Chapter 11: Reforming Development Cooperation to Attack Poverty

189

CHAPTER 11

Reforming DevelopmentCooperation

to Attack Poverty

Development cooperation is being reformed.From the relationship between donor and recipient tothe way in which aid is delivered and the frameworkfor debt relief for the poorest countries, many of theold ways of assisting development are beginning to bereplaced by new forms.

Much of this is due to a reaffirmed commitmentby the international community to fight poverty. TheWorld Summit for Social Development in Copen-hagen in 1995 set forth the goal of eradicating povertyin the world through decisive national actions andinternational cooperation. Donors have included halv-ing poverty between 1990 and 2015 and other targetsamong their international development goals (see box2 in the overview).1 In the 12th replenishment of theInternational Development Association (IDA) in 1998,donors reaffirmed their mission to support programsto reduce poverty and improve the quality of life inIDA’s poorest member countries.2 The Jubilee 2000movement helped put deeper debt relief at the heartof development cooperation strategies for poverty re-

duction. And donors are working to resolve differencesin approaches to poverty reduction through theOECD’s Development Assistance Committee (DAC),which expects by mid-2001 to agree on guidelines forpoverty reduction to help donor agencies make theirprograms more effective.3

But while the international community’s com-mitment to attack poverty was strengthened in the1990s, official development assistance shrank. This,despite the optimism at the start of the 1990s thatdevelopment cooperation would reap a post–coldwar “peace dividend” from cutbacks in military spend-ing.4 Indeed, after peaking in 1992 (in real terms),official development assistance fell consistently overthe decade—despite the robust economic growth ofDAC countries—rebounding only slightly in 1998during the global financial crisis (figure 11.1). Six-teen of the 21 DAC countries spent a smaller shareof their GNP on development assistance in 1997–98than in 1988–92.5 The regional distribution of thisaid remained roughly constant between 1987 and

1998, apart from an increase in the share going to Eu-rope and Central Asia (figure 11.2). But total develop-ment assistance fell in every region except Latin Americaand the Caribbean after 1992–93 (figure 11.3). Pre-liminary estimates show that official development as-sistance rose again in 1999, by about 5 percent, thoughit is too soon to know whether this reflects more thanthe response to the Asian crisis and indicates a much-needed real and sustained reversal of the downwardtrend in the 1990s.

The decline has been costly for many countries. Al-though it has coincided with massive inflows of privatecapital to developing countries, very little of that capi-tal goes to the poorest countries. Net private capitalflows to low- and middle-income countries reached$268 billion in 1998 and now dwarf aid flows in somecountries. Overall, private flows to developing coun-tries surged during the 1990s, from 43 percent of totalresource flows in 1990 to 88 percent in 1997, just be-fore the East Asian financial crisis. However, inflows ofprivate capital have been concentrated in relatively fewcountries; a large number of countries receive little ornothing. In 1997, before the financial crisis, the top 15developing country recipients received 83 percent of

private capital flows to developing countries, leavingsome 140 developing countries and territories (withabout 1.7 billion people) to share the remainder. Almostentirely left out were the 61 low-income countries be-sides China and India.6 For example, all of Sub-SaharanAfrica received only 1.2 percent of flows to developingcountries in 1998. These are the countries that need aidmost, and they are hit hard by its decline.

There is no single reason for the decline. Donors ini-tially cited their fiscal deficits as a large part of the prob-lem. Yet even as these deficits declined (from 4.3 percentof GDP in 1993 to 1.3 percent in 1997), official devel-opment assistance continued to shrink, dropping 14 per-cent from 1996 to 1997.7 A more likely explanation isthat donors continue to view development cooperationthrough a strategic lens rather than a poverty lens, see-ing other uses for their money as strategically more im-portant. Historically, aid flows have been determinedmore by political and strategic interests than by povertyreduction goals.8

Perhaps more noteworthy is the decline in supportfrom the traditional proponents of official development

⁄

Figure 11.1

While donor countries’ economies grew after

1992, their development assistance shrank

GNP and official development assistance per capita in

DAC countries

U.S. dollars

Note: GNP is in 1995 U.S. dollars, official development assistance in 1998 U.S. dollars.Source: World Bank staff estimates based on OECD and World Bank data.

GNP

Official development assistance

0

25

50

75

100

1998199719961995199419931992

25,000

26,000

27,000

28,000

29,000

30,000

1987–88

1992–931997–98

Figure 11.2

With the exception of Europe and Central Asia,

the regional distribution of official development

assistance remained roughly constant . . .

Average annual share of net receipts of official

development assistance

Percent

Sub-Saharan

Africa

SouthAsia

MiddleEast and

NorthAfrica

LatinAmericaand the

Caribbean

Europeand

CentralAsia

East Asiaand

Pacific

Note: Shares do not sum to 100 percent, as some aid is not allocated by country or region.Source: OECD, DAC data.

0

5

10

15

20

25

30

35

assistance. The preeminence of geopolitical interests isnot new.9 But what is new is the falloff in countervail-ing support from advocates for development assistanceon humanitarian grounds. Many fell victim to “aid fa-tigue” and were far less vocal supporters in the 1990s thanbefore.

Not every country was affected by aid fatigue—indeed,aid flows increased from some countries—but its symp-toms were clearly evident. For example, in the UnitedStates a comprehensive poll found that an overwhelm-ing majority of the population favored foreign aid inprinciple—and that only 35 percent thought it shouldbe cut from current levels.10 Yet more than 80 percentof respondents believed that waste and corruption keptforeign aid from reaching the people who need it. Thiskind of public disillusionment may have made it harderfor donor governments to maintain foreign aid, let aloneincrease it. If aid is not working, the sentiment goes, themoney could be better spent elsewhere.

In contrast to the rise in aid fatigue in some places wasa major upsurge in support and activism around debt re-duction, most notably under the auspices of the Jubilee2000 movement of religious organizations and othercivil society groups. They rallied around the cause of

cutting debt for poor countries to support poverty re-duction and human development. So there is clearlycontinuing support for the principle of providing re-sources for improving the lives of poor people in the de-veloping world, but widespread questioning of thetraditional mechanisms for providing such resources.

Is aid working? Can it work better? What is the roleof debt reduction in concessional support? Developingcountries themselves will largely determine through theirown policies whether they achieve the international de-velopment goals. But aid and debt relief can providecrucial support. So finding out how to make these moreeffective—and then doing what it takes—remains vital.

In answering these questions, this chapter outlines avision for a better system of development cooperation,one based on new thinking and new practices. This vi-sion includes a reformed framework for country-focusedaid and debt relief for the poorest countries—under-pinned by a renewed emphasis on the policy and insti-tutional environment and the fundamental priority ofpoverty reduction. Donors would work in partnershipwith countries, directing aid and debt relief along the linesof a broad-based poverty reduction framework (as ad-vanced by many donors and laid out in this report), sup-porting countries that can put these resources to good usefor poor people.

