Does Aid Work? · 2007. 10. 2. · Finn Tarp reviews the evidence from aid effectiveness studies...

Poverty

International Poverty Centre October 2007

Does Aid Work?— for the MDGs

http://www.undp-povertycentre.org

2 International Poverty Centre

F R O M T H EE D I T O R

Dag Ehrenpreis

Poverty in Focus is a regular publication of theInternational Poverty Centre (IPC). Its purposeis to present the results of research on povertyand inequality in the developing world.

EditorDag Ehrenpreis

International Advisory Board

Desktop PublisherRoberto Astorino

Front page: A crate is unloaded from acontainer ship on the collapsed quayside at theFreeport of Monrovia, Liberia. The quay wasdamaged when a boat carrying foreign aidsupplies crashed into it. The Freeport washeavily contested throughout both of Liberia’scivil wars. Photo by Tim A. Hetherington, Panos.

Editor’s note: IPC acknowledges the roots ofthe title of this issue as well as substantialinspiration from the seminal book Does AidWork? by Professor Robert Cassen, LondonSchool of Economics, Oxford U Press 1986 and1994. We are grateful to all the authors forgenerously contributing their intellectualproducts without any monetary remuneration.

IPC is a joint project between the United NationsDevelopment Programme and Brazil to promoteSouth-South Cooperation on applied povertyresearch. It specialises in analysing poverty andinequality and offering research-based policyrecommendations on how to reduce them.IPC is directly linked to the Poverty Group of theBureau for Development Policy, UNDP andthe Brazilian Government’s Institute for AppliedEconomic Research (IPEA).

IPC Director (acting)Terry McKinley

International Poverty CentreSBS – Ed. BNDES, 10º andar70076-900 Brasilia DF Brazil

[email protected]

The views expressed in IPC publications arethose of the authors and do not necessarilyreflect the views of IPC, IPEA or UNDP.

Oscar Altimir, CEPAL, Santiago de ChileGiovanni A. Cornia, Università di FirenzeNora Lustig, Universidad Iberoamericana, MexicoGita Sen, Indian Institute of Management, BangaloreAnna Tibaijuka, UN Habitat, NairobiPeter Townsend, London School of EconomicsPhilippe van Parijs, Université de Louvain

I n 2005, the G8 leaders promised to double aid to Africa and the UN WorldSummit to increase total official development assistance (ODA) by around $ 50 bn.a year by 2010 to reach at least 0.5 per cent of donor countries’gross national product(GNP). The Summit recognised that such increases in ODA were required for achievingthe Millennium Development Goals (MDGs).

Aid volume targeting is not new; even in 1970 the UN set the ODA target of 0.7 per cent ofthe GNP of each economically advanced country, so far met by only five donor countries.

The announcements of sharp increases in aid to developing countries have not beenreceived only with cheers all around. An intense debate is taking place on whether thisis the right road to take, considering the perceived macroeconomic risks and the largeamounts of aid that have already been provided during several decades, while massivepoverty persists in developing countries. The aid critique and the signs of political‘aid fatigue’ have intensified in recent years; many point to free trade and privateinvestment as better options.

Yet, the decades of ODA have seen the largest poverty reduction in the history ofmankind. The evidence indicates mostly that aid has contributed significantly to this,both via its impact on economic growth and through more direct interventions forhuman development (see page 10).

In this issue of Poverty in Focus, various features of the current international aid systemare discussed critically and constructively, with references to recent research literatureon aid effectiveness and sharing of important and policy-relevant results.

Roger C. Riddell summarises his new book explaining why the current aid system isno longer fit for purpose and needs a radical overhaul. He provides a sketch of analternative as a basis for discussion.

Stephen Browne aims his critique at the traditional volume targeting of aid and at itssupply-driven mode, proposing a shift of focus to human development at country level.

Nancy Birdsall focuses on the impact of aid on the middle class, including skilledworkers. She advocates reforms of aid to reduce its risk of undermining weakinstitutions in aid-dependent Africa.

Finn Tarp reviews the evidence from aid effectiveness studies and finds the single mostcommon result to be that aid has had a positive impact on per capita income growth.

Edward Anderson discusses how different principles for allocating aid across countries affectglobal MDG achievement, involving a trade-off between equity and effective use of aid.

Patrick Guillaumont argues that aid is most effective as volatility insurance in thepoorest and most vulnerable countries, where it helps pro-poor growth bystabilising the economy.

Rainer Thiele et al. examine the large variations across donors in the sector targeting ofaid, especially in the share of social sectors that are key to achieving most of the MDGs.

David Goldsbrough and Ben Elberger find that IMF conditionality has unduly constrained aidspending in poor countries in favour of using aid for reducing debt rather than poverty.

John Serieux stresses that such a response to the fear of Dutch disease is unjustified,since many countries have shown that aid surges can be managed and used for itsintended purposes.

Terry McKinley shows that such fears have meant that only about a quarter of disbursedaid to Africa was actually spent as intended; the lion’s share went into reserves anddebt buy-backs.

Pierre Jacquet analyses the aid grants vs. loans issue in the light of the resolution of thedebt crisis of poor countries, supporting the use of modernised models of ODA loans.

Jan Cedergren reports on the response by development cooperation partners to the manychallenges of aid effectiveness and on the intensive reform process and its monitoring.

May these articles contribute to the ongoing analytical and policy debate on aideffectiveness and thus to the changes in aid modalities and procedures that are mostlikely to enhance advancement towards the MDGs.

http://www.undp-povertycentre.org

Poverty In Focus October 2007 3

If the international community isserious about wanting to improve theeffectiveness of aid, it would focus itsattention principally on addressing thecentral systemic problems whichcontinue to impede the greater impactof official development assistance (ODA).Today’s system of raising, allocating anddeploying official aid remains effectivelythe same as that created more than 50years ago. Especially in relation to thepoorest countries, this system is nolonger fit for purpose. It needs to beradically overhauled.

International conferences at whichdonors jointly pledge to increase theiraid—and international organisations’statistics on the total amount of aidgiven and the share of ODA to grossnational income (GNI)—create theimpression of common purpose, jointaction and shared goals. The reality issharply different as the followingcharacteristics of the aid ‘system’strikingly show.

Aid is provided by individual donors onan entirely voluntary basis. Each donorchooses and determines itself how muchaid it will give. Repeated pledges atinternational conferences and summitsto increase official aid are not bindingon donor countries. No sanctions areimposed on countries which fail tohonour their promises, only mild criticismin the peer reviews of OECD’sDevelopment Assistance Committee(DAC), which they are at liberty to ignore.

There is no clear and predictable linkbetween the overall amounts of aidprovided and the aggregate aid needs,though the crude and partialassessments made suggest a huge gap.Hence, the potential impact of aid isreduced because insufficient aid isprovided. Indeed, there is no necessaryrelationship between the aid requirements

of individual poor countries andthe amount of aid each receives.

To qualify as ODA, official aid fundsmust have the promotion of economicdevelopment and welfare as their mainobjective. However, ODA continues to beinfluenced by the commercial, strategicand short-term political interests ofdonors, reducing the share of aid to thepoorest countries and raising its costs.

A major consequence is the extremevolatility in the flow of aid funds toparticular countries. Recipients do notknow whether the aid they will receivethis year will continue to be providedfive, three or even one year hence, andhistorical experience confirms that theamounts provided vary markedly fromyear to year. Aid volatility makesrecipients extremely reluctant even tospend the aid they receive, particularlyon recurrent costs, reducing still furtherits potential impact.

Nearly seven years after the UNMillennium Summit there has yetto be a country case where aid is beingsignificantly scaled up to support amedium-term programme to reach theMillennium Development Goals (MDGs),as the World Bank pointed out in theGlobal Monitoring Report 2007.

Aid effectiveness is severely reducedby the growing complexity of donor-recipient relationships. Today, there areover 200 official donor agencies, morethan double the number 40 years ago,and the numbers continue to rise.At least 30 recipient countries mustdeal with more than 40 donors.

Why does official aid continue to beprovided in ways which seriously impedeits potential impact? The answer lies inhistory. For more than 50 years, mostaid-giving has been determined by the

by Roger C. Riddell,Oxford Policy Managementand The Policy Practice

Effective Aid RequiresNew Structures

The international aid systemis no longer fit for purpose,especially in relation to thepoorest countries. It needsto be radically overhauled.

Aid is still influenced by thecommercial, strategic andpolitical interests of donors,reducing the share of aid tothe poorest countries andraising its costs.

Aid is needed most inprecisely those countrieswhich are least able touse it efficiently.

Concrete proposals arepresented to focusattention on the keyissues and encouragefurther discussionof a different aid system.


decisions of individual donors. Inrecent years, this approach has beencomplemented by what could be termedthe international co-operative approach,led by the United Nations and the OECD/DAC, both concerned at the large gapbetween aid’s impact and its potential.However, in relation to the list ofproblems just outlined, both the scopeand pace of reform of the aid system havebeen extremely limited.

