.346 (0.764) (0.247) (0.002)*** (0.012)** (0.034)** (0.007)*** (0.668) 0.635 -1.063 0.432 -0.848 0.161 -0.868 -0.535 (0.205) (0.001)*** (0.105) (0.009)*** (0.537) (0.010)** (0.116) Observations 259 286 289 284 286 287 290 R-squared 0.456 0.661 0.721 0.607 0.563 0.521 0.504 Note: OLS regressions with White corrected standard errors. P-values in parentheses, ***, ** and * denote significance at the one, five and ten percent level, respectively. All regressions include time dummies. Table VI. Nordic Aid to LICUS (Low Income Countries Under Stress)15 LICUS pilot countries (03/04) Donor (rank) average 01/02 Angola Norway (5), Sweden (7) Guinea-Bissau Sweden (6) Haiti Liberia Papua New Guinea Somalia Norway (4), Sweden (8) Sudan Norway (5), Sweden (10) Tajikistan Zimbabwe Denmark (6), Norway (8), Sweden (9) Comoros Central African Republic Burundi Togo 15 LICUS as defined by the World Bank (2004). Norway(7) Table VII: Tobit Results Ln GDP Ln Population (1) (2) (3) (4) (5) (6) (7) Multi DAC Nordic Norway Denmark Sweden Finland -1.530 -1.059 -1.485 -1.401 -1.360 -0.948 -0.534 (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.001)*** 0.263 0.349 0.436 0.432 0.479 0.444 0.430 (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Trade Policy 0.002 -0.000 0.042 0.035 0.041 0.054 0.044 (BMP) (0.796) (0.966) (0.001)*** (0.003)*** (0.005)*** (0.001)*** (0.000)*** Democracy 0.048 0.061 0.137 0.156 0.101 0.157 0.085 (0.034)** (0.020)** (0.000)*** (0.000)*** (0.010)*** (0.000)*** (0.006)*** Human Rights -0.038 -0.104 -0.234 -0.031 -0.443 -0.015 0.164 AI (0.664) (0.307) (0.049)** (0.786) (0.003)*** (0.922) (0.160) -0.984 1.764 0.376 1.332 UN-friend Egypt Israel SS-Africa Latin America Catholic Protestant Muslim Observations (0.374) (0.208) (0.833) (0.350) 0.861 2.321 2.199 0.512 3.243 0.779 2.823 (0.134) (0.001)*** (0.005)*** (0.480) (0.000)*** (0.433) (0.000)*** -1.393 3.598 -0.952 -1.858 -2.433 -0.298 -0.859 (0.025)** (0.000)*** (0.255) (0.040)** (0.058)* (0.829) (0.453) -0.286 -0.231 0.130 0.057 0.298 -0.156 0.126 (0.253) (0.420) (0.698) (0.857) (0.470) (0.722) (0.697) 0.323 0.271 -0.331 -0.221 -0.672 -0.434 -0.786 (0.221) (0.373) (0.343) (0.513) (0.120) (0.351) (0.025)** -0.611 -0.381 -1.012 -1.219 -1.515 -0.602 -0.794 (0.065)* (0.323) (0.022)** (0.005)*** (0.004)*** (0.308) (0.072)* -1.131 -1.686 -2.482 -1.708 -3.519 -0.652 -1.387 (0.043)** (0.009)*** (0.001)*** (0.019)** (0.000)*** (0.514) (0.063)* 0.260 0.047 -0.970 -1.494 -1.742 -0.880 -1.096 (0.363) (0.889) (0.012)** (0.000)*** (0.000)*** (0.087)* (0.004)*** 296 296 301 290 270 290 288 Note: Tobit regressions. P-values in parentheses, ***, ** and * denote significance at the one, five and ten percent level, respectively. Table VIII: SURE Results (1) (2) Multi DAC Multi DAC Nordic Ln GDP -1.497 -1.047 -1.495 -1.042 (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Ln Population 