Professional Documents
Culture Documents
7 7 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
.,62 ..X4T:,
. ..
7'i.
t+,ez,;
Public Disclosure Authorized
Ffs
I~~~~~~~~~~~~~~~~~~~~~~~~~~~4
Public Disclosure Authorized
Ln~~~~~~~~~~~~
I E B~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-
NO; w++ j; ; _ si r
losure Authorized
Assessing Aid
and associatedcompaniesin
BERLIN IBADAN
The boundaries, colors,denominations, and other information shown on the maps in this vol-
ume do not imply on the part of the World Bank Group any judgment on the legal status of
any territory or the endorsement or acceptance of such boundaries.
Foreword ix
The ReportTeam xi
Appendixes 121
Estimating the Effectof Aid on Growth 121
Explainingthe Successor Failureof Structural Adjustment
Programs 126
Analysisof the Fungibilityof ForeignAid 130
The Impact of Civil Libertiesand Democracyon Government
Performance 135
Estimating the Impact of AnalyticalWork 139
Boxes
i. Defining Aid 6
2. MeasuringAid 8
3. Defining Sound Management: Good Policiesand Institutions 12
4. Functions of the DevelopmentAssistanceCommittee 13
5. StakeholderViews on Aid Effectiveness 24
Text Figures
1 Financial Flows to DevelopingCountries 7
2 OfficialDevelopmentAssistanceRelativeto Gross National Product,
Major Donors, 1991 and 1997 9
CONTENTS
Box Figures
2 Total Aid: OECD Official Development Assistance and Adjusted
Official Aid 8
Text Tables
1.1 Real Per Capita Income Relative to the United States 31
Appendix Tables
A1.1 Estimating the Effect of Aid on Growth 124
A1.2 Countries in the Dataset 125
viii
Foreword
F OREIGNAID ISAS MUCHABOUTKNOWLEDGEAS IT ISABOUTMONEY.
Helping countries and communitiesgenerate the knowledgethat
they need for developmentis a prime role of assistance.And aid
itselfis a learningbusinessthat continuallyevolvesas lessonsof successand
failurebecome clear.
AssessingAidis a contribution to this ongoing learningprocess.It aims
to contribute to a larger "rethinkingof aid" that the international com-
munity is engaged in-a rethinking in two senses.First,with the end of
the cold war, there is a group that is "rethinkingaid" in the senseof ques-
tioning its very existence in a world of integrated capital markets. In
response to this trend, we show that there remains a role for financial
transfersfrom rich countries to poor ones.
Second,developingand developednations alikeare reconceptualizing
the role of assistancein light of a new developmentparadigm. Effective
aid supports institutional developmentand policy reformsthat are at the
heart of successfuldevelopment. Though tremendous progresshas been
made in the past 50 years,global poverty remains a severeproblem.
Researchinto aid, growth, and poverty reduction providesimportant
evidenceabout making aid more effective.This report summarizesthe
findingsof recentWorld Bank researchon aid effectiveness.The primary
research has been published or will soon be published in professional
journals. This report aims to disseminatethe findings to a broad audi-
cnce. It should be stressedthat there are important aspectsof develop-
ment cooperationthat havenot been coveredby the research.The report
restrictsitselfto areaswherewe have new findings.
Aid is channeled through a variety of activitiesthat bring a mix of
money and ideas. Two key themes emergein this report. One theme is
that effectiveaid requiresthe right timing and the second that it requires
the right mix of money and ideas.
The timing of assistanceis crucialin helping countries improve their
policiesand institutions. When countries reformtheir economicpolicies,
well-timed assistancecan increasethe benefits of reform and maintain
popular support for them. Similarly,at the local level,when communi-
ties organize themselvesto improve services,the availabilityof aid can
make the differencebetween successfulinnovation and failure.
Ix
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Joseph E. Stiglitz
SeniorVice-Presidentfor
Development Economicsand
Chief Economist
The World Bank
November 1998
x
The Report Team
xl
xii
OVERVIEW
1
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
2
OVERVIEW
3
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
4
OVERVIEW
5
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
f:associte withf off;cialdevlom ent assiisnce andj to procure goods anerices fro the dono crnlnt
no0grallytargetedito the porest coafuntries.Tahis
fasss
lzSaitudijeshave igshow tat ti,iedad rieducesth valn o
ir l,ancei
s theprimary
gU focus of
thiM:i:s*fi::
study
i buti milanyof a tat asistanceabyN;ii ii25"H
. abu pecet and-ithere yfiffis gi
6
OVERVIEW
TheNewInternational
Environment
cal deficits and contain growth in government spending. Even though Source:Global
Development
Finance1998.
foreign aid is a tiny fraction of budgets, it has been one of the first items
for the ax. All major donors reduced aid relative to their GNPs between
1991 and 1997 (figure 2). The decline was especially sharp in the United
States-aid was a mere 0.08 percent of GNP in 1997. Sweden and other
Nordic countries have traditionally been generous, giving almost I per-
cent of GNP. But among large countries, France is the only one that gives
more than 0.45 percent. Collectively, OECD countries contributed just
0.22 percent of their GNP in 1997. The end of the Cold War likely
influenced some countries' decisions. The strategic importance of aid
has ebbed; as a result it risks losing its broad support among donor
governments.
At the same time, there has been a surge in private capital flows to
developing countries. In the 1970s and 1980s official finance-that is,
money from bilateral donors and multilateral institutions-represented
7
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
2
Box
MeasuringteAid
tb
Thee
hrere oins orh. akngabout thisfnewal innegon ro h dvlpe oth eelpn
dat.
te ajutedfigreten
Frs, lower fianute 1 Wthpiviatfl owsepnigt
world mr hn$5
traitina
mesur. ecod, hetwoseines areaiglabety eeoigcutis
corelaed
nd prduc siilr, esutsinvatnoencaptlfosaehaiycnetae nafwcutis
anayss.Mos
ad s.in hefom o gamhow teve-n
new fosae oaie urei h ae 90 eee
sure Thmacrecoomiceffetsaftadere
the osaetothdetciiin18.A terbgusocredn
regadles
ofwhic meaueof ad used he (se 99sc utwthapfnnia rss okn Esera n19
id.fiure sow hrpergdcinvesmndrpeshrl.Tefoofriaeoeytte
1).Thid,theadustd
maidin
he 99stha th E dmeasue.l Thingiorl felb 8 ilo ewe 96ad19.I n
ntret ateissmll
fntrnaioalinerest, priatesaeflw5 ilcniu og oasalnme f(oty
percent
largecoe
and
they96areoutris
1ceipercent.et cuntifs o
frvaeinanc
available
mntte tothohe14 developing countries.
8 ~~ ~ ~ ~~~~Piaecptlfosaehaiycnetae nafwcutis
OVERVIEW
1.0
1991
0.8 _
0.6
Source-WorldDeveloppment
Indicators1998.
In a typicallow-incomecountry,foreignaidremainsfar andawaythepri-
marysourceof externalfinance,amountingto 7-8 percentof GNP.
Developments in the 1990shavesharplyalteredthe climatefor for-
eignaid.The end of the ColdWaropenedup newpossibilities: with aid
no longer constrainedby thosestrategicobjectives,
it shouldbe possible
to makeaidmoreefficientat meetingits primaryobjectiveof long-term
growthand povertyreduction. 3 Yet,givenbudgetproblemsand rising
privateflows,donorsareclearlyrethinkingthe importanceand valueof
foreignassistance.
NewThinking Strategy
on Development
10
OVERVIEW
11
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
a~n poices
lad>ohatwil
api deelpmet ad ombines hree factors that hv ensoni
povery
in redicrin
paricula counry. eveloing mpirica studies toaffect eeoig oste
4ndonsa ia 6counftrywtorisiuinlqaiy
Botswn,ihits hihqualiyisiuin,i adf
ferent story.
measurebeodtemcecnmcadistu
tionaI fetue ilmntd heFoexml,efrt
12
OVERVIEW
Box4 Functions
oftheDevelopment
Assistance
Committee
SINCE ITS INCEPTION IN 1960, THE OECD's Over the Lastdecadeand a hf, the DAC has codi-
DevelopmentAssistanceComrmittee(DAC) has finc- fiedand publisheda comprehensiveset of Principlesfor
tioned as the principal strategy-settingand policy and EffectiveAid.These guidelinesand bestpracticescover
performancerevieworganof the major bilateraldonors. keypolicyorientationsand operationalissuesin centraL
Bilateral aid programs are not conceived and areasof aid managementsuch as cootdination, project
implemented in a political vacuum. Indeed they are assistance,programrassistance,technicalassistance,pro-
subject to considerabledomesticpressuresfrom poit- curement, and evaluation-as well as for such basic
ical and commercial interest groups in the donor dimensions of development as participationand good
countries.And bilaterl aid agenciescan be subject to governance,environment,and gender equality.
the same kinds of disbursement-driven dynamics as The DAC's Working Party on Aid Evaluation
multilateral development banks, creating incentives brings together the heads of the evaluation units of
for staff to be approval-focusedrather than result- bilateraland multilateraldevelopmentagenciesto work
focused.They can alsobe more concernedto show the on evaluationcapacitiesin developingcountries.
national flag on their development projects than to At the strategic level,the DAC produced in 1996
join collectivesector improvement efforts in which the report, Shaping the 21st Century, the Role of
donor identities are merged. Development Cooperation,which provides the basis
The DAC has been the forum where the major for the kind of partnerships most likely to produce
donors haveworkedto keep their programsfocusedon development progress: on the side of the developing
developmentobjectives,to promote coordination,and country, a strong commitment to an effectivepolicy
to review aid effectiveness.Every three years, each environment and key development priorities aimed
DAC member is subject to an examination of its aid at pro-poor growth, and on the donor side, increased
policiesand performanceby the other membersof the financial support for such policieson a program and
Committee, based on studiesby the OECD staffand budget level combined with an emphasis on partici-
led by two speciallydesignated "examiners" drawn pation and capacity-building (facilitating the trans-
from the Committee. These Development Coopera- fer of "knowledge and ideas").
tion Reviews,includingthe conclusionsreachedby the
DAC, havebeen publishedsince 1994. Source:Cyntributedby DAC staff.
These goals are elaborated on in Shaping the 21st Century: The Role of
Development Cooperation, produced by the Development Assistance
Committee (box 4).
All this points to a different role for aid. Development assistance is
more about supporting good institutions and policies than providing cap-
ital. Money is important, of course, but effective aid should bring a pack-
age of finance and ideas-and one of the keys is finding the right
combination of the two to address different situations and problems.
13
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
MoneyMatters-In a GoodPolicyEnvironment
14
OVERVIEW
Figure5 ProjectPerformance
by PolicyandInstitutional
Environment Projects perform better
with better policies and
Satisfactory
projects
(percentage of total) better institutions.
so T E
w~~~~~~~~~~..+
9. ;
60
40
20 hOHg
0 * - Low Economic
High- policy
Low
Institutional
quality
1
So -ce:WbrodBank 997a.