In its 2005 Human Development Report,UNDP argued that “fixing the internationalaid system is one of the most urgentpriorities facing governments… “. Themain focus of attention needs to be theworld’s poorest countries, those whoseneed for aid is the greatest. The followingcomprise the core building blocks of anaid system for the poorest countries intune with the core principles of ourcontemporary world:

a) The acceptance by all nation states, butparticularly by the wealthier nations, ofthe obligation to provide assistanceindependent of their own commercialstrategic and short-term political interests,and channelled to those countries unable,on their own, to ensure the fundamentalrights, basic needs and core freedoms oftheir own citizens.

b) Compulsory contributions by the richcountries, provided on the basis of theirrelative wealth, and pooled into an aidfund of a size sufficient to meet the totalaid needs of the poorest countries, basedon the aggregate assessments of theirindividual aid requirements. Aid fundsfrom such a common pool would bechannelled to each of the poorestcountries to help meet their needs andaddress the financial, skills and otherrequirements upon which the assessmentof aid needs was based.

The main challenge in creating such asystem lies less in agreeing its broadprinciples and core building parametersand far more in debating and agreeingprecisely how a workable system mightbe established, and to gain the supportof all nation states, both donor andrecipient governments. Here, thegreatest hurdle is unlikely to be theswitch from voluntary to compulsorydonations or the distancing of

aid-to-the-poorest from the commercialand short-term political interests ofthe donors. Rather, it is likely to hingeon the confidence of the main donorgovernments that the recipient countriesaccessing aid funds from the commonpool will make good use of the aid fundsprovided. This issue goes to the heart ofthe following twin paradoxes of aid:

Aid is needed most in precisely thosecountries which are least able to use itwell. The less unfavourable the contextinto which aid funds are channelled, thehigher the likelihood that donors willwish to apply a range of conditions to tryto ensure that ‘their’ aid funds are wellspent. However, the impact of aid tendsto be linked directly to the commitmentand capability of the recipient to use itwell, and the greater the degree andintensity of conditionality applied by thedonors, the more recipients are likely tofeel that they are not in control, so theywill be less committed to ensuring thefunds are used as effectively as possible.

Many wise women and men havegrappled with these issues over the morethan five decades that official aid hasbeen provided. Based largely on theirwork, the following clusters of inter-related proposals provide the basic ideasfor how current aid relationships mightbe altered, most especially to addressthe core paradoxes of aid and thetensions surrounding aid conditionalities.These proposals are presented in orderto focus attention on the key issues,and encourage further discussion ofa different aid system.

A new International Development AidFund (IDAF) for the poorest countriesshould be established. The monies to beraised for the operation of the Fundwould equal the total amount of (official)development aid required by eachqualifying poor country. The resourcesrequired would be funded by compulsorycontributions made by each of the world’swealthiest countries, with the amountscontributed by each determined by theirrelative wealth.

IDAF funds would be earmarked for, butnot initially allocated to each qualifyingrecipient country in relation to their aidrequirements. In order, particularly, tosatisfy donor governments that the fundswere being spent as well as possible, anew International Aid Office (IAO) shouldbe established, whose roles andresponsibilities would include overseeingand ensuring the effective functioning ofIDAF. While, ideally, the establishmentof the IAO would be based on a ‘balanced’agreement between the governments ofdonor and poor countries, what is crucialis that it be created with the full supportof donor governments, otherwise—as discussed below—donors would beunlikely to support the new mechanismsfor overseeing the spending of aid funds.

On the basis of consultations with eachrecipient country and with key donors,the IAO would determine which of twopossible ways these earmarked fundswould be disbursed: (i) if the IAO wassatisfied that a qualifying poor countrygovernment had the commitment,competence and capacity to use the

Is aid effective and focused on poverty reduction?

The 63 poorest countries of the world receive less than half of total ODA.

Afghanistan, Iraq and Pakistan received less than two per cent of all ODA in 1999; thisshare rose to a quarter of all ODA in 2005, a thirty-fold increase.

Aid tying is estimated to add over 20 per cent to the costs of providing aid, amountingtoday to some $7 bn.

Most aid is still managed by donors: less than 40 per cent is disbursed through recipientgovernment channels; half of all multilateral aid is earmarked for particular uses; and onlya third of aid provided by UN aid agencies is aligned to national priorities.

In 2005, the number of separate aid transaction recorded , in aggregate, by all official aidagencies was estimated at 60,000—three times the number less than ten years ago—while the average size of each transaction has progressively fallen.


aid funds effectively, it would allocatethe earmarked funds directly to therecipient government; (ii) if the IAOhad doubts about the commitment,competence or capacity of the recipientgovernment to use the aid fundseffectively, a National Aid ImplementationAgency (NAIA) would be establishedfor the purpose of allocating andoverseeing the spending of aid fundsearmarked for the country by the IDAF.

As its name suggests, the NAIA would benational—recipient country-based—andnot a donor aid agency. Its operationswould be overseen by a group of eminentand competent recipient countrynationals, drawn from a cross-section ofsociety, including civil society, and staffedby a mix of nationals and internationalstaff as required to ensure aid fundsare spent well, in an accountable andtransparent way. It is likely that in manypoor countries and fragile states, fundswould be channelled into training andtechnical support as needed.

Why would a poor country governmentagree to ‘its’ aid funds being overseenby the NAIA? For two principle reasons:firstly, because otherwise it would notreceive official aid funds; and secondly,because a significant share of the aidwhich passes through the NAIA is highlylikely to end up in state-based projectsand programmes.

The IAO would play the key rolein assessing whether the NAIA wascompetent to allocate and monitorthe use of the funds provided,and, where necessary, it would drawon international expertise, usually ona temporary basis, to provide supportto NAIA—the channelling of aid intoa common pool and IAO approval forthe NAIA removing the requirementnow made by donors to apply their ownconditions to the aid they individuallyprovide. The IAO would also beresponsible for ensuring that reviewsof the status of each recipient countrytake place, and would be open torepresentations from parliamentary,civil society or other groups fromrecipient countries, and from donors,who have substantive concerns that aidfunds are not being used as effectivelyas they might.

If individual donors wanted to provideaid to recipient countries additional tothat provided through the auspices of theIDAF, they would be at liberty to do so,provided recipients were willingto accept such aid, and any conditionsattached to its provision. It would beup to the relevant donors and recipientsthemselves, and no one else, todetermine how to make use of theaid funds offered.

Under these proposals, both recipientcountry governments and, whereapplicable, the different NAIAsestablished in aid recipient countrieswould be encouraged to make use ofand draw directly upon the skills andexperiences currently residing inexisting aid agencies, internationalfinancial institutions and aidconsultancy firms to assist in ensuringthat aid funds are effectively used.Any supportive technical assistanceprovided would normally be contractedon the basis of a transparent tenderingprocess. Both recipient governmentsand NAIAs would be at liberty tocontract either one, or a number ofdonor agencies to manage and overseethe use of all, or part of a recipientcountry’s aid programme, provided theagency was willing to sign such aidcontracts and accept the conditionslaid down by the recipient countrygovernment or NAIA. The figureillustrates what an aid structure basedon these initial ideas might begin to

look like in countries where a NAIA wasdeemed necessary.

These proposals should be viewed moreas a vision of what a different aid systemmight begin to look like, and not as ablueprint for immediate implementation.Their principal purpose is to stimulatediscussion of what might be. The mostimportant proposal is for nation statesto begin the serious discussion of howthe aid system could change, andno longer continue to park suchdiscussions in their ‘far too difficult’pending trays.

Repeatedly over the past decadesstatesmen and women and internationalcommissions and scholars have called fora radical reform to the current aid systemreplacing it with one which is fit forpurpose. For too long, politicians haveavoided facing head-on the challengesinvolved in trying to change the aidsystem. They need to be encouraged, orshamed, into trying, if they really believethat extreme poverty is the most seriousform of human rights violations in theworld today and requires urgentattention. For, as all donors agree, aidcan contribute to making a difference—if it is provided in ways which seek tomaximise its impact.

R. Riddell: Does Foreign Aid Really Work?Oxford: Oxford University Press 2007.http://www.oup.com/uk/catalogue/?ci=9780199295654

http://www.oup.com/uk/catalogue/?ci=9780199295654


Traditional targeting of aidlevels follows a false scentin development terms.

Supply-driven aid is worthless, driven by the instinctsof rich country politiciansand donor self-interest.

Aid has more meaning andlegitimacy when it is focusedon human developmentoutcomes at country level.

Donors should reducethe harmful effects ofconditioned aid anddistorted trade, whilecreating a more propitiousglobal environment.

Two years ago at Gleneagles, the G8countries promised to double their aid toAfrica. Since then, they have written offa substantial part of the external debt tothe largest and oil-richest country, Nigeria.But new aid to the continent has stayedflat. In 2006, while Europe increased its aid,the two largest G8 economies, USA andJapan, reduced theirs. Africa, quite rightly,commands the growing attention ofdonors. But aid amount targets, bothfor Africa and globally, are often missed.Does that matter?

This article makes four propositions:(i) traditional aid amount targeting isfollowing a false scent in developmentterms; (ii) supply-driven aid hasquestionable value; (iii) aid should bemore concerned with genuine country-based development goals; and (iv) richcountries should use aid as a means offacilitation, not as patronage.