Supporting good policies and institutions is important,but it is not enough. We learned in the 1990s that processis as important as policy in foreign aid and the manage-ment of unsustainable debt burdens. The way donors andrecipients interact strongly influences the effectiveness ofdevelopment cooperation. Relationships have tended tofollow the preferences of donor countries, leaving recip-ient countries with little sense of ownership of the aid-financed activities. Along with advancing a broad-basedpoverty reduction framework, this report emphasizeshow much local realities matter in development. That aidrelationships have too often failed to take local realitiesinto account, undermining ownership, is an importantflaw.

If development cooperation is to attack poverty ef-fectively and efficiently, donors will need to: ■ Pay more attention to local conditions and country

ownership.■ Deliver aid in ways that intrude less on government

functions, including greater use of sectorwide ap-proaches and a movement away from old forms of aidconditionality.

1992–93

1997–98

Figure 11.3

. . . while receipts fell after 1992–93 in all

regions but Latin America and the Caribbean

Average annual receipts of official development assistance

Billions of U.S. dollars

Sub-Saharan

Africa

SouthAsia

MiddleEast and

NorthAfrica

LatinAmericaand the

Caribbean

Europeand

CentralAsia

East Asiaand

Pacific

Source: OECD, DAC data.

0

5

10

15

20

⁄

■ Provide sustained support for policy and institutionalenvironments that are strongly conducive to povertyreduction, in preference to ones that are not.The chapter begins by exploring how these new ap-

proaches can make aid more effective. It then examinesthe issues associated with relieving the debt problem ofpoor countries.

Making aid more effective in reducingpoverty

Recent studies confirm what anecdotal evidence has longhinted: the experience of aid has been mixed.11 Early pre-dictions that aid would close the financing gap that pre-vented developing countries from moving ahead have notcome to pass. If all the aid that went to Zambia between1961 and 1994 had gone into productive investment, andif investment had been as important to growth as initiallypredicted, the country’s per capita income would havebeen more than $20,000 in 1994, not $600.12

And yet there have been many aid successes. The On-chocerciasis Control Program is but one example (seebox 10.1). Aid was important, in different periods, in EastAsia’s extraordinary success in poverty reduction over thepast few decades. The rapid progress in Vietnam in the1990s is another example. So aid can work. The challengefor the international community is to understand how tomake it work consistently—and then to do what it takes.

The key problems with aidAid’s difficulties in reducing poverty go deeper than thesway of geopolitical interests over development inter-ests, which has often directed aid to countries whosepolicies were not focused on reducing poverty. Aid hasbeen hindered by the frequent differences in donors’perspectives on development policies, even though the past50 years have been punctuated by times of relativelywide consensus on the best way to pursue development.13

Donor differences have played a key role in preventingaid from achieving full effectiveness. Donors have oftenfailed to coordinate their efforts, countries have not takenownership, and there has been heavy use of condition-ality both at the project level and economywide.

In the first two decades after World War II state-led in-dustrialization was generally seen as the best way to pur-sue development, a consensus undone in the 1970s byworld events, including the demise of the fixed exchangerate system and two oil shocks, which had devastating im-

pacts on developing countries. It was widely believed thatgovernment interference in the economy had preventeddeveloping countries from adjusting to these shocks. Sub-sequently, a new consensus began to form, eventually tobe known as the “Washington consensus” (see box 4.1 inchapter 4).14 To many, including staff at the World Bankand other multilateral financial institutions, fiscal prudence,free markets, and outward orientation had clearly demon-strated their superiority as the most efficient way forcountries to grow and develop.15

But it has become clear that simple strategies for de-velopment and poverty reduction are elusive. While mar-kets are a powerful force for poverty reduction, institutionsthat ensure that they operate smoothly and that their ben-efits reach poor people are important as well. As the 21stcentury begins, donors are coalescing around a develop-ment strategy that includes investing in people throughhealth and education services, promoting inclusive andequitable growth, supporting good governance, and pro-tecting the environment.16 This strategy also recognizesthe centrality of local conditions: that the most effectivedevelopment policies will vary by situation.

Despite this growing consensus on the broad devel-opment framework, agreement on the right policies in par-ticular conditions has tended to elude donors andrecipients. Donors come to development problems withtheir own mandates, histories, ideologies, and political re-alities and often do not see situations in the same way asother donors or the recipient countries. Even in healthand education, which all donors agree are essential, theright reforms are open to debate. As an analyst com-mented, there is “a bewildering multitude of national sys-tems and experiences, with varied (and hotly debated)advantages and disadvantages associated with each.”17 Sowhile the days of adhering strictly to either state-led ormarket-led solutions are over, between these extremes liea host of options, and the debate on them is far from over.

The lack of consensus on the broad outlines and thedetails of national and local policies and projects has re-duced the effectiveness of development assistance.18 Thiseffect is especially evident in problems of ownership,donor coordination, fungibility, and conditionality—the four main issues affecting aid in the 1990s.

Ownership. Because donors and recipients often dis-agree, donors have looked for ways to ensure that theirmoney is spent as they intend. They have run their ownprojects, required detailed reports from countries on pro-jects, and attached conditions—usually policy oriented—

to the use of funds. A major study on relations betweendonors and African recipients found that “in spite ofsome improvements, donors still tend to dominate theproject cycle and pay inadequate attention to the pref-erences of the government or project beneficiaries.”19

These efforts to ensure that aid is spent effectively, evi-dence now shows, have often had the opposite effect bydiminishing ownership by the recipient country.

Analyses show that ownership is a key ingredient of aideffectiveness.20 How strongly a country believes that a pro-ject or reform will bring benefits affects the effort put intothe activity, the domestic resources contributed, and the com-mitment to the activity after the donor has left—all sub-stantial determinants of success. To succeed, reforms andprojects must foster ownership by the people for whom thepolicy or project is ostensibly being implemented.

Donor coordination. When different donor prioritiesand project-related conditions (including donor-specificreporting and procurement requirements) are multi-plied many times over, they can create an unworkable en-vironment for a recipient government. Just the sheer numberof donors and donor projects can be challenging. At onepoint there were 405 donor-funded projects in the Mozam-bican Ministry of Health alone. In the early 1990s in Tan-zania there were 40 donors and more than 2,000 projects.In Ghana during the same period 64 different governmentor quasi-government institutions were receiving aid.21 Co-ordinating these efforts to support a coherent developmentstrategy—even at the sector level—is nearly impossible.