16
OVERVIEW
17
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
18
OVERVIEW
MoneyMatters-Ina GoodInstitutional 20
Environment 15
19
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
20
OVERVIEW
21
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Source:
Isham, Narayan, and Pritchett 1995. * Analyticalwork of donor agencies,such as the World Bank, has been
shownto increasethe probabilityof successof public projects.In one
study an additional $10,000 in analyticalwork returned $80,000 in
strongerbenefitsfrom projects.
* In a sampleof donor-financedrural water supply projects, the success
rate was 68 percent for projectswith high beneficiaryparticipation,
but only 12 percent for projects with low participation (figure 9).
More important, donor efforts to promote beneficiaryparticipation
helped the adoption of this approach to serviceprovision.
* Comparing project successacross countries shows the importance of
the institutional environment for public investments.In general,more
22
OVERVIEW
Money,
butMoreIdeas,Too
than foreign financial aid. Economies that lag are held back X
more by policy gaps and institutions gaps than by a financing gap. Aid as
money has a large impact only once countries have made substantial
progress with reform of policies and institutions. Poor countries with o Mediocre Goodpolicy
good policies should get more aid than ones with mediocre policies-but
in fact they get less (figure 10). Source:Collier and Dollar 1998.
Foreign aid has concentrated too much on the transfer of capital with
(often) scant attention to the institutional and policy environment into
which resources were flowing. This approach resulted from misunder-
standings about development-overemphasizing finance at the expense
of policies and institutions-and from external and internal pressures on
aid institutions. Disbursements (of loans and grants) were easily calcu-
lated and tended to become a critical output measure for development
institutions. Agencies saw themselves as being primarily in the business
of dishing out money, so it is not surprising that much went into poorly
managed economies-with little result.
23
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
accun
n evlomet
fr uceses ssitaceChefEcNoms,GpnsOC
~
remarkable: ~ an
mtu& ateshj ih Gsindvlp petra,~~~~~~~~Coe
Natsg
urpis
hes tun ot s be pl nd~ Pnt-
hadpoe fet~i nhnigamrhp~cm
tutona
eniromen onucve oerighg
man uiypriiain h h msn h
retuns romresurcs
aplid fr dveljnrenga ailiys poet,epcal hs delvein servGeto
leain
y eciies
roe cunris i dsinigad (ns mempweingm oam omntes u h ig
managing
k Dapmnikpr gramwhihhlseii nesobohlooranzisndcpesgoemnt
commit
and
merso
open
succes
o ernance o enforcetO vleNO i rca,ani,hr,th niomn
25
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
* More selective-putting more money into economies with sound Thereis reason to be
management. optimisticthat reformof
* More knowledge-based-usingresourcesto support new approachesto aid agencieswill
servicedelivery,expandingknowledgeabout what works, and dissemi- succeed.
nating this informationas a core business.
* Bettercoordinated-results-orientedagenciesshould worry less about
plantingtheir flagson particularprojectsand more about how commu-
nities,governments,and donorsworkingtogethercan improveservices.
* More self-critical-agenciesshould be asking themselvescontinually:
Why do we do what we do?And what is the impact?
Notes
1. Even a partial list would include books by those out- mote peaceful conflict resolution. Furthermore, humani-
side aid agencies (Cassen 1986, updated in 1994; Riddell tarian aid, which accounts for less than 10 percent of devel-
1996; Mosley 1987; Krueger, Michalopoulos, and Ruttan opment assistance and focuses on short term mitigation
1989), various institutional evaluations, and large bibli- efforts, is a separate issue not addressed in this report.
ographies on specialized topics, such as aid and macroeco-
nomics (White 1992), policy-based lending (Killick 4. The oft-made suggestion that the early consensus
1991), aid and nongovernmental organizations (Riddell, was on physical investment and that the new consensus is
Bebbington, and Peck 1995), and aid and post-conflict a shift to human investment is belied by the literature and
reconstruction (World Bank 1998b). by reality. Gunnar Myrdal's classic Asian Drama, written
in the 1950s, signaled the intellectual shift to roughly equal
2. Because of data availability, the focus of this study is emphasis on physical and human capital
assistance from OECD countries to the developing world.
It does not cover South-South aid (for example, from 5. It would be wrong to be overly critical of earlier
Kuwait or Saudi Arabia) or the Soviet Union's extensive views or those who held them, as economics is an inexact
foreign aid. science and reading the evidence is difficult. A 50-year-old
development economist making policy in 1960 would
3. The report focuses on international efforts to sup- have witnessed in his or her lifetime: the Great Depression,
port long-term growth and poverty reduction in the devel- the success of World War II planning in ending unem-
oping world. There are other important types of ployment and raising wartime production, the spectacular
international cooperation not covered by the study, such rise of the Soviet Union under central planning, and the
as efforts to reduce cross-border narcotics traffic or to pro- catch-up of Japan to the European powers.
27
CHAPTER 1
28
MONEY MATTERS-IN A GOOD POLICY ENVIRONMENT
DifferentCountries,
DifferentFortunes
29
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
thousandin Botswanain
250
1994 wvhile remaining .1967
-. 1994
virtually unchanged at Mali
50 ~~TanzaniaKey
Indonesia m
1, Hondurasc
otwna
Mauritius
0
sex discrimination, social exclusion, and the like may or may not
improve with income. When people improve their status from low to
middle income, new challengesemerge for society.Consider, for exam-
ple, the choking haze-from fires set to clear plantations-that
enveloped SoutheastAsia in the summer of 1997. Or the many devel-
oping country cities that have severeproblems with traffic congestion
and air pollution.
Still for low-income countries, generating self-sustaininggrowth is a
crucial firststep in development. If Ethiopia'stotal output were divided
evenly among households, everyone would be poor. Growth (an
increase in per capita income) is clearly required to solve poverty and
its problems: malnutrition, poor health, and lack of education and basic
social services.
Overall, developing countries have made tremendous progress:
more babies are surviving, more food is on the table, more children are
in school, and there are fewer deaths from easily preventable diseases.
There has been more global improvement in life expectancy in the past
40 years than in the previous 4,000 (World Bank 1993). The revolu-
30
MONEY MATTERS-IN A GOOD POLICY ENVIRONMENT
31
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Whythe Divergence?
32
MONEY MATTERS IN A GOOD POLICY ENVIRONMENT
Figure1.2 Economic
Management
andGrowth
inSelected
Developing Countrieswith better
Countries policiescan grow faster,
and reform can get them
Annual percentage growth in GDP per capita there-as in Ghanaand
Botswana C Bolivia.
1986-89
6
El Salvador Indonesia
4 1990-93 1982-85. EThailand
Bolivia 1974-77
2 1990-93 2 . '_El Ghana
Ghana India 1986-89
0
C, ~~~~~1986-e
196g. b 1990-93
Tanzania
-2 1982-85
Bolivia Zaire Zambia
1982-85 Peru 1986-89 1986-89
1986-89
-6 Nicaragua
E31986-89
-8
Poor Mediocre Good
Economicmanagementindex
33
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
34
MONEY MATTERS-IN A GOOD POLICY ENVIRONMENT
... AidSpursGrowth
The simple relationship between aid and per capita growth in devel-
oping countriesis weak, if it existsat all (figure1.4).3Some countriesget
a great deal of assistanceand grow slowly (Zambia, for instance)while
others also get a lot and grow quickly(Botswana,Ghana).
From this simple starting point, there are two problems in analyzing
the effectof aid on growth.First, other factorsthat affectgrowth must be
Figure
1.4 AidandGrowthin SelectedDeveloping
Countries,
197093 Somecountriesreceived
Annual percentagegrowth in GDP percapita a lot of assistance,and
incomes fell-some
8 countries received little
8 c3 S z aid, and incomes rose.
6
El
-2 m D -
_.m
-2 -6 ~ ~ E
0 3 6 9 12 is
Aid as a percentageof GDP
35
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
36
MONEY MATTERS-IN A GOOD POLICY ENVIRONMENT
n Foreign aid in itself is neutral with respect to development, for its pos-
itive or negative effects depend on government policies. Effects on eco-
nomic development will tend to be positive when aid is used to build
up capital or to finance public investment that contributes to the prof-
itability of private capital, or for human capital development.
n Fiscal policy should generate government's current savings, so that
both domestic and foreign resources finance public investment. If this
does not happen, foreign resources may end up financing the govern-
ment's current expenses and not investment projects, as happened in
Nicaragua.
* The relationship between aid's positive effects and good domestic poli-
cies always holds, even during adjustment (Lopez 1997).
"Clearly, the large influx of aid has not fostered rapid economic
growth in most African countries and has been only partly success-
ful in poverty alleviation. How can one reconcile this lack of suc-
cess with the field-level evidence just cited and reach a clear verdict
on aid's economic impact? Unfortunately, a number of factors pre-
vent the establishment of a clear link between aid and the overall
performance of the economy. First, government economic policies
themselves can prevent aid from having the impact it otherwise
would have. Policies have a critical influence on the effectiveness of
aid, for the same reasons that they affect economic growth;
37
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
38
MONEY MATTERS-IN A GOOD POLICY ENVIRONMENT
... Aid"Crowds
in" PrivateInvestment
C-
The impact of aid on growth in good-managementcountries is high. Good Average Poor
This may be becauseonce good management is in place, there are high- Economicmanagement
39
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Figure1.8 MarginalImpacton for instance, have more success in a good institutional and policy envi-
Privatellnvestment
of a Percent ronment: 86 percent, compared with only 48 percent in a weak institu-
of GDPinAid tional and policy environment (see overview figure 5).
There is another possible explanation for the relationship between aid
Private investment
as a percentageof GDP and growth-that when a poor country puts good policies in place, domes-
2.0 tic and foreign private investors are uncertain about the sustainability of the
reform. If there is fear of reversal, investors will wait. And even with good
1.5 management, low-income countries are hamstrung by other impediments,
including weak infrastructure. Foreign aid to reforming governments may
1.0 | improve the environment for private investment-both by creating confi-
dence in the reform program and helping to ease infrastructure bottlenecks.
0.5 r | In a poor-management country foreign aid may crowd out private
investment: it may encourage the public sector to make commercial
o | investments
2 | | X 2 that
___ would otherwise be undertaken by private investors. In
a good-management country the effects of foreign aid would likely be
-o.s1 g ,W magnified by crowding in private investment.
Good Average Poor There is evidence to support these notions (Dollar and Easterly 1998).
Economicmanagement
One percent of GDP in assistance increases private investment an extra
Source: Dollar and Easterly 1995.