0.272 0.349 0.270 0.339 0.428 non Nordic -1.286 (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Trade Policy 0.003 -0.000 0.003 -0.003 0.042 (BMP) (0.760) (0.971) (0.749) (0.781) (0.000)*** Democracy 0.049 0.064 0.054 0.065 0.140 (0.026)** (0.014)** (0.013)** (0.011)** (0.000)*** -0.039 -0.103 -0.057 -0.111 -0.240 (0.647) (0.304) (0.494) (0.265) (0.038)** Human Rights Egypt Israel SS-Africa Latin America Catholic Protestant Muslim 0.848 2.320 0.882 2.346 2.192 (0.126) (0.000)*** (0.106) (0.000)*** (0.003)*** -1.523 3.556 -1.557 3.585 -1.388 (0.012)** (0.000)*** (0.009)*** (0.000)*** (0.088)* -0.298 -0.235 -0.283 -0.276 0.238 (0.220) (0.411) (0.236) (0.331) (0.469) 0.223 0.231 0.249 0.218 -0.348 (0.387) (0.449) (0.326) (0.469) (0.317) -0.639 -0.342 -0.614 -0.301 -1.128 (0.048)** (0.369) (0.054)* (0.425) (0.010)*** -1.223 -1.661 -1.280 -1.648 -2.437 (0.023)** (0.009)*** (0.016)** (0.009)*** (0.001)*** 0.165 0.048 0.161 0.085 -1.055 (0.554) (0.884) (0.557) (0.794) (0.005)*** Observations 292 292 290 290 290 Test of χ2(1)=180.27 χ2(3)=293.45 Independence (0.0000) (0.0000) Note: Zellner’s seemingly unrelated regressions. P-values in parentheses, ***, ** and * denote significance at the one, five and ten percent level, respectively. Appendix Donor Countries: Development Assistance Committee (DAC) consists of the following 22 member countries: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Luxembourg, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States. Recipient Countries: Part I (Developing Countries and Territories), Afghanistan, Albania, Algeria, Angola, Anguilla, Antigua snd Barbuda, Argentina, Argentina, Armenia, Aruba, Azerbaijan, Bahamas, Bahrain, Bangladesh, Barbados, Belize, Benin, Bhutan, Bolivia, Bosnia and Herzegovina, Botswana, Brazil, Brunei, Burkina Faso, Burundi, Cambodia, Cameroon, Cape Verde, Cayman Islands, Central African Republic, Chad, Chile, China, Chinese Taipei, Colombia, Comoros, Democratic Republic of the Congo, Congo, Cook Islands, Costa Rica, Croatia, Cuba, Cyprus, Côte d'Ivoire, Djibouti, Dominica, Dominican Republic, Timor (East), Ecuador, Egypt, El Salvador, Equatorial Guinea, Eritrea, Ethiopia, Ex-Yugoslavia, Falkland Islands, Fiji, French Polynesia, Gabon, Gambia, Georgia, Ghana, Gibraltar, Grenada, Guatemala, Guinea, Guinea-Bissau, Guyana, Haiti, Honduras, Hong Kong, China, India, Indonesia, Indus Basin, Iran, Iraq, Israel, Jamaica, Jordan, Kazakstan, Kenya, Kiribati, Korea, Democratic Republic of Korea, Kuwait, Kyrgyz Republic, Laos, Lebanon, Lesotho, Liberia, Libya, Macao, Macedonia (former Yugoslav Republic), Madagascar, Malawi, Malaysia, Maldives, Mali, Malta, Marshall Islands, Mauritania, Mauritius, Mayotte, Mekong Delta Project, Mexico, Micronesia, Fed.States, Moldova, Mongolia, Montserrat, Morocco, Mozambique, Myanmar, Namibia, Nauru, Nepal, Netherlands Antilles, New