Fungibility. Studies show that aid funds allocated toa particular sector tend to free up for other purposesmoney that the government would otherwise have spentin that sector.22 This means that in funding specific pro-jects or sectors, donors may actually be helping to increasespending on sectors they do not want to finance, such asthe military. This has profound implications for devel-opment cooperation. Project-level evaluations will not re-flect the true impact of aid, since aid is likely to be freeingup resources for other activities.23

Even where resources are fungible, donor support canstill have some impact, from the design of certain poli-cies to institutional development. Moreover, in countrieshighly dependent on aid, donors as a group could leadto shifts in government resource allocations, because ofthe sheer size of flows. A potentially important part of thisis the preference of donors to support development bud-gets, which can lead to a net shift in resources out of therecurrent budget—not always a good thing for develop-

ment because of the importance of recurrent spending inmaintaining basic social and economic services.24

Conditionality. Donors know that even properly im-plemented projects will have limited impact in poor pol-icy environments.25 A well-built school will be useful onlyif money is budgeted annually for teachers, books, andsupplies—and if the economic environment enables chil-dren to go to school. The role of good policies and in-stitutions in ensuring sustainable results suggests thataid should flow more to countries with a good overall pol-icy environment and good policies for poverty reduction.But the relationship between good policies and aid flowshas not been strong.26

This finding would be understandable if aid werespurring policy reform by influencing countries to changetheir policies or by helping them do so. This has been theintention of many donors, and it is one reason (fungibilityis another) that many of them have reduced the share oftheir portfolio allocated to projects and increased theshare allocated to program and policy-based aid.27 Mostprogram and policy-based aid has been tied to the en-actment of certain policy reforms. But studies in the1990s showed little systematic relationship between con-ditionality and policy changes, though case studies do findpositive effects under some conditions, especially whereconditionality supports the hand of reforming groups.28

The dynamics between aid donors and recipients ex-plain why conditionality fails. Recipients do not see the con-ditions as binding, and most donors are reluctant to stopgiving aid when conditions are not met.29 As a result,compliance with conditions tends to be low, while the re-lease rate of loan tranches remains high.30 Thus aid has oftencontinued to flow despite the continuation of bad policies.

In addition to performing poorly in influencingpolicy reform, policy-related conditions, often combinedwith project-related conditions, severely burden de-veloping country administrators—a problem that hasbecome more pronounced as conditionality has ex-panded. Conditions on World Bank adjustment loans,having mushroomed in the 1980s, continued to growin the 1990s along with the expanding developmentagenda.31 As one recent assessment put it: “Althoughmuch has been added to the conditionality menu since1981, nothing has been taken off.”32 The time gov-ernment officials spent negotiating and monitoringthese conditions is time they could better have spentanalyzing development problems and designing de-velopment strategies. Ownership has been shown to be

⁄

central to the sustainability of both projects and pol-icy reform, and the fact that the delivery of aid weak-ens it is a fundamental flaw of current developmentcooperation mechanisms.

Solutions that accommodate differentperspectivesWhile the dominant forms of donor-recipient relationshave allowed donors to pursue their own priorities, theresult has generally been a fragmented system that un-dermines their efforts. The challenge in reforming in-ternational development cooperation is to accommodatedifferent perspectives on development without overbur-dening the recipient or undermining ownership.

Achieving global uniformity in development strategiesmight be one solution, but history shows that uniformityis undesirable. Development is determined to a great ex-tent by local conditions, including social institutions, so-cial capability, ethnic fragmentation, inequality, andgeography.33 In studies these variables significantly explainthe variation in growth rates over the past 30 years.34 Stud-ies also show that external shocks—and the ability to re-spond to them—can have as much effect on growth aspolicies do.35 The approach to designing developmentstrategies should therefore be flexible enough to adjustto both internal and external conditions.

This perspective began to take hold in the developmentcommunity in the late 1990s. Combined with new think-ing on aid effectiveness, it has prompted proposals to ad-dress the problems of aid. Three prominent themes areownership and partnership, less intrusive aid deliverymechanisms that focus on the overall policy and expen-diture framework, and selectivity. Together, they form theagenda for the international community to improve de-velopment cooperation in the coming decade.

Ownership and partnership. Recognizing the importanceof ownership and the problem of donor coordination, mostdonors have embraced partnership as a guiding principlein interactions among donors, governments, and citizensin developing countries.36 Most partnership frameworkshave two parts. The first is a partnership between the re-cipient government and its citizens, who share responsi-bility for developing their national development strategy.This strategy can take shape through a consultation processinvolving government, civil society, and the private sec-tor. The second is a partnership between the governmentand donors, with donors designing their assistance strate-gies to support the government’s strategy. In the newthinking the focus is on how to shape this external part-

nership, or contract, in a way that provides the incentivesfor country-driven, long-term poverty reduction strategieswhile also strengthening the internal partnerships neces-sary for social stability and economic development.

Consultations between governments and civil societyand between governments and donors have been carriedout in a number of countries piloting the World Bank’sComprehensive Development Framework, the EuropeanUnion’s partnership approach, and other such approaches.The consultations under the Comprehensive Develop-ment Framework have proved fruitful in several countries—such as Bolivia, the Dominican Republic, and Ghana—but have also highlighted the need for government com-mitment and for capacity as key ingredients of successfulconsultations (box 11.1).

This emerging approach to development cooperationhas been incorporated into the new initiative by theWorld Bank and International Monetary Fund (IMF) tolink their support of low-income countries to nationallydesigned poverty reduction strategies, working within theprinciples of the Comprehensive Development Frame-work (box 11.2). Concessional funds and debt relieffrom the World Bank and IMF will be linked to the goalsof poverty reduction strategies prepared by governmentsin consultation with civil society organizations, the pri-vate sector, and donors. Based on a good understandingof the poverty situation in the country, the strategies willidentify actions with the greatest expected impact and setup monitoring and evaluation processes. The goal is forthese strategies, described in poverty reduction strategypapers, to form the basis for assistance not only from theWorld Bank and IMF, but from other assistance agenciesas well.37 Similar initiatives are under way in the regionaldevelopment banks.

Less intrusive aid delivery mechanisms focusing on theoverall policy and expenditure environment. Donors haveused many means to influence recipient country policies.Old forms of policy conditionality have often had dis-appointing results, depending on country circumstancesand how the conditionality was used. Policy reviewprocesses also have had limited success. Public expendi-ture reviews, for example, have evaluated the level andcomposition of countries’ expenditures and identifiedways to improve expenditure policy and use donor fundsmore efficiently (see box 9.2 in chapter 9). But several stud-ies have found this type of intervention to be ineffectivein many cases, largely because recipient countries have notbeen closely involved in the reviews—and so have felt lit-tle inclination to comply with the findings.38

Box 11.1

Learning about the consultative process through the Comprehensive Development Framework

In 1999 the World Bank announced its Comprehensive Develop-ment Framework, a tool for improving country ownership and donorcoordination in development cooperation. The framework is basedon four principles: country ownership of the policy agenda, part-nership with all stakeholders, attention to social and structural con-cerns as well as macroeconomic and financial issues, and along-term, holistic approach built on national consultations.

The country develops its national strategy in consultation withcivil society and the private sector—and then, with donors, designsa matrix linking development goals and development actors. Theactivities of actors in support of each goal are listed in the matrix,revealing any gaps or overlaps.

The framework is being implemented in 13 countries, encour-aging wide consultation between governments and their citizens andenhancing partnerships with donors in the design of comprehen-sive national development strategies. But progress has been var-ied, reflecting different starting dates and country circumstances.

Bolivia is an early case. In late 1997 the new government em-barked on an analysis of the country’s development challenges andthe preparation of a national action plan to address them. A key partwas a national consultation with a wide range of representativesof civil society—NGOs, unions, religious organizations, oppositionparties, and academics—and the private sector to discuss devel-opment constraints and propose solutions. The results of this na-tional dialogue were presented to the government as input to thenational action plan.