1.9 percent of GDP in good-management countries (figure 1.8). Thus the
combination of good management and foreign aid is welcomed by the pri-
vate sector, and this helps to explain the strong effect aid has on growth in
such an environment. In a poor-management country, however, 1 percent
andWorldBankAidPerCapita of GDP in aid is estimated to reduce private investment by 0.5 percent of
andIncoime
Level GDP, which may explain why the net effect of such aid is small.
Aid per capita These findings for the effect of aid on growth and private investment
(log scale in dollars) shed some light on the problems of the Heavily Indebted Poor Countries
~~~~terBlatel
$7 ' Muilateral AidHasNotFavoredCountries
withGood
Management
$1 WorldBank
-- T'.
ARGETING ASSISTANCE TO POOR COUNTRIES WITH
institutions and policies will make the most of scarce aid
SOUND
40
MONEY MATTERS-IN A GOOD POLICY ENVIRONMENT
41
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Figure1.10 AidandPopuilation
Aid as a percentageof GDP (log scale)
55
7 0 o 0
e0 3oo 000
10 o
0.:14 o 03 0 0 >
0 000 0O
00 0 o o0; o
0 0
0
1 10 100 500
Population(millions, log scale)
Source:Calculatedfrom Burnside-Dollardaraser.
tive to GDP (figure 1.10). This makes sense for some assistance. For
instance, it costs about the same to provide technical assistance to the cen-
tral bank of India as to the central bank of Lao PDR. But if aid is financ-
ing public services, such as education or road construction, there perhaps.
should not be such discrimination against large countries. More than half
the world's poor live in India and China: in 1990-93 they received $2
and $1 per capita in aid. Yet small countries often receive $50 per capita
or more. This discrimination against large countries is one reason the rela-
tionship between aid allocation and income is not strong.
A second factor that undermines this relationship is that aid allocation
often depends on the political or strategic interests of donors-for exam-
ple, U.S. aid in the Middle East or European aid to former colonies.
Burnside and Dollar (1997) found that political factors help explain the
allocation of bilateral aid, but that multilateral aid is not strongly influ-
enced. These factors are not important for only a handful of bilateral
donors-Canada, Denmark, Finland, Netherlands, Norway, and Sweden
(Alesina and Dollar 1998).
The dual objectives-pursuing strategic goals and rewarding good
policies-work against each other. This can be seen by looking at the cor-
42
MONEY MATTERS-IN A GOOD POLICY ENVIRONMENT
relation between bilateral aid and management (figure 1.11). So, how
have bilateraldonors treated two countries with the same income and
population but different management? Good and bad management
regimeshave receivedroughly the same averageassistance.Being a for-
mer colonyof a major donor is more valuablein attracting bilateralassis-
tance than having good management (Alesinaand Dollar 1998).
The allocation of multilateralaid has depended on income, popula-
tion, and good management.Politicaland strategicconsiderationswere
not significant.Thus, for multilateralaid, lower-middleincome coun-
tries with good management received30 percent more than the typical
poor-managementcountry with the same incomeand population (figure
1.12). For very low-incomecountriesthe differencebetweengood man-
agement and poor managementhas been minor.
Since most official development assistance is bilateral, the alloca-
tion of all aid together (multilateral and bilateral) shows little rela-
tionship with the quality of country management. Cold War aid
driven by strategic considerations may have accomplished its political
goals, but aid that went to countries with poor management did little
to reduce poverty.
Figure1.11 Allocation
ofBilateral
Aid,1970-93
Aid as a percentage
of GDP
2'
incomne
Low ~~~~~~~~~~~~~~Good
management
Lower-middle Poor
income management
Real GDPper capita
43
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
BeFocused
C- NEEDSTO BE MORECONCEN-
DEVELOPMENTASSISTANCE
LEARLY,
trated on where it can be most effective in reducing poverty. It
At thie moment plenty of needs to take more account of the environment in which it is
countries combine mass placed. Financeis most effectivein reducing poverty in those countries
poverty with good that have both mass poverty and a good policy and institutional envi-
policies and institutions. ronment. Conversely,it is lesseffectivein countries that alreadyhave rel-
atively little poverty, or have poor policies and weak institutions.
But is there much scope for a more focused approach? Perhaps mass
poverty is so strongly linked with poor policies and weak institutions that
we just don't find the combination of mass poverty and good policies in
which aid is highly effective? Not so. At the moment plenty of countries
combine mass poverty with good policies and institutions. Why? Because
of the wave of policy reform and institution building that has swept
through poor countries during the 1990s.
Figure 1.13 classifies 113 developing countries based on their inci-
dence of poverty and the quality of their policies in 1996.4
What creates a great opportunity for foreign aid is the upper-right-
hand quadrant. Thirty-two countries have poverty rates above 50 per-
Figure1.12 Allocation
of MultilateralAid,1970-93
Aid as a percentage
of GDP
3
01
Lower-middleeGood
incorne ~~~~~~~~~~~~manageme
Poor
Low management
Real GDPper capita
44
MONEY MATTERS-IN A GOOD POLICY ENVIRONMENT
Figure1.13 Poverty
andPolicy,
113Developing
Countries,
1996
~ o ~~ 03 ' 0o b9
0 0 . 4
III IV
0o 0
Poor policyGood polic
LowpovrtyLowpovrt
lll ~% lV
Dlar198
So reColiran
0 00~~~~~~~~~~~~~~~4
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Notes
1. Allthreeofthesestudiesconsiderwhethergoodpoli- 1991).In low-incomecountries,this measureof GDP is
ciesbringabout growthor growthspursgoodpolicies- often three timeshigher than GDP at domesticprices
andconcludethat policyreformleadsto moregrowth. becausepricesof nontradedgoodstend to be low in poor
countries.In the developingworldthe averageaid receipt
2. Collierand Dollar1998showthat the findingthat relativeto PPP GDPisabout2 percent.The sameamount
aid hasmoreimpactin a goodmanagementenvironment wouldbe 6-7 percentof GDP in domesticprices.
holdsin the 1990susinga muchbroadermeasureof insti-
tutionsand policies,includingissuessuchas corruption, 4. The measureofpolicyusedto constructthisfigureis
governance,equity,andsafetynets.For researchpurposes, the CountryPolicyand InstitutionalAssessmentof the
this broadermeasurehas the disadvantagethat it is not WorldBank.It hasmorecomponentsthanthe indexin the
availablebackthroughtime. Burnside-Dollar study,includingsuchareasassocialsector
policiesand safetynets.It has the advantageof coveringa
3. For the growthanalysis,aid is measuredrelativeto largenumberof countriesand beingup-to-date.But it is
GDP at purchasingpowerparity (Summersand Heston not available asa consistentmeasuregoingbackin time.
46
CHAPTER 2
47
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
48
AID CAN BE THE MIDWIFE OF GOOD POLICIES
49
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
8 Ownership-WhatMoneyCannotBuy
50
AID CAN BE THE MIDWIFE OF GOOD POLICIES
Four, Kenya pulls a placatory rabbit out of the hat. Five, the donors
are mollified and the aid is pledged. The whole dance then starts 2
again" (August 19, 1995, p. 37).
Policy
structural adjustment lending from the World Bank affected the policies Source:Dollar and Easterly 1998.
of recipients "a little, but not as much as the Bank hoped" (Mosley and
others 1995). The main problem was that the World Bank had strong
incentives to disburse funds-and thus was inclined to see a good effort
even where there was none. In fact, in the authors' sample of adjustment
loans, only 53 percent of loan conditions were met. Even so, almost all
adjustment loans were disbursed. Zambia, for example, received 18
adjustment loans between 1966-69 and 1990-93 while its policies got
worse (see figure 2.1). In Kenya the World Bank provided aid to support
identical agricultural policy reforms five separate times. Each time
reforms were either not implemented or later reversed. Yet all adjustment
loans were disbursed. The lesson? A conditional loan is no guarantee that
reforms will be carried out-or last once they are.
Yet, adjustment lending from the IMF and the World Bank has sup-
ported many successful reform programs. Bolivia is a good example of a
51
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
andFailed
ofSuccessful
Table2.1 Features Programs
Adjustment
52
AID CAN BE THE MIDWIFE OF GOOD POLICIES
The study also examined factors under the World Bank's control,
including: size of the adjustment loan, number of conditions, resources
used to preparethe loan, and resourcesdevoted to analyticalwork in the
four yearsprior to the loan. It found that these "Bankeffort"variablesare
similar for successfuland failed adjustment programs (see table 2.1). In the past the World
When all the variablesare tossed into the analysis,it becomes clear that Bank did not sufficiently
successful reform depends primarily on a countrys institutional and take into account that
political characteristics. success or failure of
In the past the World Bank did not sufficientlytake into account that reform depends largely
successor failureof reform depends largelyon a countrys own effort. A on a country's own
case in point: Zambia. In the 1980s the Bank approved four structural effort.
adjustment loans totaling $212 million, and they were almost fully dis-
bursed. After completion, the World Bank rated three of the four loans
as failures:that is, reforms supported by the loans were not satisfactorily
implemented. The Dollar-Svenssonresults suggestthat this was largely
predictable. Conditions in Zambia at the time were not conducive to
reform.The government, not democraticallyelectedand in power for a
long time, was likely a nonreformer.It may have been worth taking a
chance on the first adjustment loan, but (admittedlywith 20/20 hind-
sight) a successionof policy-basedloans for Zambia was not the best use
of resources.Donors havegraduallylearned this lessonand becomemore
selectiveabout providing policy-basedassistance.In 1990-95, the suc-
cessrate of World Bank adjustmentloans increasedto 76 percent (World
Bank 1997a).
These studiessuggestthat policyreform depends largelyon countries'
institutional and political features. Borrower ownership of reform is
increasinglyrecognizedas a prerequisitefor success.But once a serious
reform program is started, financial assistancecan help consolidateit.
Fomenting
Reform
53
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
perhaps the most important issueon their agenda.At the same time, it
is not easy to generate reform, and conditional lending into weak insti-
tutional and policy environments has failed. So, can donors promote
policy reform in the highly distorted environments of, for example,
Myanmar, Nigeria, or the Democratic Republic of Congo? Or do they
simply throw up their arms and walk away?
Clearly, to desert difficult countries is unacceptable. As poverty is
reduced in developing countries with sound policies, countries with
highly distorted policieswill account for an ever-increasingshare of the
poor. What is the answer?There is evidencethat donors can make a dif-
ferencewithout large-scalefinancing.Intangibleand low-costeffortscan
promote policy reform over the long term-by disseminatingdevelop-
ment ideas, training the next generation of leaders,and stimulating pol-
icy debate in civil society.
Development agencieshave played an important role in disseminat-
ing information about what is good policy, both at the macroeconomic
level and in individual sectors such as education, health, and pension
reform (box 2.1). But good policy is not somethingsubjectivelydecided
by the World Bank; lessonsemerge from the experiencesof developing
countries. Good management is, objectively,what has increasedgrowth
and reduced poverty in those countries.