Caledonia, Nicaragua, Niger, Nigeria, Niue, Northern Marianas Ilands., Oman, Pakistan, Palau, Palestinian Administrated Areas, Panama, Papua New Guinea, Paraguay, Peru, Philippines, Qatar, Rwanda, Samoa, Sao Tome and Principe, Saudi Arabia, Senegal, Seychelles, Sierra Leone, Singapore, Slovenia, Solomon Islands, Somalia, South Africa , Sri Lanka, St. Helena, St. Kitts-Nevis, St. Lucia, St. Vincent and Gr., Sudan, Suriname, Swaziland, Syria, Tajikistan, Tanzania, Thailand, Togo, Tokelau, Tonga, Trinidad and Tobago, Tunisia, Turkey, Turkmenistan, Turks and Caicos Islands, Tuvalu, Uganda, United Arab Emirates, Uruguay, Uzbekistan, Vanuatu, Venezuela, Viet Nam, Virgin Islands (U.K.), Wallis and Futuna, Yemen, Yugoslavia, Federal Republic of Zambia and Zimbabwe. Part II (More Advanced Developing Countries and Territories, CEECs and NISs) Bahamas*, Bermuda**, Brunei**, Cayman Islands**, Chinese Taipei**, Cyprus**, Falkland Islands**, Hong Kong, Israel**, Kuwait*, Qatar*, Singapore*, United Arab Emirates*, Belarus, Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Moldova, Poland, Romania, Russia, Slovak Republic and Ukraine. Note: Countries marked * (**) countries transferred to Part II of the list of recipients on 1st January 1996 (1st January 1997). Descriptive Statistics Variable Obs Mean Std. Dev. Min Max ln aid Multi 295 3.994493 1.691126 -2.302585 7.499106 ln aid DAC 295 4.981449 1.642774 -2.302585 8.046826 ln aid Nordic 301 1.388836 2.067621 -2.746202 5.783576 ln aid Norway 300 .1258485 1.871366 -3.267433 4.434826 ln aid Denmark 281 .0244244 2.1202 -6.242906 4.502431 ln aid Sweden 300 .1616112 2.046517 -2.302585 4.742642 ln aid Finland 298 -.7686568 1.657275 -3.286446 3.916511 ln aid USA 266 2.652011 2.248561 -2.302585 7.539036 ln aid UK 296 1.08997 2.067972 -2.688248 5.417122 ln aid France 299 2.279297 1.802724 -2.302585 6.239629 ln aid Japan 293 2.789519 1.978212 -2.302585 7.039991 ln aid Germany 294 2.990328 1.467749 -2.302585 6.384249 ln aid Netherlands 297 1.399527 1.784591 -2.302585 5.026023 ln aid Canada 300 1.31594 1.740259 -2.302585 4.843287 ln GDP 301 7.492099 .8610504 5.700444 9.904387 (ln GDP)2 301 56.87049 13.08573 32.49506 98.09689 ln Population 301 16.22056 1.515608 13.24458 20.91389 BMP 301 109.7522 701.5169 -89.16 11662.38 Democracy 301 3.033223 3.743291 0 10 UN-friend UK 290 .171524 .2066543 -.2614399 .9453666 UN-friend US 290 -.3878603 .3399373 -.8622133 .8511644 UN-friend France 290 .297018 .1528049 -.1190089 .918759 UN-friend Japan 290 .6968468 .1357362 .2490123 1 UN-friend Germany 290 .4980488 .1776126 -.0526277 1 UN-friend Sweden 290 .8333317 .1047406 .1643767 1 UN-friend Finland 290 .8332932 .1159606 .1581469 1 UN-friend Norway 290 .6881542 .1069672 .1435359 1 UN-friend Denmark 290 .6747575 .0969902 .1675813 .9809415 UN-friend Netherlands 290 .5099341 .1355987 .0410541 1 UN-friend Canada 290 .527515 .1638131 -.1033741 1 Human Rights AI 301 2.954873 .8966183 1 5 Human Rights US 301 2.728405 .9293629 1 5 Catholic 301 .3452824 .3633995 0 .966 Protestant 301 .1151296 .1532783 0 .715 Muslim 301 .285711 .3738647 0 .998 Correlation Coefficients ln aid DAC ln aid Multi ln aid DAC 0.7333 1.0000 ln aid Nordic ln GDP ln aid Nordic 0.6629 0.6134 1.0000 ln GDP -0.6633 -0.3359 -0.4822 1.0000 (ln GDP) 2 (ln GDP)2 Ln Pop. BMP -0.6776 -0.3507 -0.4930 0.9977 1.0000 ln Population 0.3685 0.4274 0.3749 -0.0040 -0.0118 1.0000 BMP -0.0120 -0.0294 0.0851 -0.0044 -0.0081 -0.0185 1.0000 Democracy -0.1316 0.0722 0.0404 0.3497 0.3367 0.0292 -0.0744 UN-friend UK -0.0749 -0.0062 -0.1177 0.1537 0.1567 -0.0861 -0.1375 UN-friend US -0.1793 -0.0314 -0.2274 0.1428 0.1488 -0.1286 -0.0780 UN-friend France -0.0353 0.0282 -0.0864 0.1210 0.1214 -0.0796 -0.1365 UN-friend Japan 0.1239 0.0359 -0.0179 0.0081 0.0039 -0.0439 -0.1602 UN-friend Germany 0.0046 0.0237 0.0563 0.1639 0.1643 0.0059 -0.1422 UN-friend Sweden 0.3254 0.1751 0.0941 -0.2682 -0.2795 -0.1225 -0.0123 UN-friend Finland 0.2822 0.1085 0.0545 -0.2639 -0.2735 -0.1282 0.0250 UN-friend Norway 0.1936 0.1534 0.0128 -0.0533 -0.0596 -0.1037 -0.1434 UN-friend Denmark 0.2269 0.2099 0.0442 -0.0640 -0.0716 -0.1156 -0.1391 UN-friend Netherl. 0.0230 0.0583 -0.0913 0.0745 0.0735 -0.1293 -0.1409 UN-friend Canada 0.0561 0.0819 -0.0064 0.0999 0.0996 -0.0650 -0.1905 Human Rights AI 0.2044 0.2198 0.0605 -0.0182 -0.0227 0.4810 0.0820 Human Rights US 0.2656 0.2237 0.2061 -0.1527 -0.1580 0.4893 0.1630 Democracy UN-friend UN-friend UN-friend UN-friend UN-friend UN-friend UK US France Japan Germany Sweden UN-friend UK 0.1772 1.0000 UN-friend US 0.0467 0.8567 1.0000 UN-friend France 0.2008 0.9677 0.8029 1.0000 UN-friend Japan 0.0876 0.8119 0.5753 0.7772 1.0000 UN-friend 0.3009 0.5846 0.2267 0.6171 0.4602 1.0000 -0.0585 0.1050 -0.0340 0.1920 0.4229 -0.1314 1.0000 -0.1371 0.0472 -0.0133 0.1133 0.3592 -0.3000 0.9692 0.1544 0.7232 0.4485 0.7511 0.8825 0.4268 0.6860 0.1998 0.5589 0.2433 0.6236 0.7235 0.4211 0.7621 0.1759 0.9467 0.7554 0.9669 0.8533 0.5792 0.3633 0.2575 0.8454 0.5199 0.8075 0.8696 0.7319 0.2151 Germany UN-friend Sweden UN-friend Finland UN-friend Norway UN-friend Denmark UN-friend Netherlands UN-friend Canada Human Rights AI -0.0605 -0.0688 -0.0441 -0.0870 -0.0334 -0.0140 -0.1074 Human -0.1006 -0.2171 -0.2523 -0.2254 -0.1578 0.0006 -0.1314 UN-friend UN-friend UN-friend UN-friend UN-friend Human Human Finland Norway Denmark Netherlands Canada Rights AI Rights US Rights US UN-friend 0.5833 1.0000 0.6389 0.9493 1.0000 0.2782 0.8710 0.7610 1.0000 0.0848 0.8364 0.7392 0.8559 Norway UN-friend Denmark UN-friend Netherlands UN-friend Canada 1.0000 Human Rights AI -0.1050 -0.0674 -0.0726 -0.0886 -0.0252 1.0000 Human Rights US -0.1386 -0.1640 -0.1309 -0.2278 -0.1036 0.8665 1.0000 Table Va: Aid Allocation by Donor (US State Department