Targeting aid amounts is nothing new. In1970, the UN set the target of 0.7 per centof rich countries’ Gross NationaI Product(GNP) for Official Development Assistance(ODA). Since then a growing number ofdonor countries have stated theirintention to reach it1. The main purposefor setting such targets for aid is to createand sustain a momentum for ODA. Whilemost donors haven’t met the target, manyhave agreed that they should increaseassistance to the poor countries. For thepoliticians of the rich countries and theirconstituents, therefore, aid volumetargeting plays a useful role in remindinggovernments of their obligations. Thetwo largest donors, however—USA andJapan—are exceptions.

But apart from providing momentum,what does such targeting really mean?Is it based on a realistic estimation ofthe resources actually required fordevelopment? When it was first set,

development—often equated witheconomic growth—was supposedto depend mainly on capital, whichdeveloping countries lacked and thedeveloped countries could supply. Today,there is no such developmental basis, forvarious reasons.

First, the resources represented by thetarget in 1970 would be a very different—and a much smaller—percentagetoday. Even if we include private externalfinance, as in the previous 1 per cent target,it was exceeded many years ago; totalprivate foreign direct investment flowsfrom North to South are several timesgreater than ODA.

Secondly, even if we still imagined thatdevelopment depended critically onexternal resource transfers, there are manysources besides aid: export proceeds andremittances from abroad, for example,which have grown appreciably.

Third, experience long ago confoundedthe facile assumption that moreresources resulted automatically in moredevelopment. If that were true, then justbased on oil revenues, Angola and Nigeriawould already be advanced countries,whereas they rank 159th and 161st out of177 countries on the human developmentscale. As the Center for GlobalDevelopment puts it (Working Paper 68):“the 0.7 per cent target was calculatedusing a series of assumptions that are nolonger true, and justified by a model thatis no longer considered credible.”

A fourth reason for questioning thecorrelation between aid and developmentderives from the quality of aid itself.There is a growing body of evidence thattraditional forms of development aid donot make a major difference. If volumetargets only help to ratchet up the sameaid on a larger scale, then more

Target the MDGs—Not Aid Amounts

by Stephen Browne,International Trade Centre, Geneva

1. Only five countries have done so: Norway,Sweden, Denmark, Netherlands and Luxembourg.


development will not result. Here aresome of the distortions, which are relatedto the supply-driven nature of aid:

Most aid is administered bymany large and expensive publicbureaucracies each with proceduresof their own.Supply is excessive, duplicativeand pre-financed.Recipient countries are chosenaccording to the instincts of northernpoliticians and donor self-interests.The content and terms of aid arestrongly influenced by the needsand interests of the suppliers ratherthan the recipients, and the solutionsdo not stick.

Foreign solutions are not fully absorbed.More often they are grafted on to localinstitutions without strengthening themfrom within. Or they remain as enclaves.The proliferation of donor-funded‘project implementation units’ provideswidespread evidence of these grafts.Such PIUs are there to ensure that thedonors’ projects are carried out to donorspecifications. They employ foreign andlocal staff who are remunerated at levelswell above local scales. When the projecthas run its course, the PIUs disbandleaving little behind in terms ofsustainable capacity. Enclaves are similar.They include, for example, some of thewithering fruits of capital assistance:empty schools and hospitals, crumblinghighways and silted dams.

More valid targets for donors would bethe eight Millennium Development Goalsto which the international communityunanimously signed up at UN summitsin 2000 and 2005. The first seven goalsrelate to poverty reduction, education,health, HIV/AIDS, gender equality andthe environment. They can be appliedat the country-level since they specifyproportionate improvements in standardsby the year 2015, e.g. halving nationalincome poverty rates. The eighth goalrelates to donor responsibilities tofacilitate the other seven.

The merits of the MDGs are the obverseof the flaws of global aid volume targeting.The MDGs are measures of progress inhuman development, albeit partial ones.They are about development ends, rather

than crude aggregates of financial means.They are relatively easy to measureand there are data available on themfor most countries.

Aid has more meaning and legitimacywhen it is focused on humandevelopment outcomes at country level.Rather belatedly, the developmentcommunity is recognizing the wisdom ofnational plans and strategies which helpto map paths towards the achievementof MDGs, taking into account theindividual circumstances of each country.The proponents of aid volume targetingwant to calculate ‘MDG financing gaps’again using the facile assumption thataid can somehow miraculously purchasedevelopment ends. The MDGs are theright focus, but aid—on condition that it isthe right kind of aid—will be only one ofthe many, many conditions that need tobe in place if development is to progress.

Beyond traditional aid, the rich countriesneed to show a stronger commitment todevelopment. That would mean easingsome of the impediments which theyplace—or continue to perpetuate—ondeveloping countries.

First, they could forgive more debt.For the poorest countries this shouldmean wiping out external indebtednessentirely, and not lending more eitherbilaterally or through the multilateralsystem, such as the World Bank.External funding should be in grantform and should be used to financemutually approved programmes aimed atpromoting growth and meeting humanneeds, including the MillenniumDevelopment Goals.

Second, they could facilitate global tradeby demantling the barriers in the way ofexports from the poorest countries. Thereis much talk of ‘aid for trade’ these days,but still developing countries face unfair,heavily subsidised competition and hightariffs on their exports of goods to whichthey try to add value.

The so-called ‘donor’ countries collectmuch more in import tariffs fromdeveloping countries than they providein aid. For example, the USA in 2006imported $37 billion from France andcollected $367 million in tariff revenues;

it collected exactly the same amount oftariff revenues from Cambodia from $2billion of imports. The USA collectedsimilar tariff amounts from the UK andfrom Bangladesh, although the former’simports were worth seventeen times asmuch. And while global trade rules tryto prevent developing countries fromsubsidising their farmers, agriculturalprotection in the richest countries stillruns at about three times the level ofannual ODA, viz. $300 billion.

In addition to lifting these impediments,rich countries could do more to financeand propagate global public goods,which poor people in developingcountries need more than hand-outs.More funds should go into vaccinesand medicines against the diseases thatattack millions of children and adults.The Global Fund for AIDS, Tuberculosisand Malaria (GFATM)—three diseaseswhich kill 6 million people per year in thedeveloping world—needs more funds, asdoes the Global Alliance for Vaccines andImmunization (GAVI) to reduce mortality.Theirs are meaningful targets, becausethey lead to monitorable results. Otherpublic goods could help eliminate thepublic ‘bads’ of conflict and violence,trafficking of people, drugs, small armsand landmines.

There is also a rapidly mountingconcern for responsible environmentalmanagement to turn back globalwarming, the most deleteriousconsequences of which will fallon the poorest countries. Funds areneeded to accelerate the developmentand deployment of technologicalsolutions which will wean the worldfrom hydrocarbons.

If donors want to do more good for theworld, they need to review their broaderdevelopment footprint: reducing theharmful effects of conditioned aid anddistorted trade, while creating a morepropitious global environment.

Traditional aid targeting won’t cut it.

S. Browne: Aid and Influence: Do DonorsHelp or Hinder? Earthscan, 2006.http://shop.earthscan.co.uk/ProductDetails/mcs/productID/587/groupID/7/categoryID/7/v/2

http://shop.earthscan.co.uk/ProductDetails/mcs/productID/587/groupID/7


The donor community is not unaware ofthe constraint to effective aid of weakinstitutions of the state, including inAfrica. Indeed much of the discussionof increasing aid to Africa has been inthe context of a ‘compact’ in which moreaid from donors would be matched byincreased attention to ‘good governance’by recipient governments.

Countries where average incomes arelow because long-run growth rates havebeen low can easily be identified and alsowhat aid should pay for: infrastructureand social investments that a country istoo poor to finance itself. Identifying thecountries that are victims of a WIT is notso straightforward. Growth is a poorindicator since there is probably somelag between an improvement ininstitutions and subsequent growth, andsince recent growth is a poor proxy foradequate institutions given the frequentgrowth reversals after periods of growthaccelerations. Moreover, absent an ex antedefinition of what constitutes weak‘institutions’—or how these can bestrengthened—it is not clear whataid should pay for.

There is strong research evidencethat institutions, variously described,matter for growth. The problem is that‘institutions’ covers many rules, habits,customs, cultural and social factors andmore; the state of a country’s institutionsis harder to describe let alone measurethan is the extent of its poverty. The linksfrom specific policy actions or aidprograms to the institutional outcomesthey measure are complex; it is hard todifferentiate among the manyinstitutions—rule of law, control ofcorruption, free press—that mightmatter now for subsequent growth, norexplain how changes in institutionalstrength in a country may affectsubsequent growth.

Donors’ implicit assumption is thatAfrica is stuck in a poverty trap and thatmassive aid is necessary to escapefrom the trap. I suggest an alternativeassumption: that Africa is caught in aninstitutional trap, signaled and reinforcedby the small share of income of itsindependent middle-income population.Theory and historical experienceelsewhere suggest that a robustmiddle-income group contributescritically to the creation and sustenanceof healthy institutions. If external aid isto be helpful for institution-building inAfrica’s weak and fragile states, donorsneed to emphasise not providing moreaid but minimising the risks more aidposes for this group in Africa.