All discussions with donors now take place in the context ofthe national action plan. At a consultative group meeting in April1998 donors pledged 45 percent more than they had in 1997.Donors have also been encouraged to formulate their strategiesin support of the national action plan. The World Bank recentlyredesigned its country assistance strategy to align it with the plan,choosing to support three of the plan’s four pillars. The govern-ment continues to lead donor coordination, chairing the consul-tative group meeting in Paris in 1999, where it presented itsversion of the Comprehensive Development Framework. It hasalso agreed with donors on intermediate indicators for monitoringoutcomes.

Other countries have not progressed as far. The difficulties ofsome highlight potential problem areas. For example, it is clear thatcountry ownership depends largely on national capacity. The coun-try must be able to hold broad consultations with all elements ofsociety and to conduct the complex analysis necessary to designnational strategies that balance macroeconomic and financial is-sues with social, structural, and institutional concerns. And, ofcourse, the country must be able to implement the strategy.

Without this ownership—and the country leadership from it—donor coordination will remain difficult. While there is someevidence that some donor countries are gradually aligning theirstrategies with those of recipient countries, stronger leadershipby the recipient government will be required to accelerateprogress.

Box 11.2

The new poverty reduction strategy initiative

The poverty reduction strategy initiative of the World Bank and In-ternational Monetary Fund seeks to link external support to do-mestically developed, results-based poverty strategies. It is alsointended to improve the effectiveness of World Bank and IMF re-lations (and those of other donors as well) with recipient countries.As important as the recipient country strategy is the process lead-ing up to it. A broad, participatory dialogue with representativesof civil society and the private sector is expected to:■ Help national authorities develop a better understanding of the

obstacles to poverty reduction and growth—and devise goodindicators of progress in poverty reduction.

■ Deepen a shared vision of desired poverty reduction goalsacross society.

■ Lead to formulation of priorities for public actions to achievethe desired poverty reduction outcomes.

■ Encourage the development of participatory processes forsetting poverty reduction goals and monitoring implementa-tion and progress.The results will be periodically reported in poverty reduction

strategy papers expected to reflect a broadly owned develop-ment strategy. The strategies will generally focus on three-year cycles, with annual progress reports in the intervening

years, all embedded in a long-term framework for poverty re-duction. While the actual form of the strategy will be decidedby the country—there is no single blueprint—most strategieswould likely include:■ Long-term goals for key poverty reduction targets, and the

macroeconomic, structural, and institutional framework forachieving them (see, for example, Uganda’s goals in box 1.7).

■ Mechanisms for monitoring and evaluating progress towardthe poverty reduction targets, linked to public actions.

■ A consistent policy and institutional framework that includes theunderpinnings for rapid, sustained growth and poverty reduc-tion (including macroeconomic policies, institutional reforms,sector strategies, and associated domestic and external fi-nancing needs).Donors can help by providing technical assistance in some

areas. Initial experience in Africa and Latin America indicates thatcountries are strong in laying out a poverty profile and a generalpoverty reduction strategy but weaker in preparing quantified tar-gets, costing the strategy, and evaluating tradeoffs under lim-ited resources. As in other aspects of development cooperation,the country should determine its own need for assistance—tomaintain ownership of this important process.

Source: Wolfensohn 1999; World Bank 1999d, 1999u.

Source: IMF and IDA 1999; World Bank and IMF 2000a.

⁄

Perhaps more surprising, donor compliance has beenweak as well. A recent evaluation found public expendi-ture reviews to have had little effect on either recipientcountry policies or donor lending practices.39 So donorsare searching for new mechanisms for strengthening pol-icy environments that encourage country ownershiprather than undermine it. They have begun, for exam-ple, to encourage countries to participate fully in thepublic expenditure review process, and they are experi-menting with new instruments.

One new instrument that has received much atten-tion is the sectorwide approach: the government designsan overall sector strategy, and donors sign on to fundthe sector, not individual projects. This resolves theproblem of donor coordination by eliminating the needfor it: all activity in the sector is conducted by the re-cipient country, using its own funds in addition tothose of donors. This instrument responds to a broaderpolicy environment while also ensuring ownership. Al-though the approach is too new to have a track record,some early experiences are promising (box 11.3).

Some proponents have suggested applying the prin-ciples of the sectorwide approach to all developmentcooperation (box 11.4). Others consider project-basedlending to be desirable and consistent with the newthinking on development cooperation for poverty re-duction. Project support can be effective for results-based sector development—if it falls within a sectorframework that systematically links investments and pol-icy and institutional development to poverty outcomes(and to intermediate indicators for tracking and in-terpreting progress). The choice of instrument will de-pend on the policy and institutional conditions ofparticular countries (or sectors within countries) andthe preferences of individual donors. But a premiumshould be placed on putting the country in charge andensuring that the mechanisms of aid delivery do notcompromise its ownership.

Selectivity. For aid to be most effective at reducingpoverty, it must be well targeted. If all aid money wereallocated on the basis of high poverty rates and reason-ably effective policies and institutions, a recent study es-timates, even today’s small aid flows could lift 19 millionpeople out of poverty each year—almost twice the esti-mated 10 million now being helped.40

Currently, about a third of aid goes to middle-incomecountries, whose average GNP per capita is roughly sixtimes that of low-income countries (figure 11.4). Whileonly a few major donors target more of their aid to mid-

dle-income countries (most donors target aid to thepoorer countries), that still means that global aid is notheavily targeted to areas where the incidence of povertyis greatest. Aid, and especially nonconcessional develop-ment flows, still has a role in reducing poverty in middle-income countries, when the policy environment is soundand the resources are well targeted.

In addition to targeting poverty, donors should allo-cate aid on the basis of the policy environment. Aid hasbeen shown to be effective in promoting growth andpoverty reduction in poor countries with sound eco-nomic policies and sound institutions—ineffective wherethese are lacking.41 Aid driven by political and strategicinterests rather than by the recipient country’s develop-ment policy environment is largely wasted from a povertyreduction perspective. Several instruments have been de-veloped to assess the policy and institutional environmentin recipient countries, generally covering macroeconomicmanagement, structural policies, policies for social in-clusion (poverty, gender), and public sector management(box 11.5).

Factoring in the level of poverty and the quality ofpolicies should make aid much more efficient in reducingpoverty, and there is evidence that donors began to dothis in the 1990s.42 In replenishing IDA in 1998, for

1997–98

1987–88

Figure 11.4

Aid does not go only to poor countries

Average annual net receipts of official development

assistance as a share of total

Percent

Note: Least developed countries are the 48 countries so defined by the United Nations on the basis of certain thresholds for income, economic diversification, and social development.Source: OECD, DAC 2000.

0

5

10

15

20

25

30

35

40

High-incomecountries

Middle-income

countries

Otherlow-income

countries

Leastdevelopedcountries

Box 11.4

The common pool for development cooperation

Seeing the potential of the sectorwide approach, some proposeextending the idea to the country level (Kanbur, Sandler, and Mor-rison 1999). Donors would cede complete control to the recipientcountry government—advancing their own perspective on de-velopment strategy through dialogue with the country and withone another rather than through specific programs and projects.Rather than fund their own projects, donors would give central bud-get support to countries with good development strategies (andthe capacity to implement them).