Some key developmentpolicies (macroeconomicstability and open-
ness) have become known as the "Washington consensus." This is an
ironic designation because in the 1970s agenciessuch as the World
Bank were not promoting these policies. In fact, the Bank was taken
t0report,2
donrs}hdave
helped mmaycutiies
study0thet way for the governmettto convineegthe pubi that
kingWS-term
fis&alCan dstrtibutioniaBl
cn jncesof 00kSrefrm isAuessay.SX 00000fF00X0
tUW00k'l..;s 00000
54
AID CAN BE THE MIDWIFE OF GOOD POLICIES
55
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
56
AID CAN BE THE MIDWIFE OF GOOD POLICIES
57
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
58
AID CAN BE THE MIDWIFE OF GOOD POLICIES
tance can boost investor confidence.Collier points out that the way in
which policy-basedlendinghas beenmanagedtendsto undermineits use-
fulnessas a restraintand signalingmechanism.These loanshavenot been
usefulas restraints(sincefunds have been disbursedregardlessof whether
reform has been carried out) and have not been good signalsof the seri- To be effective,
ousnessof reform (before1990 one-third of adjustment loans failed). adjustment loans need to
The role of agenciessuch as the World Bank is not to arm-twist gov- focus on fewer, but truly
ernments to do what they are reluctant to do. Nor is it true that if the important measures.
World Bank works harder or puts more resourcesinto an adjustment
loan, a failed reformer can somehow be turned into a successfulone.
Instead, the role of international institutions should be to disseminate
information that might influencepublic dialogueabout policyreform-
and to learn to read the signalsabout whether governmentsare seriousor
not. Mistakesare inevitable,but the successof adjustment loans should
be closerto 85 percent than 67 percent. This target could be achievedby
being more selectiveabout the environments in which donors become
involvedin financing adjustment.
Collier arguesfurther that for this approach to be effective,adjustment
loans need to focus on fewer, but truly important measures. (In the
Dollar-Svenssonsample the mean number of conditions is 46, and the
maximum 193.) For such loans to be effectiveas a restraint,there has to
be clear agreementbetween government and donor agenciesabout what
is important and the grounds for nondisbursement.To be effectiveas a
signal,loanswould haveto havea high successrate (say,85 percent).They
would then be accurate signalsto the private sector that the government
receivingthe loans is seriouslycommitted to reform.
59
CHAPTER 3
60
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
the lifetime of a project, but judging what would have happened had
there been no aid.
The first section of this chapter exploresin detail what is meant by
fungibility.We then look at the availableevidenceon severaldimensions Most aid goes to
of fungibility. If a donor finances an investment for education, for projects-and most
instance,what confidence does it have that the resources: project aid goes to
social, economic,and
* Stick with the government rather than finance, say, tax reductions? administrative
* Finance investment rather than expandingconsumption spending? infrastructure.
* Financeeducation rather than some other sector?
Figure ofAid,
3.1 Distribution
Donors cannot be certain that resourcesactuallygo to the area favored byTypeandSector
by the project. So, we examine the implication of fungibility for evalu-
Program
ating projects. Foodassistance
One possibilityis that in practice the actual effect of financing tied 3a/ 6%
to projectsis to raisegovernmentspending exactlyas if generalbudgetary
support were given. This implies that, even when donors are financing
projects, an assessmentof aid's effect as a financial transfer depends on
a judgment of overallgovernment spending-whether expendituresare
allocatedto activitieslikely to promote development and how effective
the spending is. This involves assessingthe rationale for government
spending. Again, what you seeis not necessarilywhat you get, as the key
question is: what would have happened without government
interventions?
What does fungibility mean for analyzing the allocation and effi- Industry and
ciencyof government spending and for managing and evaluatingdevel- other production
7%
opment assistance? In countries with sound policies, appropriate Agriculture
allocations of expenditures, and effectiveservices,donors can provide 12%
large amounts of assistance as general budget support, knowing the
resourceswill be well used. In caseswhere there is agreementabout allo-
cation but efficiencyis low, aid projects and financing should be evalu-.
ated not just as money for a particular project but also for what they Economic
61
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Fungibility?
62
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
or if the item to which the financingis tied is specificand only a tiny frac- Roads
tion of the government budget (for example, project aid made available
for an entirelynewtypeof spending). Sar. Hypothetical
illustration.
Most foreign aid goes to projects in which the donor finances either
a single large investment (such as a highway) or, more often, a cluster
of related investments (rural roads, urban housing, systems of irriga-
tion canals, water supply for a number of villages,programs of school
construction). Yetthe fungibility of aid financing has been recognized
as an issue from the beginning of large-scaleaid.' The analytical pos-
sibility is not new, but what is new is the evidence on fungibility.
Suppose that a donor provides $1 million to a government for an
investment in a specific project in, say, education. The three relevant
questions are:
63
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
DoesAidIncreaseGovernment
Spending?
64
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
40
30
20
10
-10
ODA disbursements Concessional loans Bilateral loans
38 countries 38 countries 46 countries
1971-90 1971-90 1969-80
100
so
-50
14 countries 14 countries Sri Lanka Dominican Indonesia
ODAdisbursements Concessional Republic
loans
65
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
DoesAidIncreaseDevelopment
Spending?
66
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
100
s0
60
40
20
Source: Feyzioglu, S-aroop, and Zhu 1998; Pack and Pack 1990, 1993, 1996.
Figure3.6 BilateralAidandGovernment
Consumption ... most of the rest goes
Governmentconsumptionas a percentage of GDP
to government
consu mp n.
consumption.
40
35
4 4
30
25 4 4 4 4
5 *++ ~ ~ *
0 2 4 4 6 10
Source:
Calculated fromn the Burnside-Dollar data set.
67
ASSE';SING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Empirical studies of Indonesiaand Sri Lanka also find that aid has a
smallereffecton developmentspendingthan on total governmentspend-
ing, sothat not allaid passesinto investment.Again,the differencesacross
countries are large. In Sri Lanka an additional aid dollar raised govern-
ment spending by 52 cents. Of that, 38 cents was devoted to develop-
ment spending and 14 cents to consumption spending. In Indonesia a
dollar'sworth of aid raised spending by $1.58, of which development
expendituresaccounted for 89 cents.
The evidencesuggeststhat, evenwhereaid goesmostlyto government
spending, the spending funded by aid is largelyfungible between con-
sumption and investment, regardlessof whether the aid is administra-
tivelytargeted to specificinvestmentprojects.
DoesProjectAidFinanceParticular
Sectors?
80-t
60
40~~~~~~~a
40
_20
-20
Dominican Indonesia SriLanka 14 countries 14 countries
Republic (public investment) (government
expenditure)
68
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
69
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
1o
10
-I10
Indonesia Dominican Sri Lanka 14 country
Republic sample
Source:Feyzioglu,Swaroop, and Zhu 1998; Pack and Pack 1990, 1993, 1996.
BudgetaryMechanicsof Fungibility
70
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
Table3.1 Donor
Financing
ofInvestment
Budgets
inTwoAfrican
Countries
(percent)
Ghana Uganda
Projectfinancing/capital
expenditures 71 87
Projectfinancing/total
expenditures 27 43
Alldonorfinancing/total
expenditures 32 67
S&urcr:Campos and Pradhan 1996-
71
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
ProjectAssessment
withFungible
Money
theEffect
Table3.2 Evaluating of Projects
withFungibility
Base
case .. A..B......C...
tax &~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~et1 ,
~ on,
~~
IIW'~~~~~~~~~~~~~~r ~ ~ ~ ~ ~ areo
72
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
First, the government agreesthat the donor will finance the highest-
return project (A).Thus the apparent rate of return is 100 percent: that
is, when the donor assessesthe benefits it will find that its aid has been
hugelysuccessful.What reallymatters in evaluatingthe effectof aid, how-
ever,is what the governmentdoes with the resourcesit otherwisewould While the project
have devoted to the project. It could do the next best project not yet approach has some
financed, which would be F (with a return of 40 percent), since it is advantages for donors, it
alreadycommitted to B,C,D, and E. Or the governmentcould chooseto does not ensure that aid
do a white elephant project (K) with political payoffsbut with a return is put to good use.
to the economy of zero. It could also choose to reallocatethe resources
from investment to consumption. What is the return on additional con-
sumption expenditures,high or low?Often consumption spending, such
as that on recurrent inputs or operations and maintenance, is not
financedby donors but has enormouslyhigh returns in keeping existing
investments productive. So a shift from investment to consumption
could have a huge return, say,200 percent. Alternatively,the consump-
tion could be the conspicuous consumption of the elite and have zero
return. Or the aid could havefinancedtax relief,which as a "project"may
have very high returns if it reduces distortionary taxes.
Delivering aid through projects has severaladvantages for donors.
There are administrative advantagesto organizing tasks into discrete,
manageable units. Investment projects fit well with the notion that
increasinginvestment is the key to development.And by financingpar-
ticular items (familyplanning, education, and so on), projectsappeal to
single-interestcoalitionsin donor countries supporting foreignaid.
But while the project approach has some advantagesfor donors, it
does not ensurethat aid is put to good use. Nor does it facilitatethe eval-
uation of aid effectiveness.If, say, aid finances a road project with a 20
percent rate of return, this information tells us next to nothing about
aid effectiveness.Does fungibility imply that projects are irrelevantand
should be abandoned as a way of organizingand deliveringaid, in favor
of broader financing mechanisms?Not at all. Project financing can be
used to conveyother valueadded. In many casesthe project'svalueadded
was never intended to be the investment itself.Donors often are not try-
ing to influence project selection or directly alter the sectoral composi-
tion of spending but are using projects as a vehicle for technology
transfer or to build institutional capacity or reform sectoral policies.
Taking the road example, donor involvement may mean that those
institutions responsible for roads are improved-or road policies are
73
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
changed for the better. Projectsare a cauldron of money and ideas. Since
the money does not necessarilystick with the sector, the design and
evaluationof projectsshould focuson the contribution of ideas to alter-
ing sector institutions and policies (seechapter 4).
Donorsare, more or less,
finanicingwhateverthe
governmentchooses Public notQuantity
Spending-Quality
to do.
Equityand Efficiency
74
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
Quality
andQuantity
The quality of government expenditures is typically more important
than the quantity. While there is evidence about what kinds of public sec-
tor outputs are crucial to growth and poverty reduction (basic education
and infrastructure, for instance), there is often little or no evidence that
increasing public spending would have an impact on these objectives. Just
as the effect of aid differs across countries, so too does the effect of
increased public spending on improving outcomes. Cross-national and
sectoral evidence, as well as individual cases, suggests that inputs like
money are often not the most important issue for public sector outputs.
The empirical growth literature finds that government consumption
has no positive effect on growth. Some studies find a negative effect, others
75
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
76
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
77
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Forma View...