The key to understanding Africa’spoor long-run growth record—andlimited poverty reduction—may be theweakness of its ‘institutions’, includingits institutions of the state. With neitherthe institutional structure that providessufficient ‘autonomy’ from interest groupsnor the democratic arrangements andhabits that make the state accountable tothe citizens, the state in many Africancountries fails to protectthe property rights—except of the fewinsiders—that sustain productive privateinvestment and risk-taking; indeed in theworst cases the state actually abusesthe property rights of citizens.

Thus, rather than a poverty trap I wouldcall this a ‘weak-institutions trap’—a WIT!The poverty trap idea supports anemphasis on increasing the quantity of aidas the key to sustained growth. The idea ofa weak-institutions trap, and the failuresof past aid programs that led to highdebt burdens without growth in manycountries, supports emphasis on thequality of aid, including the nature of aidprograms and their impact on political andother institutions in the recipient country.

by Nancy Birdsall,Center for Global Development Do No Harm: Aid and the

Missing Middle in Africa

Africa is caught in a‘weak-institutions trap’signaled by thesmall share ofmiddle-class income.

The state in many casesfails to protect the propertyrights—except of thefew insiders—that sustainprivate investment.

Donors need to avoidharm to the fragile middlethat fuels growth andstrong, accountabledemocratic institutions.

The impact of volatile aidon skilled labour, interestrates, investor confidenceand domestic revenuegeneration must beclosely monitored.


In short, we have no reasonablypredictive measure of a country’sinstitutions and thus its WIT status.But recent political and economictheories of the role of state institutionsin growth suggest possible currentindicators that combined with thesystematic measures of governanceshould be warning signs.

Among likely underlying causes ofa country’s suffering a WIT are:

Heavy dependence on mineral andoil exports, leading to Dutch diseaseand limited investment in humancapital, high income inequality,limited creation of productive jobs,and little state accountability to themajority of citizens.Low natural openness; landlocked,limited access to the sea and non-trading neighbours. That reducesopportunities for trade andthus for export-driven growth.Problematic borders combined withethnic heterogeneity, which togetherundermine the legitimacy of the state.

Among symptoms of a WIT likely toreinforce such causes are: Primarycommodity dependence and associated

economic volatility, recent experience ofconflict and/or current inability of thestate to effectively control all of its ownterritory, low non-trade tax revenues,prevalent corruption, and lack ofexecutive accountability.

Some evidence suggests an additionalsalient indicator: an unequal distributionof income with a small ‘middle’, i.e. wherethe 60 per cent of households in middlereceive a small proportion of nationalincome. This should more systematicallyaffect decisions about the amount ofaid a country can absorb, and the typeof aid that makes sense.

Most middle-income households in Africaare actually poor by internationalstandards, or at risk of becoming poor.While maintaining their concern for the‘poor’ as conventionally defined, donorsneed also to avoid harm to the fragile‘middle’. Of special concern should be theimplications of high and unpredictable aidinflows for small entrepreneurial activityand job creation in the private sector.

In the more than 20 countries alreadyhighly dependent on aid—where aidconstitutes 10 per cent or more of GNPand as much as 50 per cent of total

government spending—donors andrecipient governments should monitormore closely the effects of aid and ofplanned aid increases on the labourmarket, particularly for skilledworkers; on interest rates and othermacroeconomic variables; on domesticinvestor confidence, given the volatilityof past aid; and on incentives fordomestic revenue generation.

Many countries in Africa that are highlydependent on aid have the symptoms ofa weak-institutions trap. Strengtheningtheir local institutions, particularly stateinstitutions, is key to their sustaininggrowth. Proposals for reform of aid tomake it compatible with institution-building exist. But even in the case of‘capacity development’—the first andoften last resort of donors—caution iswarranted given the failures of the past.In the end experience suggests that theinstitution-building process is a localtask; it is not particularly amenable tooutside help.

A first step for donors is to reduce therisk that aid from outside underminesexisting and incipient institutions.Broad reforms of aid practice wouldhelp—harmonisation would reducepoaching, and greater predictabilitywould avoid the volatility thatdiscourages domestic investment.

But ambitions and rhetoric aboutbroad donor reform are way aheadof the reality of how donors actuallynow behave. Donor reforms may yettake hold, but in what may be a long‘meantime’ official donors active inaid-dependent countries ought tofocus on a more modest goal: whiledoing good for the poor, do not do harmthe productive middle. After all, in theadvanced economies and in developingcountries that have sustained growthfor several decades, it is the productive‘middle’ that not only fuels sustained,private sector growth but provides theballast of accountable, democratic andstrong state institutions.

Nancy Birdsall: Do No Harm: Aid, WeakInstitutions, and the Missing Middle inAfrica. Center for Global DevelopmentWorking Paper No. 113, March 2007.http://www.cgdev.org/content/publications/detail/13115

Payments for progress...

... is a proposed hands-off approach to scaling up foreign aid that would explicitlybe aimed at strengthening local capacity and institutions, including in fragile states.It would link additional aid to clear evidence of progress already achieved on theground. This approach would give flexibility and autonomy to local institutions,providing an opening for experimentation, while at the same time ensuring that aidpays only for real, measurable achievements. Donors would bind themselves as a groupto pay a specific amount for clear evidence of progress against one or more agreedgoals in low-income developing countries.

Developing country governments would present an independently audited statementreporting their progress on the measures, and donors would pay the agreed amount.Payments would be determined as a function of the outcomes, and not linked to theimplementation of any particular policies, any other intermediate outputs, or ‘tied’to purchases from particular suppliers or companies. Governments that found ways toprovide services efficiently and so reduce the costs of providing them would benefitfrom a larger surplus. Such an approach raises issues in setting the benchmarks againstwhich progress is measured, in avoiding cheating, and in managing unintendednegative consequences of an incentives-based approach. But it entails importantadvantages for both donors and recipients.

Owen Barder and Nancy Birdsall: Payments for Progress: A Hands-Off Approach toForeign Aid. Center for Global Development Working Paper No. 102, December 2006.http://www.cgdev.org/content/publications/detail/11550

http://www.cgdev.org/content/publications/detail/11550http://www.cgdev.org/content/publications/detail/13115


Aid Does Work,but Beware ofGreat Expectations

by Finn Tarp,University of Copenhagen

Looking for stable andpredictable impacts ofaid on growth anddevelopment is nosimple task.

A careful review of theevidence indicates apositive impact overall,but aid alone will notturn the wheels of history.

‘Good policy’ can be amisleading criterionfor aid allocation, andsimplistic rules-of-thumbmay make things worse.

There is much tocriticise in current aidpractices, but a lot ofprogress is being made.

Foreign aid remains squarely on mostpolicy agendas concerned with growth,poverty and inequality. Yet controversyabout aid impact is rampant. This is notsurprising. Even a cursory look at historyshows that development over the past 50years has been a complex and variegatedprocess. There have been interrelatedchanges in resource accumulation,population, knowledge and productiontechnology—all in the context of frequentand dramatic change in policies andinstitutions. Looking for stable andpredictable impacts of foreign aid ongrowth and development is no simpletask under such circumstances. Analystsand policymakers are well advised to bewary of the dangers of oversimplification.

In particular, one cannot concludethat aid has been a fiasco just becausegrowth in Africa has been less thandesired, or because some developmentprojects have failed. Development is,and will always be, a risky business.Unsuccessful investment projects andpublic sector activities abound in eventhe best of political, social and economiccircumstances. If all investments weresuccessful, investors are likely to havebeen too risk-averse.

Serious aid impact analysis must (i) tryharder to uncover whether foreign aidhas on average had a positive impacton development in recipient countriesor not; and/or (ii) identify themechanisms through which aid impactse.g. on growth, including all potentialpositive and negative effects. A recentpaper (see end-reference) reviews theevidence on whether aid has beeneffective in furthering economicgrowth and development.

It is easy to find a negative associationbetween aid and growth in simplecorrelation analysis. Yet, there is no

inconsistency between little growth andlow aid inflows. Over the past decadethe mode of annual aid per capita inrecipient countries amounted to lessthan $20 and the median was under$35. Moreover, the simple correlationcoefficient between growth and aidmay be insignificant, or even negative,because donors allocate more aid topoorer countries, subject to difficultiesand shocks of many kinds, includingnatural and man made calamities. Whencountries have done well for a while soaverage income has gone up, donorstend to transfer less aid and eventuallythey withdraw. While such ‘graduation’may take a while, simple correlationsare thus likely to show a negativerelationship, but they do not revealthe ‘true’ impact of aid.

Aid allocation matters, as does themajor changes that have taken placein the global economy and affectedthe environment in which aid isimplemented. Targets for aid havealso been changing from one decadeto the next. Hence, simple correlationanalysis or story telling cannot—andshould not—settle the causalitydebate about aid’s potential impacton development.