A country would first develop its own strategy, programs, andprojects in consultation with its people and with donors. It wouldthen present its plans to donors, which would put unrestrictedfinancing into a common pool of development assistance, to beused along with the government’s own resources to finance thedevelopment strategy. Earmarking would disappear. Donor mon-itoring and control of specific projects and programs would notbe permitted. And no conditions would be placed on donor aid.

How much donors give would depend on their assessmentsof the country’s policy environment, including how the countrycame to agreement on the strategy and how capable it is of im-plementing the strategy and monitoring progress. In this way thecommon pool approach would be a more rigorous form of condi-tionality, because donors would need to evaluate the overall pol-icy environment, direction, and capacity of countries. Theseassessments would be made known to the country and to otherdonors during the dialogue leading up to the financing decision.

This approach would entail many of the same challenges fac-ing the sectorwide approach, including the need for recipientcountries to have both the capacity to implement their strategy

and the confidence to follow through even if donors do not sup-port it. In addition, donors might resist common pools at the na-tional level because they would likely mean a reduction in donorstaff, since donor agencies would no longer be developing and mon-itoring projects or negotiating and monitoring conditions.

However, like the sectorwide approach, the common poolapproach would ensure full ownership by the country and elimi-nate donor coordination problems. It would also preserve two im-portant benefits of the current development cooperation approach: ■ The knowledge transferred in donor-implemented projects, an

important side effect of aid. A road building project, for example,might transfer knowledge of engineering or even project ac-counting to local workers. This transfer would not be lost in a com-mon pool arrangement. Recipient countries could still ensureknowledge transfer through their choice of companies and theterms of contracts.

■ The support that conditionality gives to reform factions in gov-ernments. Support for reform elements in a country is perhapsthe only effective part of the present system of conditionality.Donor-imposed conditions can strengthen the position of re-formers in national debates or serve as a “self-imposed” con-straint on government officials. The approach to conditionalityin a common pool arrangement would be far different, but itwould not sacrifice this benefit. Donors could strengthen the handof reformers by publicizing the criteria used to assess countrystrategies and adjusting the volume of their assistance. This wouldform the basis for a more open and honest relationship betweendonors and recipients and preserve the benefits of the currentconditionality while eliminating its problems.

Box 11.3

Sectorwide development cooperation

To address problems of ownership, donor coordination, and fun-gibility, donors are experimenting with pooling their resources tosupport sectorwide strategies designed and implemented by therecipient government. The country, in consultation with key stake-holders, designs a sector strategy and a budget framework extendingseveral years forward, and donors put their money into the cen-tral expenditure pool for the sector. The approach encouragescountry ownership of sector strategies and programs. It also linkssector expenditure with the overall macroeconomic framework. Andit ensures coordination of donor and recipient activities.

Some benefits of a sectorwide program are evident in the Zam-bian health sector. In 1994 the government presented its nationalhealth policy and strategy to donors and—to ensure equitable dis-tribution of services and coherent implementation of the strategy—asked them not to fund specific provinces or projects but to fundthe Ministry of Health centrally. Hesitant at first, donors began tocomply. An independent evaluation in 1997 found that “healthworkers are better motivated; clinics are functioning; funds areflowing to the districts; some modicum of decentralization is in place;[and] an important part of the private sector has become formallyinvolved.”

The approach ensures full ownership by the country and elim-inates problems of donor coordination. With the country having moreownership and control over what happens, the use of resources canbe much more efficient. But it also means great changes in donor-recipient relations and perhaps greater difficulties in implementa-tion. Several sectorwide programs have stumbled because of therecipient country’s inadequate institutional capacity. Lack of con-sistency with the macroeconomic program has been another prob-lem. And donors often have too many requirements and thus toomuch of a problem (or too little interest) in harmonizing them (Har-rold and associates 1995). Furthermore, these arrangements greatlydiminish donor control and monitoring of exactly how money is spent.

The changes required imply that gaining support for the ap-proach will be difficult. The recipient government has to be very con-fident, because strict adherence to a sectorwide approach meansdonors that do not participate in common implementation arrange-ments are not allowed to act in the sector (that is, they do not havetheir own projects). The result may be less donor funding for a sec-tor. Governments might therefore opt for less strict sectorwide pro-grams, choosing instead to allow donors to implement projects aslong as they fit into the overall sector strategy.

⁄

example, donors called for allocating funds on the basisof each country’s policy performance.43

How selectivity is applied will likely evolve as the in-ternational community continues to learn about the en-vironments in which aid is most effective.44 Some analystsstress that the level of poverty in a country is more im-portant to aid effectiveness than the policy environment,though both are crucial.45 Others show that externalshocks—such as declining terms of trade, volatility inexport prices, and even climate change—can impedecountries’ efforts in growth and poverty reduction (chap-ter 9).46 It has been argued that aid can make a larger dif-ference in these countries (and therefore be more effective)than in countries not experiencing shocks.47 Refining thecriteria for selectivity should continue. Adhering to the basicprinciple that aid should go where it is most effective inreducing poverty will be key if the international communityis to achieve the international development goals.

Implementation difficulties and practical stepsThese three components—ownership and partnership, aiddelivery mechanisms that are less intrusive, and selectivity—

provide the framework for substantially improved inter-national development cooperation. But progress towardthat vision will not be easy. Each component of improveddevelopment cooperation brings great challenges inimplementation.

For example, while almost everyone agrees that part-nership is a good idea, there is no consensus on how toimplement it.48 Some analysts note that ownership is rel-ative and that reaching consensus on strategies is essen-tially a political process, involving the same powerrelations that exclude poor people from discussions ordiscriminate against them (as seen in chapter 6).49 Oth-ers voice doubts that donors will really come to termswith the implications of ownership and partnership fortheir actions: that donors should interfere less in recip-ient country policymaking.50 Many donor practices—such as maintaining control over resource monitoring andtying aid to specific procurement requirements—run con-trary to the idea of partnership.51 The recipient coun-try’s capacity to design and implement developmentstrategies and its ability (and willingness) to hold broadconsultations with all elements of society also pose sig-nificant challenges.

Box 11.5

Assessing country policies and institutions

The World Bank has designed a measure of policy and institutionalsoundness—the Country Policy and Institutional Assessment, whichgives equal weight to 20 components that have evolved as the mea-sure has been refined. Each component is rated by country specialistson a scale of 1–6 using standard criteria. Although care is taken toensure that the ratings are comparable within and between re-gions, the scores include an irreducible element of judgment. Butwhen the measure has been included in regression analyses ofgrowth along with other commonly used policy variables, it has hadstatistical significance, while other policy measures have not. Itthus appears to be a good summary indicator of the overall policyenvironment for economic development. The 20 components:

Economic management

Management of inflation and the current accountFiscal policyManagement of external debtManagement and sustainability of the development program

Structural policies

Trade policy and foreign exchange regimeFinancial stability and depthBanking sector efficiency and resource mobilization

Competitive environment for the private sectorFactor and product marketsPolicies and institutions for environmental sustainability

Policies for social inclusion and equity

Equality of economic opportunityEquity of public resource useBuilding of human resourcesSafety netsPoverty monitoring and analysis

Public sector management and institutions

Property rights and rule-based governanceQuality of budgetary and financial managementEfficiency of revenue mobilizationEfficiency of public expendituresTransparency, accountability, and corruption in public services

Developing a consistent basis for rating economic and struc-tural policies has been relatively straightforward, but doing sofor social inclusion and public sector management has provedmore challenging. Work to refine the indicators and referencepoints continues.