78
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
79
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
80
MONEY MATTERS-IN A GOOD INSTITUTIONAL ENVIRONMENT
81
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Note
1.TheWorldBank'sfirstChiefEconomist,PaulRodenstein-Rodan, empha-
sizedin the 1950sthat evenwhenfinancingwastiedto projects,moneyoften
remainedfungible.Assessments of aid as earlyas the 1960sraisedthe issueof
fungibility(Littleand Clifford1965).
82
CHAPTER 4
83
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
BefterPublicProvision
84
AID CAN BE THE MIDWIFE OF GOOD INSTITUTIONS
Accountability
85
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
BoxK
4 I Aidu et'lzto
"STREGTHEING
HE R,E O LOCl GOERNMENT, bcLDnrgeisarenrnomly t
esp&i$llylini te dilY ofPbic rvics-for exam-0 lcd ir4c+
d tlytolcalovrnens withuti
pie, heahs&iA tei 4iciiturextension,a1nd tee frotheen al ge t. But pro 'idissisuch
rural-ifatuture>0 provis ion,0 al0lof which hae all
0><0;$ -agarne run cut t;0^f
- prain.:
th<z :0 g < , oA:
jo ivi
beenSdeole
Win tah Phalilipies-i iantitiel deasiri000nrespon; s ibilityC toite,kloal goenmn.I th b0e c
;tin <prac0ic4~~iZ44O
0tablie.But skd+ pb-g
hlorE satisfactoryloan icondu <iring
of000j mechanBisB)ji~n
fit locali0
86
AID CAN BE THE MIDWIFE OF GOOD INSTITUTIONS
changeson donor agenciesfrom project to process:the fund must specify Source: Isham,Narayan, and Pritchett 1995.
87
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
88
AID CAN BE THE MIDWIFE OF GOOD INSTITUTIONS
EffectivePublicInstitutions
89
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
90
AID CAN BE THE MIDWIFE OF GOOD INSTITUTIONS
91
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
PublicProvision
withoutPublicProviders
92
AID CAN BE THE 4MIDWIFE OF GOOD INSTITUTIONS
93
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
NGOsas ServiceProviders
Box'
B4.:2:i A Useful' Oaizationst T W
Su'rgeiin tNongoermentalt k ii l tRi S jE S
A RECENTREPOI4T JAPx4sOERSES EONOMtCai .SNG. theJapan WdildBlird Society o u ami:
Cooperatiton Fud(EF j:,O ntsta ,ES ped
ii ih
envronmenalsaurveon itheiihait ofcansa
jjofjdeMiocacyanSdgthe' ibroaieirpoitical:pariciptRa-iij patt of the OEW:CFsj;
jspecial assi;iistance fi;
p Sijec
jiii'iontha hazscomewith thei end,iof th ColU W ar fomto fo the.: DeveopmenSitAiii Pr jeOfii
94
AID CAN BE THE MIDWIFE OF GOOD INSTITUTIONS
creating sustainability and local ownership (World Bank 1998 a). So, while
NGOs can be part of the service delivery system, they cannot replace gov-
ernment and cannot be a permanent substitute for public sector capacity.
UserFees,Effectiveness,
andDemand-side
Financing In recent years donors
have financedinnovative
Governmentservices(particularlythose financedby aid) are often pro- ways to increase service
vided free to users-health care, irrigation, water, extension services, effectiveness.
schooling, roads, and so on. In many cases this approach is socially desir-
able but raises three problems-two serious, one less so. Least important,
free services tend to be overused, clogging the system-be it a road net-
work or health clinic. The two more severe problems are institutional and
closely related to aid. First, since the service does not generate revenues,
its provider depends on the general treasury for funding. Especially in
poor countries with chronic or recurrent fiscal crises, that is cold com-
fort. There is little assurance that funds needed for, say, medicine or irri-
gation maintenance will be forthcoming. Second, because funding comes
from the treasury, it is difficult to empower beneficiaries or instill in them
the sense of ownership so necessary for the success of aid projects. Users
are frozen out, in setting priorities and planning and delivering services.
In recent years donors have financed innovative ways to increase service
effectivenessby creating charges directly linked to improvements in quality
-from irrigation to education, health to agriculture. Health is a good
example. Providing universal health care in poor countries is a laudable
aim. But because of insufficient funding, morale is in the depths, and more
important, clinics often have no needles, drugs, or medicine. That leads
to a vicious cycle: patients, knowing the poor quality, seek care elsewhere,
leading to underuse of clinics, which makes it harder to justify increased
budgets to improve services (Filmer, Hammer, and Pritchett 1998).
In 1988 the Bamako Initiative sought to reverse that cycle by intro-
ducing user fees in Cameroon. A project financed by the U.S. Agency for
International Development carried out a controlled evaluation of the
combination of charging user fees and using them to improve quality by
creating a reliable drug supply. The study found that not only did the use
of health centers increase but that there were proportionately more poor
users than rich (Litvack and Bodart 1993). Similar experiments have had
similar results in other sectors, such as education (Birdsall and Orivel
1996). Examples like these reveal the valuable role donors can play in
financing and evaluating innovations in service delivery.
95
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
MakingAidWorkfor BetterPublicServices
ProjectFinance
We have not yet discussed a topic that accounts for most of the analy-
sis of aid effectiveness: the technical and administrative details of making
aid projects more effective. There are two reasons for this. First, and most
important, this area has been well covered in many past evaluations,
reviews of experience, and research, and we have little of value to add.
Many donor agencies review the performance of completed projects. The
combination of individual project experience and comparisons across
countries allows these evaluations to provide critical and constructive sug-
gestions from the lessons learned to improve the organization's opera-
tions. In addition, some donor agencies' technical departments and
evaluation departments produce sector specific reviews of both the prob-
lems faced and promising approaches in each sector.1 This voluminous
literature on project performance and sectors provides a wealth of infor-
mation needed to implement successful projects.
The second reason that we do not go further into the issue of effective
aid projects is that we are more interested in the indirect than the direct
effects of projects. Aid-financed projects can have spillover effects from
individual projects to the general operation of the public sector.
Almost from the beginning, projects financed by foreign aid were
thought to have important indirect and broader benefits. Hirschmann
(1967) argued that perhaps the greatest were institutional and personal
learning from the conception, planning, and implementation of projects.
He believed that projects were the "privileged particles" of development
that created innovation by confronting problems as they arise. These ben-
efits emphasize the value added of donor involvement, which allows a
supportive "infrastructure" of institutional and technical support for
implementation that would otherwise not be present.
96
AID CAN BE THE MIDWIFE OF GOOD INSTITUTIONS
97
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Finance
Nonproject
mpnrecigp,fl .. sa ccuesta:
e oAresettki
ol .tlmen i geerlla
wis o no cirie
n hv oi4ettiwell!
.
j roe
pte ;p Ban misin or 0the4
t ireadls S+ of;:t frquneofi
ment Mne" niain f0irst
bementhe
Having00 resultf
'r 'te Bakc atlyi 0thr0et.0l
renranol'evelotsmewnt0l0Bank ats0| tosuspndiXasburem
g0gggjggg
teman.0gngEmS theBauk00 Ban hso b enarmoe
"The00<)g ffciveSover
98
AID CAN BE THE MIDWIFE OF GOOD INSTITUTIONS
99
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
100
AID CAN BE THE MIDWIFE OF GOOD INSTITUTIONS
Methodsof Assistance
andPublicSectorEfficiency
101
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Notes
1. Evena partiallistof the literatureproducedby justthe WorldBankgives
someindication.FromtheOperationsEvaluation Departmenthavecomereviews
of,amongothersectors,irrigation,agricultural extension,and adjustmentlend-
ing.Sectoralpolicypapersreviewing withBankinvolvement
the experience have
been producedon, amongothers,energy, water,education, and adjustment
lending.
102
CHAPTER 5
103
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
that has improved the livesof people living in countrieswith weak gov-
ernance and poor policies.In some casesthis assistancehas contributed
to a larger transformationthat made a major difference.
Here we present four casestudies of effectiveaid under difficult con-
In difiFicultenvironments, ditions. Running through them are four themes:
effective assistance is
more about ideas than * Find a champion.Countries, governments,and communitiesare het-
money or projects. erogeneous.While it is fair to characterizeBurma overall as "poorly
managed,"there are likelyto be reform-minded elementsin the com-
munity and evenin the government. If aid can find and support these
reformers,it can havea big impact.
* Havea long-termvisionofsystemicchange.Successfulreformershave a
vision of how things could be different in, say, 10 years-different
both in outcomes(more kidsgoing to school,graduating,gettingjobs)
and in process(communityinvolvementin schools,broad public sup-
port for reform policies).
* Supportknowledgecreation.While reformerstypicallyhavea long-term
vision, they often need to developthe detailsof reform through inno-
vation and evaluation. Moreover,for reform to take root requires a
demonstration that it works. Financingand evaluatinginnovations is
a key role for developmentassistance.
* Engage civil society.In highly distorted environments the govern-
ment is failing to provide supportive policies and effectiveservices.
That is why government-to-government financial transfers pro-
duce poor results. Effective aid in such an environment often
involves supporting civil society to pressure the government to
change or to take service provision directly into its own hands (or
to do both).
104
MONEY, BUT MORE IDEAS, TOO
Vietnam:Adjustmentwithout Lending
105
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
support from the IMF or the World Bank becauseof its politicalestrange-
ment from the major shareholdersof these institutions. The two were,
however, contributing some staff time to advise the government.
Furthermore,UNDP and SIDA invited the Bank and the Fund to man-
Vietnam's reform opened age some of the technicalassistancethey financed.So, the IMF managed
more space for civil UNDP-financed technical assistance for the central bank and con-
society, and donor tributed some of its own staff time. In a similar arrangement the World
activities encouraged Bank managed UNDP-financed assistance for the State Planning
this aiswell. Committee (now the Ministry of Planning and Investment).
In the early stages of reform this partnership of agenciesprovided
important inputs. In 1991 the UNDP and the World Bank organizeda
meeting in Kuala Lumpur at which Vo Van Kiet and his team met with
economicministersfrom Indonesia,the Republicof Korea,and Malaysia.
Each country laid out some key policiesthat had worked for them, and
the Vietnameseteam asked detailed questions about stabilization,trade
liberalization,foreigninvestment,and other economicpolicies.At a more
technical level, the same group of donor agencies organized training
courses,policy workshops,and study tours for staff from Vietnam'seco-
nomic ministries.
All this was aimedat helping Vietnamlearn about its neighbors'expe-
riencesand developthe knowledgeit neededto manage its own economy
successfully. The generationof knowledgewassupported at a more micro
level,too. UNDP,SIDA, and the World Bankteamed to help the govern-
ment carry out the first representativehousehold survey ever done in
Vietnam-generating a wealth of information about household produc-
tion, incomeand poverty,education, health,and child nutrition. This sur-
vey,conductedin 1992-93, revealedthat 55 percentofthe populationlived
belowa widelyused internationalpovertyline (Dollarand others 1998).