Similarly, it is never straightforward togeneralise from case studies; this helpsexplain the surge of macroeconomiccross-country studies of the aid-growthlink from the mid-1990s. Such anapproach makes it possible to movebeyond simplistic aid-growthcorrelations, where the analysis of causaleffects is rather primitive. Much of themodern empirical aid-effectivenessliterature has focused on whether theimpact of aid is conditional on policy orwhether aid can be expected to have aseparate and positive impact,independent of policy.


The single mostcommon result in themodern aid-growthliterature is actuallythat aid has had apositive impact onper capita growth.

This has involved a mixture of concerns,ranging from technically demandingeconometric modelling issues tofundamentally different approachesto development strategy and policy.Thus, it is clear that the ‘true’ aid-growthrelationship is far from easily established.In any case, aid is of much too limited asize to turn the wheels of history.

Yet, this does not justify rejecting aid asa useful instrument in the fight againstpoverty. Nuanced and subtleassessments are advisable withthe empirical evidence in hand; andthe single most common result in themodern aid-growth literature is actuallythat aid has had a positive impacton per capita growth. No excessiveclaims about aid impact should bemade on this basis—complacency isnot called for. Still, these establishedresults should not be overshadowedby the aid criticism.

Turning to the debate about aidallocation, there appears to be meritin more sophisticated versions ofarguments for selectivity. Structuraladjustment lending makes little sensewhen the macro policy environment is‘bad’ and there is little possibility forpolicy reform. However, the empiricalwork on aid effectiveness has madeit clear that macro criteria shouldnot stand alone in assessing aideffectiveness and determiningaid allocation.

‘Good policy’ can be dangerouslymisleading as the fundamental criterionfor aid allocation, and simplistic macrorules-of-thumb may make things worse.Many of the world’s poorest people sufferunder substandard governance and lackthe means for changing it. It would begravely ironic for aid agencies tocompound the misfortunes of thesepeople with discriminatory aid allocation.

Overall, based on the research results it isjustified to argue for increased aid, butexpectations about its impact on growthshould be kept at reasonable levels.It would be unfortunate if unrealisticexpectations about aid impact are onceagain built up as in the early stagesof international aid. At the same time,asserting that aid has a positive impact

and should be ‘scaled up’, and that itsimpact does not appear to be conditionalon ‘good policy’, in no way contradictssuggestions that future aid should becarefully redesigned.

Much is indeed already happening onthe foreign aid scene. Major shifts havetaken place in aid modalities over thepast 15 years, and the general rise of aculture of transparency andaccountability is more than superficial.

The March 2005 Paris Declaration onAid Effectiveness codifies best-practiceexperience and it is now beingimplemented (see below, page 26).

Three sets of critically important butunresolved issues remain:

Foreign aid has been associated withboth development successes andfailures. The basic analytical problemin assessing its impact is to identifythe underlying development model.We therefore continue to work withdebatable reduced form models,while the data suggest that aideffectiveness is very uneven. Theconditions for aid having a morepositive development impact remaindisputable. It remains a challenge tobetter understand what drives thedifferent impacts of foreign aid, e.g.the potential interaction with policiesand structural characteristics.

We lack the necessary understandingof the complex, country-specific linksbetween aid, growth and developmentobjectives such as poverty reduction tojustify selectivity as the basic approachto aid allocation. This does not meanthat I favour old-fashioned ex-anteconditionality, but a betterunderstanding of the intricacies of the

donor-recipient relationship in theoryand in practice would be valuable.

This would as key elements includeaddressing issues such as how to best(i) channel resources to the poor whennational governments fail to do so;(ii) ensure that aid delivered directlyto national governments does notundermine local accountability;(iii) establish the appropriate balancebetween aid to the governmentand to private sector entities; and(iv) strengthen incentives for genuinedomestic policy leadership withlearning-by-doing adjustments.Furthering accountability andtransparency vis-à-vis localpopulations is a demandingbut crucial task.

In the present drive to scale up aid, itis critically important to avoid makingthe mistake of the past of promisingtoo much, i.e. contributing to themisconception that aid can on itsown turn history. Based on historyaid has much to offer, but managingexpectations is far from easy. A majorchallenge is to make sure promisesmade are actually kept. There aremany unresolved issues here,including how to design incentivesin aid agencies to meet this challengealongside topics such as the role ofindependent evaluation, of donorcoordination, and of the need tosharpen the incentives for recipientsto maximise reform efforts.

In conclusion, it would be gravelyironic if disagreement about overalldevelopment strategy and themacroeconomic impact of aid got in theway of pursuing practical and useful aidfunded activities in poor countries. Thereis much to criticise in foreign aid, butpossibilities for constructive and forwardlooking action should be kept in mindthroughout. There are in my experiencelots of them out there in practice. Theydeserve to be uncovered more preciselyand implemented effectively to thebenefit of those in need.

F. Tarp: Aid and Development, SwedishEconomic Policy Review, vol. 13, 2006,pp. 9-61. http://www.ekonomiskaradet.se/Panda_ekonomiska/Data/Documents/sepr2006_2/Tarp.pdf

http://www.ekonomiskaradet.se/Panda_ekonomiska/Data/Documents/sepr2006_2/Tarp.pdf


The global povertyreduction impact of aidvaries with its allocationacross countries.

This affects donors’allocation criteria: shouldthey go for maximum globalpoverty reduction orfor reaching the MDGsin each country?

Should aid compensatedisadvantaged countriesand try to create equalopportunities for all?

These alternatives have asubstantial cost in termsof foregone globalpoverty reduction.

by Edward Anderson,University of East Anglia Aid Allocation and

the MDGs

How donors decide to allocateaid across countries is an importantand highly political policy issue.It concerns fundamental principlesfor the allocation of aid as well asthe evidence on the extent to whichaid promotes development objectivesin different country contexts. This articlediscusses these issues and how theyaffect approaches and strategies forachieving the MDGs.

One principle which can guide aidallocation decisions is that of ‘poverty-efficiency’. This principle can be statedsimply: aid should be allocated so as toachieve the largest possible reductionin poverty at the global level, whichis consistent with the view that theover-riding purpose of aid is toeradicate poverty. Nevertheless,it does have its critics.

One concern is that its strict applicationwould cause donors to neglect poorcountries in which the effectiveness ofaid at reducing poverty is consideredto be very low, due for example tosubstandard governance.

Another concern is that the principlefails to take into account the differentopportunities that developing countriesface for reducing poverty. For example,poor countries which are geographicallyisolated, or which have few naturalresources, will typically find it harder toachieve sustained growth and povertyreduction than other countries. Manywould argue that such countries shouldbe allocated additional aid, so as tocompensate for their relative lackof opportunities.

Alternatives to the poverty-efficiencyprinciple, which address these concerns,do exist. One alternative would be toallocate aid so as to achieve a targeted

reduction in poverty in each recipientcountry. Another alternative wouldbe an ‘equal-opportunity’ aid allocation,which would allocate aid in a way whichcompensates countries facing feweropportunities for growth andpoverty reduction.

Which of these different principles—poverty-efficiency, country-by-countrytargets, or equal-opportunity—is right isof course a normative question involvingphilosophical and moral considerations.It is, however, a question certainly worththinking more about, not least to seehow much consensus exists on whichprinciple is considered to be right.

To apply any aid allocation principlein practice, estimates are requiredof the likely effects of aid on thingslike economic growth, poverty andgovernance, in different countrycontexts. Some studies have shown thatthe effect of aid on poverty, via its effecton economic growth, is greater incountries with a more favourable policyand institutional environment. However,there are now upwards of 60 differentestimates of the effect of aid oneconomic growth. Many of these findthat aid’s effect does not vary muchacross recipient countries, or thatit varies according to countrycharacteristics other than the policyand institutional environment.

Because there is relatively little consensusabout precisely how aid affects poverty,or other relevant outcomes, and abouthow these effects vary across recipientcountries, there is a lot of disagreementabout what a poverty-efficient aidallocation, or an equal opportunitiesaid allocation, would look like in practice.Possible ways of resolving this dilemmaare proposed below. First however, it isworth considering the conflict between


Notes: Figures refer to the 41 low-income countries with recent $1-a-day poverty estimates (89% of total populationof low-income countries in 2002). The additional amount of poverty reduction is calculated relative to a country-by-country approach to meeting MDG 1. References: See ODI Briefing Paper No. 19.

alternative aid allocation principles inthe context of attaining the MDGs.

In debates building up to the MDGReview Summit in 2005, the unevenprogress towards each goal acrosscountries and regions becameimmediately apparent. For instance,while East and South Asia were eitheron-track towards, or had already met,the target of halving $1-a-day povertyby 2015, progress in Sub-Saharan Africahad been negative.

The main reaction to this unevenprogress was to recommend a largescaling-up of aid, precisely to accelerateprogress in those countries and regionswhere progress was lagging. This wasbacked up by an interpretation of theMDGs as country-level targets as well asglobal ones, most prominently by theUN Millennium Project, which in itsmain report Investing in Developmentinterprets the MDGs as country goals,“since this is the spirit in which they arepursued the world over”.

Clearly, a country-level interpretationof the MDGs requires donors to allocateadditional aid to lagging countries, evenif that aid could have a larger impacton poverty elsewhere. It thereforedeparts from the principle of globalpoverty efficiency, and involves asacrifice—an ‘opportunity cost’—interms of the amount of progress

towards the MDGs achieved at theglobal level.