Human Rights data) Ln GDP Ln Population (1) (2) (3) (4) (5) (6) (7) Multi DAC Nordic Norway Denmark Sweden Finland -1.552 -1.188 -1.344 -1.104 -1.026 -0.664 -0.350 (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.001)*** (0.009)*** 0.279 0.376 0.426 0.366 0.390 0.374 0.362 (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Trade Policy 0.005 0.003 0.042 0.031 0.039 0.047 0.039 (BMP) (0.223) (0.684) (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Democracy 0.047 0.051 0.117 0.132 0.059 0.119 0.054 (0.061)* (0.104) (0.000)*** (0.000)*** (0.082)* (0.001)*** (0.022)** -0.168 -0.226 -0.151 0.012 -0.312 0.051 0.163 (0.128) (0.064)* (0.195) (0.911) (0.031)** (0.692) (0.157) -1.685 0.298 -0.762 0.548 Human Rights UN-friend Egypt Israel SS-Africa Latin America Catholic Protestant Muslim (0.120) (0.798) (0.572) (0.585) 0.800 2.178 2.125 0.351 2.961 0.685 2.781 (0.000)*** (0.000)*** (0.000)*** (0.206) (0.000)*** (0.045)** (0.000)*** -1.450 3.733 -1.150 -1.933 -1.886 -1.171 -1.158 (0.051)* (0.000)*** (0.061)* (0.001)*** (0.000)*** (0.193) (0.074)* -0.403 -0.467 0.219 0.067 0.313 0.031 0.075 (0.050)* (0.092)* (0.526) (0.831) (0.388) (0.938) (0.781) 0.193 0.028 -0.259 -0.309 -0.631 -0.314 -0.741 (0.531) (0.944) (0.390) (0.282) (0.028)** (0.306) (0.001)*** -0.399 -0.107 -1.099 -1.221 -1.607 -0.574 -0.750 (0.303) (0.826) (0.007)*** (0.001)*** (0.000)*** (0.239) (0.041)** -1.164 -1.513 -2.364 -1.668 -2.959 -0.951 -1.430 (0.057)* (0.063)* (0.003)*** (0.042)** (0.000)*** (0.313) (0.024)** 0.310 0.100 -0.996 -1.387 -1.717 -0.811 -1.118 (0.179) (0.703) (0.008)*** (0.000)*** (0.000)*** (0.083)* (0.002)*** Observations 308 308 313 301 281 301 299 R-squared 0.599 0.447 0.531 0.492 0.462 0.318 0.400 Note: OLS regressions with White corrected standard errors. P-values in parentheses, ***, ** and * denote significance at the one, five and ten percent level, respectively. Table Va continued … Ln GDP Ln Population (1) (2) (3) (4) (5) (6) (7) USA UK France Japan Germany Netherlands Canada -1.272 -0.788 -0.304 -0.742 -0.617 -0.951 -1.111 (0.000)*** (0.000)*** (0.008)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** 0.015 0.537 0.623 0.713 0.566 0.468 0.567 (0.893) (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** (0.000)*** Trade Policy -0.049 -0.002 0.013 -0.018 0.006 0.019 0.013 (BMP) (0.000)*** (0.666) (0.002)*** (0.000)*** (0.107) (0.000)*** (0.029)** Democracy 0.121 0.043 0.008 -0.026 0.070 0.104 0.003 (0.002)*** (0.084)* (0.695) (0.279) (0.003)*** (0.000)*** (0.915) 0.198 -0.091 -0.166 -0.278 -0.092 0.064 -0.327 (0.263) (0.471) (0.042)** (0.022)** (0.297) (0.572) (0.005)*** 2.637 2.327 3.674 6.083 5.008 1.657 2.634 (0.003)*** (0.001)*** (0.000)*** (0.000)*** (0.000)*** (0.061)* (0.007)*** 0.580 0.536 0.794 0.000 0.158 0.785 (0.000)*** (0.000)*** (0.000)*** (.) (0.082)* (0.000)*** 0.063 -0.084 0.156 0.125 0.048 0.202 (0.517) (0.253) (0.003)*** (0.069)* (0.366) (0.000)*** 4.842 0.099 2.550 0.547 1.939 1.993 1.492 (0.000)*** (0.782) (0.000)*** (0.022)** (0.000)*** (0.000)*** (0.000)*** Human Rights UN-friend Own colony Other colony Egypt Israel SS-Africa Latin America Catholic Protestant Muslim 3.108 -4.713 0.142 -1.281 2.284 0.980 -1.095 (0.009)*** (0.000)*** (0.784) (0.045)** (0.000)*** (0.035)** (0.074)* -0.182 -0.296 0.336 -1.484 0.317 0.229 -0.339 (0.695) (0.286) (0.127) (0.000)*** (0.212) (0.460) (0.258) 0.747 -0.492 -0.115 0.387 0.161 0.927 0.820 (0.153) (0.077)* (0.658) (0.180) (0.577) (0.003)*** (0.007)*** 0.411 -0.603 1.038 -1.210 0.013 -0.376 -1.179 (0.509) (0.083)* (0.002)*** (0.001)*** (0.968) (0.332) (0.003)*** -0.985 -1.078 -1.887 -1.413 -1.293 -2.036 -0.681 (0.338) (0.120) (0.000)*** (0.030)** (0.014)** (0.002)*** (0.380) 0.595 -0.921 0.374 -0.836 0.160 -0.849 -0.670 (0.232) (0.004)*** (0.136) (0.008)*** (0.533) (0.014)** (0.045)** Observations 270 297 300 295 297 298 301 R-squared 0.438 0.644 0.736 0.620 0.589 0.532 0.515 Note: OLS regressions with White corrected standard errors. P-values in parentheses, ***, ** and * denote significance at the one, five and ten percent level, respectively. Data Sources: Aid Average aid received in constant 1995 US Dollars (millions) over each five year period. Following Alesina and Dollar we added 0.1 to each observation before taking logarithms. Source: OECD 2001. Population Total population. Source: WDI 2001. Trade Policy (Black Market Premium) We measure openness to trade by taking the parallel (‘black’) market premium. In order to make the coefficients easier to compare we multiplied the premium by 100. Source: Global Development Network. www.worldbank.org\research\gdn Democracy We used the democracy score from the Polity IV data set which measures openness of political institutions on a scale of 1 to 10, with higher values indicating more democratic regimes. For further documentation see Jaggers and Gurr (1995). Human Rights We use two measures of human rights violations, one is based on information published by Amnesty International the other is based on US State Department information. The index ranges from 1 to 5, with higher values indicating worse human rights. Source: Cornett and Gibney, 2004. UN-friends Based on UN voting patterns, the UN-friend index ranges from -1 to 1, with higher values indicating more similar voting patterns. Source: Gartzke and Jo (2000). Religious Affiliations Catholic, Protestant and Muslim denote the percentage of the population in a country who declared in a survey that they follow a particular religion. We used data from 1980. Source: Barratt (1982) Colonies Dummies indicate whether a recipient was a former colony. Source: Burnside and Dollar (2000). Regions Regional dummies for Sub-Saharan Africa and Latin America were taken from Burnside and Dollar (2000).