Vietnam'sreformopened more spacefor civil society,and donor activ-
itiesencouragedthis as well.Twoexamples:The AsiaFoundation and the
World Bankorganizedthe firstpublic meeting between the domesticpri-
vate sector and leading officialsto discusseconomicpoliciesand keydis-
tortions holding back private investment. And a group of donors
organized rural workshops to discussthe poverty analysisthat emerged
from the household survey and how donors could support local initia-
tivesaimed at poverty reduction.
It is impossibleto measure the exact effect of these donor activities.
But there undoubtedly has been dramatic improvement in economic
policies in Vietnam and that has made people's livesbetter. Vietnam's
106
MONEY, BUT MORE IDEAS, TOO
Poor 0
1988 1989 1990 1991 1992 1993 1994 1 995
107
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
ElSalvador,
Pakistan,andBrazil:Education's
andReform
Decentralization
G OVERNMENTSIN DEVELOPINGCOUNTRIESUSUALLY
PLAYA MAJOR
role in the allocationand managementof educationalresources.
- Even when some authority is delegatedto provincesor munici-
palities, individual school administrators and parents have a limited
voice. This centralized approach has supported many achievementsin
education, but it has not alwaysreached groups that have traditionally
had low levelsof education (the poor and girls, for instance).And over-
centralizationcan stifleinitiativeamong those critical in affectingschool
outcomes-teachers, principals,and parents.
In recent years many developingcountries have experimented with
decentralization of education in two ways: first, allowing more local
autonomyand input into public schools,and second,creatingmore space
for the nongovernmental sector. In some cases donors have helped
finance and design these innovations. Most important, there has been
support for government efforts to evaluate these innovations seriously,
followingpilot groupsand comparingthem with control groupsnot par-
ticipating in the pilot. For reforms in El Salvador,Pakistan,and Brazil,
there is now solid evidencethat they work.
El Salvador'scommunity-managedschoolsprogram-more popularly
known by the acronym, EDUCO, or Educacioncon Participationde la
108
MONEY, BUT MORE IDEAS, TOO
109
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
110
MONEY, BUT MORE IDEAS, TOO
Cameroon:
HealthFinancing
andDelivery
111
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
112
MONEY, BUT MORE IDEAS, TOO
Africa'sRoadMaintenance
Initiative
A FRICAKS
ROADSECTORPROVIDESGOOD EXAMPLES
OF HOW MONEY
-in a weak institutional and policy environment-can have lit-
tle impact. Almost a third ofAfrica's $150 billion in roads (much
of it aid-financed) has been lost due to lack of maintenance. Entering the
1990s about half the region's paved roads and 70 percent of the unpaved
roads were in fair to poor condition. The poor state of the transport net-
work was a serious bottleneck that lessened the value of reforms or other
public investments.
Institutional and policy weaknesses prevented governments from pro-
viding necessary road maintenance despite ample funding. In most coun-
tries there was an almost exclusive reliance on public road maintenance
companies, which tended to have a large capital stock (equipment) at low
levelsof use. The public sector companies were also overstaffed. Thus funds
were directed at equipment and employment, but often were not available
113
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
114
MONEY, BUT MORE IDEAS, TOO
Road boards have also provided input into other institutional changes,
moving away from exclusive reliance on the public sector to carry out
road maintenance. There now are commercially run equipment pools,
and both road maintenance tasks and equipment maintenance are being
contracted out (Carapetis and others 1991).
One key lesson is the importance of having a long-term game plan for
reform. The initiative began in 1987, but it came to show major results
only in 1992. It also demonstrates the importance of "knowledge creation,"
as we use the term in this report. In highly distorted environments people
stop expecting services from the public sector and become skeptical about
possibilities for reform. What is needed in this situation is not just good
ideas, but demonstrations that they work. Now that there is tangible evi-
dence of their value, good ideas from the African road maintenance initia-
tive are spreading quickly from country to country. Also needed for the
reform of public institutions is the participation of civil society.
Conclusion
OVERTYHASALWAYS
BEEN WITH US. DEVELOPMENTASSISTANCE
IS
only 50 years old. It is encouraging that there have been huge suc-
cesses in the aid business-and it is not surprising that there have
been some spectacular failures. Many failures involved investments
115
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
116
MONEY, BUT MORE IDEAS, TOO
117
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
inthe l9i90*
Bo 5.1 refnnng AidAgenciZess WX:S0Xs
RSRS0XS;;0S
W6
Ai0MAORV CHLLNE FOR1AID AGENIES I1N tRECENT0 goals.Inlstitutionlfctrs critica forsustaii a1-devel
been to iagdapt
years hG3asX therr institutoa strtgies,0 opan t lncgimpact had been ineglected.0 Cges areno
c tllres, andi ctapctietls t towor with participatoryo .9lnder u wa ingeahof thiemultilaterals,gand s o e twide
New parterisg
dg0everlCop}fmenlt. jof
Sstalg,
grec ruitment,lapp esH are rel
lroa4t wacinwg individalg proj,tal-Te W
ant;al0personnlli
perorm3;ance; 'assessmient, nllewprjc&ti i Worl
0 Banikhas emboied ts nlewobjEetives an orga-
managemtentapproachlest, ~dne<w <critrita for)judg.ing ga
nizatinnalcl,tureinpa,tnXshp istrategy thait~ahes
S succes-l haeemerge progressvelydHowfarShas 0ourtoconterparts iwithiinand outside govenet ;0i
thislprocesgoe, ndw hat remWains to bedone? Simiiiiiiiiiilar,tFchangesare evident inbiatra ;>s, a;;
* matilateadelopmn hanks,
self comisiwh
hOr havs,oten b hasem thiesC]ur
sto
sned reorts led t afurndnenta recounsideraion jofR th0
Weir maininstrumenathas b3eergrants no lais, wth;
Wthefetiveess oftheirtraditionarlintruments-pro- mr e flexbibilityin
prprton iandtchial ssrsra
Ject
lons Beinn ~th ti;fheWold Bdank' The empasis ha hiftEWe from capitaSlprets t
Wap)enhanis Repot, flE;lolwedbysiilreprts of
iitcNac buMiingxi an upr for
poicy rfoi
Intet-America2
**n 0tDeve; , tlotnet B!ank,l the Asianlltltl 00 In allgthese can;gles tefoclWusiIson dev toent
Developmnirt ;Bank,andthe Afican D;l:evelopmfenitft 'Urelus-on thei human progress an thetoomic;,iC
Ei.atikLthe conlsion gwasthtla %qtuat hadbee00
jity 0Pt 2 0social adanegeneraedby th furs oft
giv.Men
00primc over. loanBquality. Seur.ing oan the deeloping conries andther prnrs'.' , Bv, ; gi
. waosia
morepiowefeti 'motivaror forstaffthani S ogE000SV Sf S0 ii tSlSfMXSxs<i 0EA
woringto nsurWprect siicesseorlarge de r elomet S AssitaeetCommitteest
uce Dimopmenr aff S>SL(200W
must fir into a larger game plan for policy reform and institutional change.
In the same vein, the focus of evaluation has risen above the levelof the pro-
ject to overall country program reviews. These essentially ask: Have agen-
cies used their resources to stimulate institutional and policy changes that
have led to improved services and a better quality of life?This is not an easy
question to answer. But it is the right one to ask. The Overseas
Development Council recently sponsored reviews of the overall impact of
aid in eight African countries, conducted joindly by scholars from the devel-
oping countries in question and from major donor countries-which
seems to be a fruitful approach. In the same vein, the shareholders of the
IMF recently turned to a group of outside scholars to assessthe Fund's sup-
port to low-income reformers. And the World Bank has initiated country
assistance reviewswith input from a wide range of stakeholders.
Called for are independent reviewsof development agencies with strong
input from developing countries and focusing on two questions: Has the
bulk of financing gone to sound institutional and policy environments?
And have agencies contributed to policy reform and institutional change?
Evaluating the right things should feed back into the management and
118
MONEY, BUT MORE IDEAS, TOO
119
Appendix 1
Estimating the Effect of Aid on Growth
122
APPENDIX I
123
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
'0
mana00000ge
Eonom;<C1SXfic L03bS
0.500A0giH00iEiE-i
i00inStgf0k2t4 tl 0 ,7 9 b00 t2Ob
000iSl0.01000 ; ' 2058aU
.Tradc
openness~~~~~~~~~~~.......
~
Inibtina ~ ~ ~ M 2
,MAMdGEWM M%
Instititton.quM~ M5M.
igS,MXS1'[',2.S,,EX,1Cliyle,'(27,'17)'',' 1 (70I .1 '. ;L3 (060 :{~ ;34)(.0:
Hm~~~~~~~~~~~3
Govrnmntcnsu.prori -2
124~~~~~~~~~2
~
124 ~ ~ ~ ~~t~551. 05). 068 02) (.51 (.41 02)>1241
APPENDIX I
TableA1.2 Countries
inthe Dataset
MiddleEastand
S.b-SaharanAfiica LatinAmerica NorthAfrica EastAsia SoutbAsia
Botswana Argentina Algeria Indonesia India
Cameroon Bolivia Egypt Korea Pakistan
Coted'lvoire Brazil Morocco Philippines SriLanka
Ethiopia Chile Tunisia Thailand
Gabon Colombia Syria Malaysia
Gambia CostaRica Turkey
Ghana Dominican
Kenya Republic
Madagascar Ecuador
Malawi El Salvador
Mali Guatemala
Niger Guyana
Nigeria Haiti
Senegal Honduras
Sierra Leone Jamaica
Somalia Mexico
Tanzania Nicaragua
Togo Paraguay
Zaire Peru
Zambia Trinidad
Zimbabwe and Tobago
Uruguay
Venezuela
125
Appendix 2
Explaining the Successor Failure of Structural
AdjustmentPrograms
126
APPENDIX 2
127
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
128
APPENDIX 2
TableA2.1 Estimating
theOutcome
ofAdjustment
Loans
Reessaios
- 1 2 3 4 5 6
Dependent vasiabke Outeome Outcome Outcome Preparation Supervision Supervision
Regressiorm
iesod ... Probit Probit Probit/lVa OLS OLS Iva
Observations 220 182 179 179 179 .179
Countries 67 60 . . 60 60 60
Constant ' -0.098 -0.762 --0.366 3.311 2.685 3.272
(0.32) (0.72) (0.25) (438) (4,02) (3.11)
Ethnic fractionalizarion 5.930 8.176 7.763 0.D18 -0.134 -0.144
{4.16) (4.40) :(404) (0.04) (OA2.1 (0.46)
2
Ethnlicfactionralization -6.513 -8:501 ,8.:046 0J043 0.213 0.254
(4.27) (4.32) (3.79) (0.10 (0.59) (0.73)
Government crisis -1.301 -2.372 -2285 -0.223 -0.029 -0,017
(3.94) (4.46) (4-29) (2.48) (0.39) (0.18)
Democraticaly elected 0.585 0.887 0.912 0.124 -6.1E-3 -0.009
(2.61) (3'.il (33)9) (1.98) (0.0) (0.18)
Tine in power .. 089 -0.118 -0.113 0.004 0.003 0.004
(2.07) (2.23) (2.09) (0.36) t( 29 (0.48y
Time in pow0cf 0.003 0.004 0.004 -.-3.7E-3. 3.6E-3 -4.7E-3
(2.15) (2.17) (2.02) (0.99. - (1.141 (1.47)
Preparationstaff weeks(Iogn 0.966 0.323 0.339 0.364
(2.31) (0.24) (5.14) (I.34)
Supervisitin stalfweeks (g) -1.410-O.