Source: Collier and Dollar 2000; World Bank 1999h.

The combination of greater selectivity and a broader,less intrusive approach to delivering development assistancepresents its own challenges. Determining how much sup-port to give to a sector or national budget is difficult—andlikely to prove contentious. Some country expenditures maynot seem to fit into a “best” poverty reduction strategy, butdonors will have to evaluate the poverty reduction impactof the overall program, not the individual expenditures.

A more fundamental problem arises when a countrydoes not have an overall policy environment worth sup-porting, so that aid is largely ineffective. How shoulddonors proceed?

Most important, they must understand that policies aredriven primarily by the domestic political economy—andthat donors are simply not very effective in influencingthem.52 But donors can have some influence by tailoring theirinvolvement to a country’s commitment to reform. Until acountry commits seriously to reform, the best that donorscan do is to provide technical assistance and policy dia-logue, without large budget or balance of payments support(box 11.6). If donors pour large amounts of aid into poorpolicy environments, they are likely to sustain poor policieslonger. When the country finally commits to reform, evi-dence shows that finance should be increased as policies im-prove.53

In addition to this more nuanced approach to influ-encing policy reform, donors can address the challengesof the new development cooperation framework by tak-ing several other steps:■ Move the donor-recipient dialogue to the country and turn

its leadership over as well. Donor-recipient consultations—consultative groups or roundtables—have traditionallytaken place in donor countries, chaired by the WorldBank, the United Nations Development Programme, oranother donor institution. Meetings are now beginningto be held in recipient countries and chaired by their gov-ernments, to foster ownership.

■ Continue to experiment with sectorwide approaches.National capacity—and donor-recipient partner-ships—can be built up sector by sector. While manycountries will for some time not have the overall tech-nical capacity, accountability, and transparency tomonitor funds to the satisfaction of donors, thesemay be more advanced in some sectors than in oth-ers. The advanced sectors could be funded through thesectorwide approach as soon as possible, taking intoaccount the lessons from experience with this ap-proach.54 And donors should continue to improve their

own practices—for example, by harmonizing proce-dures and reporting requirements among themselves—so that they can contribute effectively to these new aidrelationships.

■ Strengthen monitoring and evaluation practices. Donors’systems of monitoring and assessing the impacts oftheir own projects have failed to focus on how poor peo-ple benefit.55 But doing this will be even more impor-tant (and challenging) when looking at a sectorwide ornationwide program. Donors should encourage localmonitoring by participants, to ensure ownership of theresults. Furthermore, donors tend to be weak in dis-seminating information and incorporating knowledge

Box 11.6

How aid can help in countries with a weak

policy environment

When a country has poor policies and no coherent politicalmovement to change the situation, aid can have a limited buteffective role, as Ghana, Uganda, and Vietnam all illustrate.In their prereform periods (before 1983 for Ghana, 1986 forUganda, and 1991 for Vietnam), these countries received verylittle aid, probably reflecting their governments’ political es-trangement from the West. But the aid was instrumental inlaying the foundation for policy reform.

For example, when Ghana was dealing with a macro-economic crisis in the early 1980s, its well-trained econo-mists found the policy dialogue with international financialinstitutions to be helpful in working out plans. A few yearslater, when Uganda’s leaders were trying to design new poli-cies, donors financed helpful study tours to Ghana. In 1991the United Nations Development Programme and WorldBank organized a meeting for Vietnamese leaders with eco-nomic ministers from Indonesia, the Republic of Korea, andMalaysia, who laid out some key policies that had workedfor them and also some of the detailed issues in stabiliza-tion, trade liberalization, foreign investment, and other eco-nomic policies.

In successful cases political leaders learn from othercountries and from their own mistakes. Low-key assistancecan help with this policy learning, which generally has to takeplace at a country’s own pace. Even in countries that do notreform for a long time, technical assistance can lay the foun-dation for policy learning. In Kenya, for example, donors aresupporting the Institute of Policy Analysis and Research tohelp develop local capacity in research and policy analysis.This kind of capacity building is not going to have large ben-efits as long as vested interests resist serious reform. Butit is an essential foundation if a political movement for changedevelops.

Source: Devarajan, Dollar, and Holmgren 2000; World Bank 1998b.

⁄

from evaluations.56 Feedback and learning are essentialto successful aid practices, and donors must ensurethat they happen effectively. As part of this, donors andrecipients should continue to strengthen their effortsagainst corruption, a major obstacle to economic per-formance that occasionally also affects donor agencies.

■ End tied aid. In 1998 almost a quarter of official devel-opment assistance was tied, meaning that the procure-ment contracts were limited to the donor country or agroup of countries. Driven by domestic political inter-ests, this practice goes against the very free-market prin-ciples that most donors are trying to encourage indeveloping countries and results in inefficient use of aid.It has been estimated that tying aid reduces its value by15–30 percent.57 The practice should be ended as quicklyas possible, and contracts should go to the best bids.58

■ Make technical assistance demand driven. Turning moreresponsibility over to recipient countries for designingnational development strategies and leading consulta-tion meetings will require rapid capacity development.Recipient countries will also need strong auditing andaccounting skills if donors are to relinquish monitor-ing and control of projects. But technical assistance, theobvious choice for building capacity, has a spottyrecord at best, particularly in countries where capac-ity is already weak. The main reason is that it has oftennot been demand driven—it has often been tied aidand designed to develop capacity only in donor-sup-ported activities.59 Instead, technical assistance shouldbe incorporated into a national strategy and expendi-ture plan, with the recipient government decidingwhat assistance it needs and who should provide it. Thisis likely to require initial support to countries on howto use the market for technical assistance.

■ Continue to learn about how to work effectively withNGOs. Relationships between donors and NGOs arecomplex, with much room for improvement.60 Gooddata on the extent and effectiveness of donor-NGOrelationships are scarce, but an estimated $5 billion inaid is now channeled through NGOs, either in sub-sidies to their activities or in contracts to implementdonor activities (figure 11.5). NGOs appear to be aneffective channel for aid when they are involved earlyin projects (at the design phase), when they are cho-sen for their proven capacity and experience, andwhen they are treated as partners rather than con-tractors.61 The long-term impact of NGO projects re-mains unknown, perhaps because so little money has

gone into funding their evaluation and monitoring ef-forts.62 Even with better monitoring, though, NGOprojects face the same problems of fungibility as donorprojects, and policy environments strongly influencetheir effectiveness. Donors and NGOs should continueto improve their working relationships, sharing bestpractices for making aid more effective in the long term.

■ Relieve more debt. Debt relief for the poorest countriesis essential for effective aid. Heavy debt burdens reduceincentives for policy reform, while debt negotiations andthe constant circulation of new aid money to service olddebt distract government officials from the needs of theircitizens. The next section turns to this issue.