How large is this opportunity cost?Recent research has estimated its sizein relation to MDG 1, that of halving$1-a-day poverty by 2015. It calculated(i) the total amount of aid required ifeach low-income country is to achieveMDG 1; (ii) the amount of aid each low-income country would receive if thistotal amount were instead allocatedon a global poverty-efficient basis; and(iii) the overall level of the $1-a-daypoverty measure for all low-incomecountries through to 2015 under thetwo allocation systems. The differencebetween the two estimates gives theopportunity cost of a country-by-country approach to meeting MDG 1,viz. the amount of poverty reduction atthe global level which is sacrificed asa result of departing from a poverty-efficient allocation.

The results of this research show that theopportunity cost of a country-by-countryapproach to meeting MDG 1 could be verysignificant: it is unlikely to be less than 10million people, and could be as high as70 million people (see figure). Theadditional reduction in poverty under apoverty-efficient allocation would beachieved in two main ways. The firstwould be by allocating relatively moreaid (compared to a country-by-countryallocation) to South Asia, and relatively

less aid to Sub-Saharan Africa. Thesecond would be by allocating relativelymore aid within each region to countriesin which the effectiveness of aid atreducing poverty is seen to be higher.

Of course, these findings do notnecessarily imply that interpreting theMDGs on a country-by-country basis, andallocating aid accordingly, is wrong. Suchan approach may be justified, for instancebecause it avoids the possibility thatcountries in which the effectiveness of aidis considered to be very low are by-passedby aid. In this case, the opportunity cost ofthe approach would be a price worthpaying. The fact remains, however, thatdonors should be aware of the size of thiscost, even if they regard it to be offset byother normative considerations.

Two main implications arise from theissues discussed here. First, donorsseeking to implement any allocationprinciple in practice should consideralternative estimates of the effectivenessof aid in different country contexts.Different estimates may well suggestdifferent ‘optimal’ allocations, though itmay still be possible to identify countrieswhich are currently under-aided (or over-aided) whichever estimates are chosen.

Second, there is a need for more clarityand debate about the underlyingprinciples on which aid allocationdecisions should be based. One issue iswhether aid should be allocated so as tohelp achieve the MDGs on a country-by-country basis or at the global level only.Another is whether an equal-opportunityaid allocation is preferable to a poverty-efficient allocation.

Overall, the large scaling-up of aidvolume called for and agreed during2005 need not come at the expense of asound approach to its allocation acrosscountries, as long as donors engage inan open discussion about the principlesaccording to which allocation decisionsare made, and apply those principleswisely in practice.

E. Anderson: Aid allocation and the MDGs.How much should different countries getand why? Briefing Paper No. 19, OverseasDevelopment Institute, London, April 2007.http://www.odi.org.uk/publications/briefing/bp_april07_aid_allocation.pdf

http://www.odi.org.uk/publications/briefing/bp_april07_aid_allocation.pdf


by Patrick Guillaumont,CERDI1/Université d’Auvergne Aid Works Best in

Vulnerable Countries

Aid effectiveness is linkedmuch more to the economicvulnerability of countriesthan to their policies.

Export volatility has anegative impact on growth,while aid tends to have theopposite effect.

Aid increases stabilityand growth, while alsomaking growth morepro-poor by protectingfrom poverty traps.

Aid should increase more,balance directly productiveand social sectors, and serveas insurance against externalshocks and export volatility.

1. Centre d’Études et de Recherchessur le Développement International.

The overall impact of aid onincome poverty has been examinedmainly through its effect on economicgrowth. According to the most influentialparadigm, this impact is considered asmainly depending on policy. The resultingmessage—that priority in aid allocation isto be given to countries with ‘good’ policiesand institutions—met a moral andpolitical sentiment not always groundedon a robust assessment of aid effectiveness.

The sometimes intense debate on suchselectivity in allocating aid, initiatedby World Bank research some 10 yearsago, has at least made clear that aideffectiveness is likely to depend on somespecific features of the recipient country.But which features are most important?

One strand of empirical research hasshown that a major factor conditioningaid effectiveness at the macro levelin recipient countries is the economicvulnerability they face. In vulnerablecountries, foreign aid has a high marginalproductivity in avoiding collapses whenshocks occur, sometimes followed bylong lasting recessions. Aid can smoothpublic expenditures and lower the risk offiscal deficit. Consequently the marginalcontribution of aid to economic growthis shown to have been significantlyhigher in recipient countries exposedto external shocks.

There is not much debate about thenegative impact of ‘one-sided’ naturalnegative shocks such as earthquakes,typhoons or floods. The damage causedby these events is often huge, first bythe number of deaths, second by thedestruction of physical capital. The debateis rather about the measurement of thesize of these losses. Many shocks are‘two-sided’, in particular externaleconomic volatility, such as a successionof booms and slumps of export prices,external demand, rainfalls etc. Therefore,

to assess vulnerability over a long periodit is appropriate to consider the impact ofinstability or volatility rather than theimpact of separate shocks.

The effects of export instability, a mainsource of structural vulnerability indeveloping countries, have beendiscussed for many years in the literatureon growth regressions. There seems tobe now a consensus emerging fromseveral studies to conclude that exportinstability, or in some studies terms-of-trade instability, has a negative effect ongrowth. This effect comes mainly throughthe instability of the rate of investmentand that of the real exchange rate, eitherby its impact on public finance whenretained at the government level or by itsimpact at the producer level.

Economic instability also affects growthby triggering political instability thatsometimes can lead to armed conflict,which is a very important cause ofeconomic decline. The instability ofexports is higher for countries exportingunprocessed primary commodities subjectto large price volatility. Other exogeneousshocks have been shown to have similareffects on the risk of conflict, for examplerainfall instability.

At the project level, aid has been shownto have stabilising effects that are notdirectly perceived by the evaluators ofeach project, since it cannot be easilyobserved at that level. Indeed thevulnerability of a country harms thesuccess of the projects, but less so whenthe level of ODA received by the countryis high, which leads to the conclusionthat aid reduces the negative effects ofeconomic instability. Research resultssubstantiate the macro-micro linkagesand provide further support that moreaid should be given to vulnerablecountries because it can dampen thenegative effects of shocks. This can be


done by taking into consideration ameasure of vulnerability in the criteriaused for aid selectivity.

By lowering growth, instability hasdeleterious consequences on the pace ofpoverty reduction. It also has direct socialeffects independently of its effects ongrowth. Two reasons make these directeffects likely.

One is the feeling of frustration generatedby a shortfall of income following a rapidexpansion which creates new needs andexaggerated expectations, as illustratedabove by the risk of civil war or of crime.The other reason is due to poverty traps,linked to the asymmetry of reactions ofhealth, education, employment to incomefluctuations. As far as instability lowersgrowth, it slows down poverty reductionnormally expected from growth, but alsoresults in an anti-poor bias for a givenaverage rate of growth.

There is evidence that instabilityof income worsens the social situationin low income countries, and tends toincrease—or hamper the otherwiseoccurring reduction of—the mortalityof children under five. Furthermore,instability of income pushes peopleinto poverty traps; poor people contractlasting health handicaps, their childrenleave school, laid-off workers getexcluded from the labour market, etc.Thus, the poverty impact of a rise ofaverage income is less than the impactof a fall. This assymetrical effect lowersthe average impact of growth on poverty,and increases poverty independentlyof income growth and inequality change.

The current concern about aid volatilitynotwithstanding, aid seems to have astabilising impact in the medium term; firstwith respect to exports volatility, secondand more generally as a dampening factorof income volatility. Aid volatility maylower and possibly cancel this effect whenit is procyclical with regard to exogeneousshocks, and even, but less likely, when it iscontra-cyclical and very high compared toother sources of shocks. Pro-cyclical aid canstill be stabilising if it is relatively lessvolatile than exports.

To assess the impact of aid on stability,we consider an index corresponding to

the difference between the volatility of(i) exports and (ii) the total inflow of aidplus export revenues. If the difference ispositive, aid is stabilising; if it is negative,aid is destabilising. The results for asample of more than 100 developingcountries over the period 1970–99 showthat, on average, it has been stabilisingand more clearly so during the 1990sthan previously; the indicator represents18 per cent of the average value of theinstability of exports, and 28 per cent fora sub-sample of African countries. On thewhole, aid was destabilising in less thanone tenth of cases.

The graph above illustrates the result interms of the net stabilising impact of aid,measured by the above-mentionedvolatility difference index. It shows thatfor any level of aid volatility overallstability improves with higher aid/GDPratios. At the same time, overall stabilityis not significantly influenced by thelevel of aid volatility.

The stabilising effect of aid is due to themix of: (i) a level effect; (ii) a contra-cyclicaleffect; (iii) a relative volatility effect.It explains why aid seems more efficientin terms of growth in countries moreaffected by export instability.

Aid is likely to promote poverty reductionby its stabilising impact through twodifferent channels: by increasing the pace

of growth and by avoiding poverty trapsand thus making growth more pro-poor.