(2.92) 0t.67)
Financerconditions (percent) i217 1.423 -0.149 - -0.078 -0.120
-(i.84) . (2.02) (1.07) (0.67) (0.99)
Macro and fiscal conditions (percent) 0.910 0.766 0.260 -0.323 -0-.256
(1.04)4 (0.89) - (1.33) (1-97) (1.40)
Sectoralconditions (percent) 1.386 1.161 0.002 a.180 . .1.75
(2.26) (1i83) (0.02) (1.65) (1.59)
Trade conditions (percent) 13067 0.961 -0021 -0.141 -0.14 I
(1.70) S(i.46) (0.15) (1.25) (1.23)
Number of conditions (percent) 0.153 0.074 0.077
(329) (1.85) (1.28)
Loan sine (log) 0.281 0.210 0.220
(5.29) (4.37) (2.50)
Structural adjustment loan -0.145 -0.062 -0.105
(216) (1,10) (1.58)
Sub-SabaranAfrica -0.08Q 0.093
- 0.78) (1.09)
Latin America& Catibbean -0.284 0.020
(3.06) (0.25)
East Asia -0.148 -0.118
(1.39) (1.33)
InitialGDP.percapita,(log), -0.064 -0.153 - -0.184
(1.04) (2.11 . (3.39)
Initial population (log) -0O147 -0.099 -0. i24
(3.90 (00) . (2.66)
I'recte@biiy. 0.750as07
AuwwSr R2 . .34 0.45
a. Regtession3 is estimated by a trwo-stageproceduredescribed in Dollar and Svensson(1998), with preparation,and supervisionspecifications
given in regressions4 and 6. Regression6 is ctin-ated by 2SLSwith preparation specificationgivenin reresrion4; r-sra istics in pareatheses.
129
Appendix3
Analysisof the Fungibility of ForeignAid
130
APPENDIX 3
131
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
132
APPENDIX 3
133
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
TableJA3.2 Impact
ofSectoral
Concessionary
Loans
onSectoral
Government
Expenditure,
1971-90
.Dependentvari I .2 A061 4 5 9
2
Contan 121. 9-3.07 .... O&mtOME
Strdl.n'sahro D, ,
(1
8) 0.3) 0.0) . 31d(038
1e342 *. -3 . .7 .5 .. 1.0 . 59
Appendix 4
The Impact of CivilLibertiesand Democracyon
GovernmentPerformance
135
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
136
APPENDIX 4
TableA4.1 BaseSpecification
for the NongovernanceDeterminants
of
theEconomic
Rateof Returnof Government Projects,197447
Estimatewithout Estimatewith
regionaldummies regionaldummies
Exogenous variable
In (capital/labor) -1.09 -1.66
(0.067)' (0.060)a
Dummyfor projectcomplexity -4.29 -4.23
(0.0 I 7)b (0.016'b
Termsof tradeshock 0.0015 0.001
(0.889) (0.922)
Policyvariable
Blackmarketpretnia -0.046 -0.037
(Oo
.O)b (Q.)00)
Fiscalsurplus 0.197 0.266
(0.149) (0.063)a
GDP growth 0-193 0.013
(0.357) (0.949)
Regional
dummyvariable
East Asia -3,33
(0.154)
Latin American and the -4.74
Caribbean (0.072)a
Europe, the Middle East, and -4.93
North Africa (0.l00)
Sub-Saharan Africa -10.8
(o,00G)b
Sectoral dummy variable
Agriculture 0.027 1,39
(0.992) (0.602)
Energy and public utilities -3.92 -3.18
(0136) (0.220)
Transport and tourism 3.85 6.24
(0.137) (0.016)b
Urban 10.1 11.9
(0.01 1)b (0 003)b
a. p-levellessthan 0.10.
b. p-level lessthan 0.05.
Note:We report p-levelsof the testwhether the coefficientis 0 rather than test statistics them-
selves.The p-levelis the significancelevel at which the null hypothesiscan be rejected, hence
a p-level lessthan 0.05 indicates a rejection of the nlull hypothesisat (at least) the 5 percent
level. Tbe p-levelsate in parentheses.The sample size is 761.
137
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
civil lSDgiberXt
ieisS0
, 198S7SQOOO000ia0
Note
a$ampelle sitesarei>))-649Wt0
ford
hetAglA
Freedo
f House.X
civiiberf Uind~ addU236iDj,
tties W;g} jEfor theW
138
Appendix 5
Estimating the Impact of AnalyticalWork
139
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Issues,Approach,
andData
Results
140
APPENDIX 5
141
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
142
Selected Bibliography
Alesina, Alberto, and David Dollar, 1998, "Who Gives Burnside, Craig, and David Dollar. 1997. "Aid, Policies,
Aid to Whom and Why?" NBER Working Paper and Growth." Policy Research Working Paper 1777.
6612. National Bureau of Economic Research, World Bank, Development Research Group,
Cambridge,Mass. Washington, D.C.
Amemiya,T. 1978. "The Estimation of a Simultaneous . 1998. "Aid, the Incentive Regime, and Poverty
Equation GeneralizedProbit Model." Econometrica46. Reduction." Policy Research Working Paper 1937.
World Bank, Development Research Group,
Bauer, P.T. 1971. Dissent on Development. London: Washington, D.C.
Weidenfeld and Nicolson.
Campos, Ed, and Sanjay Pradhan. 1996. "Budgetary
Berg, Elliott. 1993. Rethinking Technical Cooperation: Institutions and Expenditure Outcomes: Binding
Reformsfor CapacityBuilding in Africa. New York: Government to Fiscal Performance."Policy Research
United Nations Development Programme. Working Paper 1646. World Bank, Policy Research
Department, Washington, D.C.
Birdsall,Nancy, and FrancoisOrivel. 1996. "Demand for
Primary Schoolingin Rural Mali: Should User Fees Be Carapetis,Steve,Hernan Levy,and Terje Wolden. 1991.
Increased?"EducationEconomics4(3). TheRoadMaintenanceInitiative:BuildingCapacityfor
PolicyReform.EDI SeminarSeries.Washington, D.C.:
Boone, Peter. 1994. "The Impact of Foreign Aid on World Bank.
Savingsand Growth." London Schoolof Economics.
Cashel-Cordo, P., and S. Craig. 1990. "The Public Sector
Branson, William, and Carl Jayarajah. 1995. Structural Impact of International Resource Transfers."Journal
and Sectoral Adjustment: World Bank Experience, of DevelopmentEconomics32(3).
1980-92. A World Bank Operations Evaluation
Study.Washington, D.C. Cassen, Robert. 1994. Does Aid Work? Oxford:
Clarendon.
Bruno, Michael,Martin Ravallion,and Lyn Squire. 1998.
"Equity and Growth in Developing Countries: Old Chang, Charles, Eduardo Fernandez-Arias, and Luis
andNew Perspectiveson the PolicyIssues."In V. Tanzi Serven. 1998. "Measuring Aid Flows: A New
and K. Chu, eds., Income Distribution and High- Approach." World Bank, Development Research
QualityGrowth.Cambridge, Mass.: MIT Press. Group, Washington, D.C.
143
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Collier, Paul. 1997. "The Failure of Conditionality." In C. Edgren, Gus. 1996. "A Challenge to the Aid Relation-
Gwin and J. Nelson, eds., Perspectives on Aid and ship." In SIDA (1996).
Development. Washington, D.C.: Overseas Develop-
ment Council. Ehrenpreis, Dag. 1997. "The Changing Global Frame-
work." In SIDA (1997a).
Collier, Paul, and David Dollar. 1998. "Aid Allocation
and Poverty Reduction." World Bank, Development El Salvador, Ministry of Education. 1995. EDUCO
Research Group, Washington, D.C. Learns and Teaches.San Salvador: Algier's Impresores,
SA de CV.
Deininger, Klaus, Lyn Squire, and Swati Basu. 1998.
"Does Economic Analysis Improve the Quality of Fallon, Peter R., and Luiz Pereira da Silva. 1995.
Foreign Assistance?" World Bank Economic Review "Recognizing Labor Market Constraints: Government
12(3). -Donor Competition for Manpower in Mozambique."
In Lindauer and Nunberg (1995).
Devarajan, Shantayanan, Vinaya Swaroop, and Heng-fu
ZZou. 1996. "The Composition of Public Expenditure Feyzioglu, Tarhan, Vinaya Swaroop, and Min Zhu. 1998.
and Economic Growth." Journal of Monetary "A Panel Data Analysis of the Fungibility of Foreign
Economics 37. Aid." World Bank Economic Review 12(1).
Dollar, David, and William Easterly. 1998. "The Search Filmer, Deon, Jeffrey Hammer, and Lant Pritchett.
for the Key: Aid, Investment, and Policies in Africa." 1998. "Health Policy in Poor Countries: Weak Links
World Bank, Development Research Group, in the Chain." Policy Research Working Paper 1874.
Washington, D.C. World Bank, Development Research Group,
Washington, D.C.
Dolla:r, David, Jennie Litvack, and Paul Glewwe,
eds.1998. Household Welfare and Vietnam 's Transition. Filmer, Deon, and Lant Pritchett. 1997. "Child Mortality
Washington, D.C.: World Bank. and Public Spending on Health: How Much Does
Money Matter?" Policy Research Working Paper 1864.
Dollar, David, andJakob Svensson. 1998. "What Explains World Bank, Development Research Group, Wash-
the Success or Failure of Structural Adjustment ington, D.C.
Programs?" Policy Research Working Paper 1938.
World Bank, Development Research Group, Fischer, Stanley. 1993. "The Role of Macroeconomic
Washington, D.C. Factors in Growth." Journal of Monetary Economics
32(3).
Easterly, William. 1997. "The Ghost of Financing Gap."
Policy Research Working Paper 1807. World Bank, Fraker, Thomas, Alberto Martini, and James Ohls. 1995.