Relieving the debt burden of poor countries

The most prominent issue in development cooperation atthe end of the 20th century and the beginning of the 21sthas been debt relief for the poorest countries. There has beena steady increase over two decades in the indebtedness ofa group of poor countries now referred to as the heavily

Figure 11.5

NGOs channeled some $10 billion to

developing countries in 1998, about half

of it from the official sector

Billions of U.S. dollars

Note: These data cover money used for development or relief purposes. The amount of official development assistance administered by NGOs is a rough estimate only, because many donors do not report this information.Source: OECD, DAC data.

0

1

2

3

4

5

6

Officialdevelopment

assistanceadministered

by NGOs

DAC contributions

to NGOs

NGOs‘ own grants to

developing countries

indebted poor countries (figure 11.6). Public attention hasbeen drawn to their plight in large part through the tire-less efforts of NGOs in developed and developing coun-tries, whose campaign for debt cancellation by 2000 hascaptured the world’s interest.63 At the 1999 annual meet-ings of the World Bank and IMF, member countries agreedon an enhanced plan for debt relief, an acknowledgmentof the detrimental effects of debt on country policy envi-ronments and overall expenditure frameworks (box 11.7).

The effects of heavy debt burdensMany heavily indebted poor countries spent as much asa fifth of their annual budgets on debt service in the 1990s,and some spent much more.64 Because this is often morethan the amount spent on social programs, debt servic-ing is viewed by many as a severe impediment to im-proving the lives of the world’s poor.

It has been argued, however, that debt servicing isnot really a problem because heavily indebted poorcountries receive more money from donor countriesthan they pay back. Actual debt service payments are

Figure 11.6

As per capita income in the heavily indebted

poor countries has gone down, debt has gone

up—and vice versa

Median per capita

GNP

1997 U.S. dollars

Note: The observed association between declining income and rising debt should not be viewed as implying that debt reduction will automatically result in higher incomes. Government policies are the key to growth and poverty reduction, and bad policies can lead to both higher debt and lower incomes.Source: Easterly 1999c.

Median present value of external

debt as a share of GDP

Percent

1997 1995 1990 1985 197940

50

60

70

80

250

300

350

Per capitaincome

External debt

Box 11.7

The Enhanced Heavily Indebted Poor Countries

Debt Relief Initiative

The Heavily Indebted Poor Countries (HIPC) Debt Relief Ini-tiative was announced in late 1996. Realizing that the initia-tive did not go far enough, leaders of the Group of Seven (G-7)countries endorsed an Enhanced HIPC Initiative at a summitin Cologne, Germany, in July 1999. The enhanced initiativewas approved by the full membership of the World Bank andInternational Monetary Fund in September 1999 as an inte-gral part of the new poverty reduction strategy initiative (seebox 11.2). The Enhanced HIPC Initiative changed the eligi-bility requirements for debt relief and the timing of relief.

Eligibility

To be eligible, a country must be very poor, have an unsus-tainable debt burden, and pursue good policies. ■ Poor is defined as both eligible for support under the IMF’s

Poverty Reduction and Growth Facility (the reformedand renamed Enhanced Structural Adjustment Facility, orESAF) and eligible only for concessional financing fromthe World Bank, through the International DevelopmentAssociation.

■ An unsustainable debt burden is defined as a stock of debtthat is more than 150 percent of exports in present valueterms after the full use of traditional debt relief mecha-nisms or (for countries with certain structural character-istics) a ratio of debt to government revenue of more than250 percent.

■ Good policies are interpreted to mean macroeconomic,structural, and social policies consistent with poverty re-duction and sustained growth.

These new eligibility criteria increase from 26 to 33 the num-ber of countries likely to qualify for relief.

Timing of relief

The Enhanced HIPC Initiative provides for the possibility ofinterim relief for countries after they pass the decision point,when the World Bank and IMF determine a country’s eligi-bility. A reduction in debt service payments is therefore pos-sible even before a country reaches the completion point,when the stock of debt is reduced. Under the earlier HIPCagreement the debt stock was reduced only after comple-tion of two full ESAF programs—a minimum of six years. Nowthe completion point can be moved up if the country’s per-formance is particularly good. Relief is intended to be front-loaded as much as possible.

Combined with traditional debt relief arrangements, theEnhanced HIPC Initiative is likely to cut by half the net pre-sent value of public debt for the 33 countries likely to qual-ify. As many as 20 countries may reach a decision point ondebt relief by the end of 2000, depending on progress in de-veloping their poverty reduction strategies and on how muchfinancing is available from donors.

Source: World Bank (www.worldbank.org/hipc).

⁄

almost always far less than scheduled payments, becausethe countries cannot make the full payments. Thedebts are serviced by rescheduling some loans and fi-nancing the rest through a combination of new loansand grants.65 Overall, while net transfers of noncon-cessional resources tend to be negative because new non-concessional borrowing is strongly discouraged, transfersof concessional resources tend to more than compen-sate (figure 11.7).

However, heavy debt burdens bring additional prob-lems that can affect a country’s growth performance andability to focus government action on social priorities.Debt service is financed largely by scarce domestic bud-getary resources and thus competes with domestic re-current spending, while concessional assistance goes tonew investment projects. This mix can mean resourcesfor new health centers and roads but not for nurses ormaintenance. In addition, many grants go to donor-managed activities that are not included in the budget.These are subject to all the problems of ownership anddonor coordination discussed above and can contributeto the further institutional weakening of an already weak-ened, insolvent state.66 And debt negotiations and mon-

itoring take up much of the already stretched time andcapacity of government officials.

These resource inflows can also be unstable, makingit difficult for governments to manage their spending andmaintain sound fiscal policies.67 Furthermore, if resourceflows are positive because countries have to rely on con-tinuous recheduling and on grants and concessional lend-ing, their access to private capital flows will remain verylow. And where debts are not serviced in full, countries’debt stocks continue to grow, creating a potential disin-centive to investment, since investors may fear that fu-ture profits will be affected by debt-related macroeconomicproblems or higher taxes to service debt.68

Debt is therefore as much a problem of how gross flowsand debt management affect ownership, policy, and ca-pacity as it is a problem of net flows. In this, it sharesmany of the problems that have diminished the effec-tiveness of aid. Debt relief can play an important role hereby reducing the burden on recurrent budgets and allowinggovernment officials to focus on sound spending strate-gies rather than continual renegotiation of debt. And itcan be particularly crucial for countries emerging fromcivil conflict and war.

Figure 11.7

Concessional transfers largely compensate for negative net transfers of nonconcessional resources

Net transfers to heavily indebted poor countries by creditor, 1988–97

Billions of U.S. dollars

Source: Easterly 1999c.