There are three major implicationsfor aid policy:

First, there is a need to balance theutilisation of aid between directlyproductive and social sectors, in orderto avoid poverty traps and transitoryloss of competitiveness.

Second, aid should be designedmore to function as insuranceagainst exogeneous shocks, in orderto contribute to a faster and morepro-poor long term growth.

Third, due to the higher marginalimpact of aid in vulnerable countriesand in particular the Least DevelopedCountries (LDCs), where the need ofa big increase of aid is the largest,priority should be given to thesecountries in aid allocation. Toimplement such a new selectivityit is possible to use the EconomicVulnerability Index (EVI) set up at theUN for the identification of LDCs.

P. Guillaumont, S. Jeanneney Guillaumont:Big Push versus Absorptive Capacity: How toReconcile the Two Approaches, UNU-WIDERConference “Aid, Principles and Policies”Helsinki, June 2006. CERDI, Document detravail de la série Etudes et DocumentsE 2006.14. http://cerdi.org/Publi/ED_Detail.asp?Id=806

Note: The level of stabilising effect of aid as defined in the text is represented for 16 groups of developing countriesclassified by the quartiles of the aid to GDP ratio and the quartiles of aid volatility (Hodrik-Prescott measure).

http://cerdi.org/Publi/ED_Detail.asp?Id=806


The prospects for achieving theMillennium Development Goals (MDGs)look bad in various developing countries,notably in Sub-Saharan Africa. To turnthe tide, recent reports by the UNMillennium Project and the Commissionfor Africa issued urgent calls to increaseofficial development aid substantiallyand, thereby, close the gap betweendonor rhetoric and reality.

Qualitative aspects of aid allocation havereceived considerably less attention eventhough they may be as important foreffectively meeting recipient needs. Inparticular, a needs-based targeting of aidin priority sectors such as health andeducation should have an important sayon whether donors help achieving theMDGs. Hence, in addition to the usual

The overall share of thesocial sectors has nearlydoubled in 15 years, butthis varies a lot across donors.

Unless the sector targetingof aid is better focused onMDG needs, even muchlarger aid may not be enough.

Sectoral Aid Priorities:Missing MDG Targets?

by Rainer Thiele1, Peter Nunnenkamp1

and Axel Dreher2

ranking of donors according to theiroverall ‘generosity’, the structure of aidportfolios offers interesting insights asto whether aid has been prioritised inline with the MDGs.

The sectoral composition of aid by alldonors taken together has changedquite dramatically since the early 1990s.Most notable in the context of theMDGs, the share of aid devoted tothe social sector has almost doubled(to about 35 per cent in 2002-04), withhigher spending on education, healthand population programs, though noton water and sanitation.

However, this overall pattern maskssubstantial variation across donors.In recent years, social sector aid rangedfrom 23 per cent of total aid in Japanto 50 per cent in Norway. France andGermany put a strong focus oneducation but spent very littleon primary education, even thoughthe MDGs would require donors toconcentrate on this sub-category. TheUnited Kingdom stands out in thateducation and health-related aid wasfocused on basic services from whichpoor population segments might benefitmost. Denmark and Germany were theonly donors that provided a non-negligible share of total aid for basicwater and sanitation. Environmentalprotection and gender equality, whichboth explicitly correspond to MDGs, werelargely neglected by almost all donors.

Different aid priorities of donors mustnot necessarily imply inappropriatetargeting. While a multi-dimensionalobjective function follows from the MDGproject, coordinated donor efforts mayhave resulted in a division of labour withspecific donors concentrating on specificMDGs. Coordination and harmonisationindeed figure high on the policy agendaof donors. The Paris Declaration on Aid

1. The Kiel Institute for the World Economy.2. ETH Zurich – KOF Swiss Economic Institute.

Matching MDGs, indicators of need and aid categories

Indicators of need Aid categoriesTarget 2: Hunger

undernourishment developmental food aidmalnutrion of children emergency food aid

Target 3: Primary schoolingnet primary enrolment educationprimary completion rate basic educationaverage years of schooling

Target 4: Gender disparity in educationratio girls/boys in education educationliteracy ratio, males/females basic education

Target 5: Under-5 mortalityunder-5 mortality rate healthimmunization, measles basic health

Target 6: Maternal mortalitymaternal mortality ratio healthbirths attended basic health

Target 7: HIV/AIDSprevalence of HIV health

population programsTarget 8: Malaria, other diseases

incidence tuberculosis healthmalaria ecology basic health

Targets 10/11: Water & sanitation/ slum dwellersaccess to improved water water supply & sanitationaccess to improved sanitation basic drinking water


Effectiveness emphasized the need“to eliminate duplication of effortsand to rationalize donor activities tomake them as cost-effective as possible.”

Yet donor coordination remains elusive.Donors tend to favor the same ‘aiddarlings’. Comparing pairs of majordonors with regard to the allocation ofaid across 140 recipient countries, mostof the correlations turn out to be positiveand very few are significantly negative.This applies not only to total aid percapita of the recipients’ population, butalso to sector-specific aid for health,education, and water and sanitation.

It is against this backdrop that we assesswhether donors allocated aid accordingto specific needs of recipients. Weanalyze the impact of ‘indicators of need’related to the non-income MDGs on thedistribution of sector-specific aid across140 recipient countries. We control forthe per-capita income of recipientcountries and democracy indicatorsin order to isolate the impact of theindicators of need listed in thematching table below.

It is important to note that allexplanatory variables are weighted bythe recipient countries’ population. Thisimplies that the unit of observation isthe individual, rather than the country.This approach is taken because of theglobal character of the MDG concept;success or failure depends on thepercentage of the worldwidepopulation, rather than the number ofcountries reaching a particular target.Obviously, the results may be drivenmainly by China and India. Therefore,we perform additional estimations byexcluding these two heavyweights totest for the sensitivity of results.

We compare eleven bilateral andmultilateral donors on the basis of theiraid commitments in 2002-2004. The aidcategories under consideration aresupposed to be most relevant for aid tobe effective in contributing to the MDGs.This is not to deny that other aidcategories such as humanitarian andmulti-sector aid may also promote theMDGs. In addition, it should be stressedthat addressing the question of whetherdonors paid sufficient attention to the

MDGs by allocating aid according torelated indicators of need does not allowstrong conclusions on the effectivenessof aid. Well targeted aid is a necessary,though not a sufficient condition foraid to be effective.

We rank the eleven donors underconsideration as follows. Each donormay achieve a maximum of 64 ‘creditpoints’: one point for each significantcoefficient of the specific indicator ofneed in the allocation of sector-specificaid (see table); 0.5 extra points when asignificant coefficient is robust to theexclusion of China and India; another0.5 extra points when the per-capitaincome of recipients enters significantlynegative at the same time, revealing ageneral poverty orientation of aid.A similar procedure is applied forcomparing the targeting of sector-specific aid across the MDGs.

No donor comes close to the maximum.Yet there are striking differences(see graph). It appears to be in linewith conventional wisdom on particulardonors’ performance that Japan ranksat the bottom and Norway at the top.However, the group of donors withweak targeting includes not only another donor widely blamed to be low-performing—the United States—but alsocountries usually regarded as superiordonors like Denmark, IDA and Sweden.The latter donors are more MDG-orientedin terms of granting more aid to poorerrecipients, including humanitarian andmulti-sector aid. However, more sector-specific targeting has played a minorrole for them, too.

On the other hand, France ranges closeto the top in allocating aid to MDG-related priority areas. The fine-tuningof French aid according to specificindicators of need qualifies earlierverdicts that the poverty orientationof its aid allocation is particularly weak.Likewise, there are striking differences intargeting sector-specific aid across theMDGs under consideration (not shownin the graph). The fight against HIV/AIDS(Target 7) clearly stands out, with almosthalf of the maximum of possible ‘creditpoints’ being reached. This implies thatalmost all donors focused on this target.Target 2 comes second, though at aconsiderable margin—22 per cent ofpossible credit points. By contrast, varioustargets were largely neglected, namelyTargets 3, 4, 5 and 10/11. The targetingis particularly poor with respect to theobjective of achieving universal primaryeducation (Target 3). The allocationof aid in education was shaped by thecorresponding indicators of need forjust three donors—France, Germanyand Norway—and only weakly so.

All this invites the conclusion that thecurrent focus on substantially increasingaid in order to turn the tide and try toachieve the MDGs misses an importantpoint: Unless the targeting of aid toMDG-related priority areas is improved,increasing the amount of aid is unlikelyto have the desired effects.

R. Thiele, P. Nunnenkamp, A. Dreher:Do Donors Target Aid in Line with theMillennium Development Goals? A SectorPerspective of Aid Allocation. Forthcoming inReview of World Economics. http://www.uni-kiel.de/ifw/pub/kap/2006/kap1266.pdf

http://www.uni-kiel.de/ifw/pub/kap/2006/kap1266.pdf


A major problem facing both theInternational Monetary Fund and itscritics is the limited knowledge about keyeconomic relationships that determinehow macroeconomic policies willinfluence growth and poverty outcomesin a particular country, e.g. how differenttypes of public spending will affect futureeconomic capacity and competitiveness,how private investment might respondto lower fiscal deficits or how long anyincrease in aid flows will last.

What key actors assume about theserelationships influences fiscal policies.For example, whether higher aid-financed spending may cause adversemacroeconomic effects of concern forthe longer term depends critically on thelikely supply response to such spending.If higher aid-financed spending on non-traded goods pushes up the realexchange rate in the short-term—i.e.causes some temporary ‘Dutch disease’effects—we should not be too concernedprovided the spending improvescompetitiveness in the longer term.

So, the most important challenge facingaid-dependent countries is often not a

The IMF and Spendingfor the MDGs

by David Goldsbrough and Ben Elberger,Center for Global Development (CGD)

IMF conditionality haserred on the conservativeside, leading to lessaid delivery and lesseffective use of aid, e.g.for MDG-related spending.

It tends to favour domesticdebt reduction or externalreserve increases overadditional public spending.

It has not done enoughto help countries exploreand adjust to the macro-economic consequencesof higher aid.

The current approachis untenable. IMF shouldfocus more on analysis andless on specific conditionality.

‘macro’ one at all. It is to ensure thatadditional spending is used effectively,which requires good governance, soundpublic financial management, andstrong sector-level policies. If thoseare right, the more narrowly ‘macro’challenges will be manageable.

In fact, different judgments about thelikely effectiveness of higher spendingmay underlie many of the disagreementsbetween the IMF and its critics on macroframeworks with respect to theMillennium Development Goals (MDGs).With some recent exceptions, the IMF hasnot done a good job of incorporatingsupply-side responses into its country-level macroeconomic analysis and canimplicitly assume that additional publicspending will have little impact ongrowth. In contrast, those who produceMDG-costing scenarios often assumethat the ‘technical’ costing ratios aroundthe MDGs can be scaled up without thehuge governance problems and‘leakages’ common to other typesspending that have been a fact of lifein many low-income countries. Neitherassumption is likely to be correct.

In practice, we can never know, withouttrying, what the impact of moreambitious spending strategies will be.So countries are forced to make policychoices under considerable uncertaintyand balance the costs of differentpotential mistakes; between threats tomacro ‘stability’—i.e. inflation, Dutchdisease, and fiscal deficits all worsen due toweak supply responses—and foregoingexpenditure opportunities that mightimprove the lives of the poor and raisegrowth. How these risks are balancedshould depend on the macroeconomicsituation—which in most countries ismuch better than a decade ago—as wellas on governance and sector-specificpolicies that influence the likelyeffectiveness of additional spending.


Since the IMF does not have the expertiseto judge these sector-specific issues, itshould be humble in its macroeconomicpronouncements if sufficient informationis unavailable. Furthermore, even if theseeconomic relationships were fullyunderstood, many fiscal policydecisions—especially for the healthsector—involve fundamental socialchoices that should be left to nationalpolitical processes. The IMF job, therefore,is to help countries explore theconsequences of feasible policy optionsto clarify the trade-offs involved.

In this context, the IMF has not doneenough to explore more expansionarybut still feasible options for highergovernment spending. As a consequence,the initial fiscal content of some ‘IMFprograms’ has been too conservative.In particular, programs tended to favordomestic debt reduction or externalreserve increases over additionalspending, even when macroeconomicconditions were quite favorable. Facedwith huge uncertainty about keyeconomic relationships, many IMFprograms implicitly assumed withinsufficient justification that the balanceof risks was against additional spending.

The problem is not that the IMF ispursuing a ‘one size fits all’ approach,imposing the same squeezes on deficitsand spending in all countries. In fact, IMFprograms vary considerably in the size oftargeted changes in fiscal deficits andpublic spending (see graph). Reflectingbetter macroeconomic starting conditions,recent programs, on average, targetedsmall increases in fiscal deficits and overallspending, compared with targeted cuts inearlier programs. For example, programsnegotiated between 2003 and 2006targeted average expansions in the fiscaldeficit before grants and in spending ofabout ½ - 1 per cent of GDP in the firstprogram year.

Despite this shift toward moderate fiscalexpansion, both a recent study of IMFprograms in Africa by the IMF IndependentEvaluation Office (IEO) and the detailedcountry case studies for the CGD WorkingGroup (see end-reference) found that theIMF has tended to favour additionaldomestic debt reduction or external reserveincreases over additional public spending.

The IEO study found that, across allprograms in Africa, each additional dollarof expected aid was associated with atargeted fiscal expansion, i.e. additionalspending, of only 27 cents. Only whenexternal reserves were quite high, above2½ months of imports, and domesticmacroeconomic conditions highlystable—represented in the IEO study byinflation under 5 per cent—did programsaim to channel the bulk of additional aidto higher spending.

While the IMF is right to take accountof the level of reserves and domesticmacroeconomic conditions whendesigning fiscal policies and howaid will be used, the degree to whichthese factors influenced the fiscalstrategy seems too conservative. A widerrange of paths for the fiscal deficit andspending is now possible, followingdebt relief and with the prospect ofhigher aid, but there was often littleanalysis—at least in publicly availableIMF documents—of the rationaleunderlying the specific path chosenfor the fiscal deficit and overallgovernment spending.

With some commendable recentexceptions, the IMF has also failed toexplore the macroeconomic consequencesof scenarios for scaling up aid. In someearlier programs in the case studycountries, IMF aid projections wereoriented around goals of reducing aiddependency—e.g. Mozambique in 2002-2003—or avoiding borrowing even onconcessional terms—e.g. Rwanda in 2003-2004—without convincing macroeconomicarguments for the prescribed approach.The IMF programs did eventually adaptwhen substantially higher aid wasforthcoming but it is not possible tosay whether the initial negative signalsdiscouraged any aid.

The IMF approach seems to be slowlychanging: in the last couple of years,in-depth analyses of alternative scenariosfor ‘scaling up’ aid have been undertakenin several countries, e.g. Ethiopia,Madagascar, Zambia and, very recently,Mozambique and Rwanda. However,expectations of IMF staff by the IMFBoard remain unclear and much seemsto depend on the initiative of individualmission chiefs.

This ambiguity causes two problems.First, the IMF has not done all it coulddo to help countries explore themacroeconomic consequences of higheraid. Second, it risks sending confusedsignals to donors. For example, if onlyconservative scenarios are presented,does this mean the IMF thinks moreresources cannot usefully be absorbedfrom a macroeconomic perspective oronly that the IMF does not think moreresources will be forthcoming?

In fact, projections of aid to Africa inIMF programs remain conservative,reflecting skepticism by IMF staff—whichmay be justified—on donors’ resolve todeliver on their commitments to doubleaid by 2010. Of the 27 IMF programs andreviews in Sub-Saharan Africa that werecompleted in the 18 months after theGleneagles Summit, aid projections inonly two were as optimistic as theGleneagles commitments.

A fundamental message of the WorkingGroup is that if the IMF is to continuebeing heavily involved in thesecountries once macroeconomic stabilityhas been achieved—advisinggovernments on longer-termmacroeconomic challenges andsignaling to donors on the suitabilityof macroeconomic frameworks—itneeds to adapt its analytical approachand way of doing business. Clearly, analternative division of labour is possiblein which the IMF confines itself to short-term stability issues and makes nopretence of pronouncing on the longerterm challenges of scaling up aid andexpenditures. The IMF Board couldmake clear that the IMF role in post-stabilisation low-income countries willbe much more limited and scale back itsinvolvement and policy pronouncementsaccordingly. But, if the IMF is to continueto play the broader role that theinternational community seems to want,its current approach is untenable and theinstitution should transition to a rolefocused more on exploration and lesson specific conditionality.

Does The IMF Constrain Health Spending inPoor Countries? Evidence and an Agenda forAction. Report of the CGD Working Groupon IMF Programs and Health Spending.http://www.cgdev.org/section/initiatives/active/ ghprn/workinggroups/imf

http://www.cgdev.org/section/initiatives/_active/ghprn/workinggroups/imf


The prospect of large increases inthe flow of resources from wealthy topoor countries in support of efforts toachieve the MDGs is viewed with sometrepidation by Ministries of Finance andCentral Banks in some poor countries. Infact, in the past few years, many countrieshave actively tried to reduce the rate atwhich resources have been transferredinto their economies. While there maycertainly be real concern about severaldifferent kinds of negative side-effects oflarge aid inflows, the dominating fear inthe recent economic literature is that ofaid-induced ‘Dutch disease’.

What do economists mean by ‘Dutchdisease’? The theory is that high andsustained aid inflows have much thesame effect as a natural resourcewindfall, which can lead to anappreciated exchange rate and wageinflation, and thus to a loss of marketsand umemployment in export andimport-competing sectors. This effectappeared in the Dutch economy whenmassive North Sea gas revenues upsetthe macroeconomic balance—hence thename. The fear is that high inflows ofaid may thus prove to be a ‘curse’ bycontributing to the immiserisation oftheir export sectors.

If the Dutch disease scenario were likely,there would be good reason for fear. Itwould not only mean that countries maybecome unable to produce significantamounts of exportable goods—theremay also be very real adjustment costs.

Date post:	13-Feb-2021
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times