–10

–5

0

5

10

15

20

25

International Bank for Reconstructionand Development

Bilateralnonconcessional

PrivateInternational Monetary Fund

Othermultilateral

nonconcessional

Othermultilateral

concessional

Bilateralconcessional

InternationalDevelopmentAssociation

There is also some evidence that high debt serviceobligations (including those to international financialinstitutions) tend to weaken the link between conces-sional flows and the quality of the policy and institu-tional framework—and so the effectiveness in reducingpoverty. This could be because donors try to avoid de-faults on loans, and as countries become more in-debted, donors give new loans to cover the old ones.(Between 1989 and 1997 debt relief for the 41 heav-ily indebted poor countries totaled $33 billion andnew borrowing $41 billion.)69 Not only does this com-promise the ability of donors to target aid to where itwill be most effective, but it may also deter reform incountries with poor policies, because they have lessincentive to reform if they can expect relief and resourcesanyway.70

Debt relief can ease all these problems by reducingthe gross flows and, if structured correctly, encouraginga structure of new inflows that is more effective forpoverty reduction.

An improved initiative for debt reliefTo be effective, debt relief needs to be delivered in waysthat encourage country ownership, using instrumentsthat provide incentives to use the resources for povertyreduction. This is the same issue as for traditional aid flows,but in the context of a one-time decision to reduce debt.How much impact debt relief has on net transfers to acountry depends, of course, on what happens to gross aidflows—on whether the resources for debt relief are ad-ditional or not. But even if the resources are not entirelyadditional, debt relief can ease policy and budgetary con-straints for the recipient country, since it frees up resourcesfrom the recurrent budget. What will guarantee thatthese resources are used for poverty reduction? There aretwo related challenges:■ Linking resources from debt relief to results in

poverty reduction.■ Strengthening accountability in the use of public re-

sources, to minimize diversion to other uses (especiallythrough corruption).The lessons from the past—including those from the

experience with aid outlined above—indicate that bothare best tackled through their links to the overall policyand institutional environment, especially for public re-source use. Experience also shows that debt relief alonewill not improve policies. Twenty years of gradually in-creasing debt relief have not improved policies in heav-

ily indebted poor countries.71 That is why the principleis to grant debt relief on the basis of reputation—an es-tablished track record in using resources effectively forpoverty reduction.

The design of the Enhanced Heavily Indebted PoorCountries (HIPC) Debt Relief Initiative incorporatesthese lessons. Debt relief will be granted to eligible coun-tries with a viable and comprehensive poverty reductionstrategy and a framework for linking public actions tomonitorable results in poverty reduction. The strategy isto be defined through a participatory process involvinggovernment, the private sector, and civil society. Theparticipatory process is important for the design of thestrategy—and to help ensure good use of external (andinternal) resources. Debt relief will be integrated with othersources of external finance in the country’s overall bud-getary framework for poverty reduction, rather thanbeing earmarked for certain expenditures. The goal of theEnhanced HIPC Initiative is to contribute directly topoverty reduction and to ensure that countries that re-ceive debt relief do not have policies that will lead themdeeply into debt again.

In May 2000 Uganda became the first country to re-ceive debt relief under the Enhanced HIPC Initiative(box 11.8). The relief was based on several years ofprogress in the participatory formulation of its povertyreduction strategy, results in key areas (getting childreninto school, reducing income poverty through agricul-tural and aggregate growth), and mechanisms to helpincrease accountability for public funds and reduceleakages.

The cost of the Enhanced HIPC Initiative is esti-mated at $28 billion. If the debt relief is to be addi-tional, financing must come from outside the normalaid and concessional lending budgets of donor insti-tutions. Under current plans the cost will be financedroughly equally by bilateral and multilateral credi-tors. Although many donors have endorsed the En-hanced HIPC Initiative and made politicalcommitments for funding, the mobilization of re-sources has been slow, and some donors have not yetcommitted to the initiative. Because a key principle un-derlying the initiative is that debt relief should be co-ordinated among all creditors, with broad and equitableparticipation, this lagging of resources and commit-ments seriously endangers the initiative. Donors needto give high priority to securing sufficient fundingfor the Enhanced HIPC Initiative.

⁄

• • •

Many questions remain about the implementationof debt relief and of the new development cooperationframework advanced in this chapter. Despite the fi-nancing difficulties of the Enhanced HIPC Initiative, someobservers call for even deeper and faster debt relief, ar-guing that the debt deemed “sustainable” under the En-hanced HIPC Initiative is still too burdensome.72 Howto move quickly to relieve debt while still allowing enoughtime to build country ownership of the poverty reduc-tion strategy is another concern. Some countries wonderabout their capacity to prepare their own poverty assess-ments and poverty reduction strategies. Others questionwhether donors can support the formulation and im-plementation of poverty reduction strategies withoutundermining country ownership. Questions also remainabout the participatory process—how best to consultwith poor people, how to fit consultative processes intothe context of national political processes, and how to de-velop effective feedback and monitoring systems. Andcountries wonder how well donors will be able to realign

their procedures and interventions along the lines laid outin their poverty reduction strategies.73 All these issues re-flect the state of international development cooperationat the turn of the 21st century. There is profound, on-going change in the way developing and developed coun-tries work together to fight poverty.

While many issues remain, the right direction for theinternational community is clear. Country-focused as-sistance should incorporate a greater emphasis on part-nership between donors and developing countries. Itshould apply less intrusive mechanisms of aid delivery thatfocus on the overall policy and expenditure environ-ment. And it should exercise greater selectivity in allocatingaid where it will be most effective. More aid and debt re-lief need to be available to countries with effective povertyreduction programs. Donor evaluations of these pro-grams must be informed by an awareness of the condi-tions each country faces and by the new approach topoverty reduction presented in this report. And to relievethe burden of the heavily indebted poor countries, donorcountries should finance the Enhanced HIPC Initiativewith money additional to their aid budgets.

Box 11.8

How debt relief fits into a poverty reduction strategy: Uganda’s Poverty Action Fund

Fundamental in the fight against poverty is improving the overallallocation of resources, including those from debt relief, throughmore poverty-oriented and transparent budgets. There are manyways of achieving this end, and in Uganda a special fund to usethe savings from debt relief is proving useful.

The government chose to create the Poverty Action Fund asa conduit for the savings from debt relief under the HIPC Initia-tive (about $37 million a year; the Enhanced HIPC Initiative is ex-pected to double this amount). The fund has been earmarked forpriorities of the poverty eradication action plan adopted in 1997to address poverty and social conditions. The plan emphasizes main-taining macroeconomic stability while increasing the incomes andthe quality of life of poor people by developing rural infrastructure,promoting small businesses and microenterprises, creating jobs,and improving health services and education. The Poverty ActionFund focuses on schools, rural feeder roads, agricultural extension,

and district-level water and sanitation. Specific outcome targetshave been identified, such as the construction of 1,000 additionalclassrooms to support the primary education program.

Two crucial features of the Poverty Action Fund are its integrationinto the overall budget and the Ugandan government’s effort to cre-ate a transparent and accountable structure of management. Re-ports on financial allocations are released at quarterly meetingsattended by donors and NGOs. The Inspector General’s officemonitors the use of funds at the district and national levels. Thisself-imposed conditionality reflects the government’s strong com-mitment to tackling corruption. But it is also an attempt to addresscreditor concerns about the capacity of a debtor country to link debtrelief to poverty reduction. Several measures have been proposedfor improving monitoring, ranging from including district-level offi-cials in the quarterly meetings to having local NGOs do community-based monitoring of the poverty fund’s spending.

Source: UNICEF and Oxfam International 1999.