Development Research Group, Washington, D.C. "The Effect of Food Stamp Cashout on Food
Expenditures: An Assessment of the Findings from Four
Easterly, William, Michael Kremer, Lant Pritchett, and Demonstrations." Journal of Human Resources30(4).
Lawrence Summers. 1993. "Good Policy or Good
Luck? Country Growrh Performance and Temporary Freedom House. 1997. Ereedom in the World: Political
Shocks." Journal ofMonetary Economics 32 (3). Rights and Civil Liberties. New York.
Easterly, William, and Sergio T. Rebelo. 1993. "Fiscal Haltiwanger, John, and Manisha Singh. Forthcoming.
Policy and Economic Growth: An Empirical Investi- "Cross-Country Evidence on Public Sector Retrench-
gation." Journal ofMonetary Economics 32(3). ment." World Bank Economic Review.
144
REFERENCES
Hanushek, Eric, and Dongwook Kim. 1995. "Schooling, . 1978. Labor Absorption in Asian Agriculture.
Labor Force Quality, and Economic Growth." NBER Bangkok:ILO-ARTEP.
Working Paper 5399. National Bureau of Economic
Research,Cambridge,Mass. Jha, Shikha,and VinayaSwaroop. 1997. "FiscalEffectsof
Foreign Aid: A Case Study of India." World Bank,
Heggie, Ian. 1994. Managementand Financingof Roads: DevelopmentResearchGroup, Washington, D.C.
An Agendafor Reform.World Bank Technical Paper
275. Washington, D.C. Jimenez, Emmanuel, and YasuyukiSawada. 1998. "Do
Community-Managed Schools Work? An Evaluation
Hilton, Rita. 1990. "Cost Recoveryand Local Resource of El Salvador's EDUCO Program." Impact
Mobilization: An Examination of Incentives in Evaluation of Education Reforms Working Paper 8.
Irrigation Systems in Nepal." Burlington, Vt.: World Bank, Development Research Group,
Associatesin Rural Development. Washington, D.C.
. 1992. "Institutional Incentives for Resource Jones, William. 1995. The WorldBank and Irrigation.A
Mobilization: An Analysis of Irrigation Schemes in World Bank Operations Evaluation Study.
Nepal." Journal of TheoreticalPolitics4(3). Washington, D.C.
Isham, Jonathan, and Daniel Kaufmann. Forthcoming. Knack, Stephen, and Philip Keefer. 1995. "Institutions
"The Forgotten Rationale for Policy Reform: The and Economic Performance: Cross-Country Tests
Impact on Projects."QuarterlyJournalof Economics. Using Alternative Institutional Measures." Economics
and Politics7(3).
Isham, Jonathan, Daniel Kaufmann, and Lant Pritchett.
1997. "Civil Liberties, Democracy, and the Krueger, Anne O., Constantine Michalopoulos, and
Performance of Government Projects." World Bank Vernon Ruttan. 1989. Aid andDevelopment.Baltimore
EconomicReview11(2). and London:Johns Hopkins UniversityPress.
Isham, Jonathan, Deepa Narayan, and Lant Pritchett. Levy, Victor. 1987. "Does ConcessionaryAid Lead to
1995. "Does Participation Improve Performance? Higher Investment Rates in Low-IncomeCountries?"
Establishing Causalitywith Subjective Data." World ReviewofEconomicsand Statistics.69(1)
Bank EconomicReview9(2).
Li, Hongyi, Lyn Squire, and Heng-fu Zou. 1998.
Ishikawa, Shigeru. 1960. EconomicDevelopmentin Asian "ExplainingInternationaland IntertemporalVariations
Perspective.
Tokyo: Kinokuniya. in Income Inequality."EconomicJournal108.
145
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
LindaLuer,David, and Barbara Nunberg, eds. 1995. Mosley, Paul. 1987a. "Conditionality as a Bargaining
Rehabilitating Government: Pay and Employment Process: Structural Adjustment Lending, 1980-86."
Reformin Africa.Washington, D.C.: World Bank. Princeton Essays in International Finance 168.
Princeton University,Princeton, N.J.
Lindert, Peter. 1994. "The Rise of Social Spending,
1880-1930." Explorationsin EconomicHistory31(1). . 1987b. OverseasAid: Its Defenceand Reform.
Brighton, England:Wheatsheaf.
Little,I.M.D., and J.M. Clifford.1965. InternationalAid.
London: GeorgeAllen and Unwin. Mosley, Paul, Jane Harrigan, and John Toye. 1995. Aid
and Power.Vol. 1. 2d ed. London: Routledge.
Litvack,Jennie, and Claude Bodart. 1993. "UserFeesPlus
Quality Equals Improved Access to Health Care: Musgrove, Philip. 1996. Public and Private Roles in
Results of a Field Experiment in Cameroon." Social Health: Theoryand Financing Patterns.World Bank
Scienceand Medicine37(3). DiscussionPaper 339. Washington, D.C.
Ljunggren, Borje. 1993. "Market Economies under Narayan, Deepa. 1995. The Contribution of Peoples
Communist Regimes:Reform in Vietnam, Laos and Participation:Evidencefrom 121 Rural Water Supply
Cambodia." In B. Ljunggren, ed., The Challengeof Projects.Environmentally Sustainable Development
Reform in Indochina. Cambridge, Mass.: Harvard OccasionalPaper 1. Washington, D.C.: World Bank.
UniversityPress.
North, Douglass. 1990. Institutions,InstitutionalChange
Lopez, Cecilia. 1997. Impacto de la Ayuda Externa en and EconomicPerformance.Cambridge: Cambridge
America Latina, 1972-1992 (Impact of External Aid UniversityPress.
on Latin America, 1972-1992). Bogota: Tercer
Mundo. OECD (Organisation for Economic Co-operation and
Development).1998. DevelopmentCooperation:1997
Lucas, Robert. 1988. "On the Mechanics of Economic Report.Paris.
Development." Journal of Monetary Economics
22(1). -. 1996. Shaping the 21st Century: The Role of
DevelopmentCooperation.Paris.
Marshall, Katherine. 1997. "Vision and Relationshipsin
the International Aid World: The Gambia-World . 1992. PrinciplesforEffectiveAid. Paris.
Bank Kairaba Partnership Forum." HIID Discussion
Paper 602. Harvard Institute for International OECF (OverseasEconomic Cooperation Fund) ofJapan.
Development, Cambridge, Mass. 1996. The Roleof OECFtoward2010. Tokyo.
Mauro, Paul. 1995. "Corruption and Growth." Quarterly OECF of Japan and World Bank. 1998. A New Visionof
Journal ofEconomics110. Development Cooperation for the 21st Century.
Proceedingsof a symposiumheld in Tokyo, September
McCarthy, Desmond. 1995. "Review of Public 1997. Tokyo
Expenditure Work." World Bank, Policy Research
Department, Washington,D.C. Ostrom, Elinor. 1996. "Incentives, Rules of the Game,
and Development." In Michael Bruno and Boris
Moffitt, Robert. 1989. "Estimating the Value of an In- Pleskovic, eds., Annual WorldBank Conferenceon
Kind Transfer: The Case of Food Stamps." Development Economics 1995. Washington, D.C.:
Econometrica 57(2). World Bank.
146
REFERENCES
Pack, Howard, and Janet Rothenberg Pack. 1990, "Is Sachs, Jeffrey, and Andrew Warner. 1995. "Economic
Foreign Aid Fungible? The Case of Indonesia." Reform and the Process of Global Integration."
EconomicJournal100(March). BrookingsPaperson EconomicActivity,no. 1.
1993. "Foreign Aid and the Question of SIDA (Swedish International Development Authority).
Fungibility." Reviewof Economicsand Statistics75(2). 1996. Aid Dependency.Stockholm.
Pritchett, Lant. 1996. "Mind Your P's and Q's: The Cost . 1997b. SIDA LooksForward.Stockholm.
of Public Investment Is Not the Value of Public
Capital." Policy Research Working Paper 1660. Summers, Robert, and Alan Heston. 1991. "The Penn
World Bank, Development Research Group, World Table, Version V." Quarterly Journal of
Washington, D.C. Economics106.
1998. "Patterns of Economic Growth: Hills, Swaroop, Vinaya. 1997. "Experience with Public
Plateaus, Mountains, and Plains." Policy Research Expenditure Reviews." World Bank, Development
Working Paper 1947. World Bank, Development ResearchGroup, Washington, D.C.
ResearchGroup, Washington, D.C.
Tan, Jee-Peng, and Alain Mingat. 1992. Education in
Ranis, Gustav. 1995. "On Fast-DisbursingPolicy-Based Asia: A ComparativeStudy of Cost and Financing.A
Loans."YaleUniversity,EconomicsDepartment, New World Bank Regional and Sectoral Study.
Haven. Washington, D.C.
Ravallion,Martin, and Shaohua Chen. 1997. "What Can Tendler, Judith. 1975. Inside ForeignAid. Baltimore:
New Survey Data Tell Us about Recent Changes in Johns Hopkins UniversityPress.
Distribution and Poverty?" World Bank Economic
Review11(2). Thomas, Vinod, Ajay Chhibber, Mansour Dailami, and
Jaime de Melo, eds. 1991. RestructuringEconomiesin
Riddell, Roger. 1996. Aid in the 21st Century.ODS Distress:PolicyReformand the WorldBank. New York:
DiscussionPaperSeries.United Nations Development Oxford UniversityPress.
Programme, Office of Development Studies. New
York: United Nations. Thompson, Louis S., and Karim-JacquesBudin. 1997.
"Global Trend to Railway Concessions Delivering
Riddell, R.C., A. Bebbington, and L. Peck. 1995. Positive Results." Viewpoint 134. World Bank,
PromotingDevelopmentby Proxy:An Evaluationof the Finance, Private Sector, and Infrastructure Network,
DevelopmentImpactof GovernmentSupportto Swedish Washington, D.C.
NGOs. SIDA Evaluation Report 1995/2. Stockholm:
SwedishInternational DevelopmentAuthority. United Kingdom, Secretary of State for International
Development. 1997. Eliminating World Poverty:A
Rodrik, Dani. 1996. "Understanding Economic Policy Challengefor the 21st Century. White Paper on
Reform."Journalof EconomicLiterature34(1). International Development. London.
Sachs, Jeffrey. 1994. "Life in the Economic Emergency van de Walle, Dominique. 1995. "The Distribution
Room." In Williamson (1994). of Subsidies through Public Health Services in
147
ASSESSING AID: WHAT WORKS, WHAT DOESN'T, AND WHY
Indonesia, 1978-87." In Dominique van de Walle and . 1995. World Development Report 1995: Workers
Kimberly Nead, eds., Public Spending and the Poor. in an Integrating World. New York: Oxford University
Baltimore: Johns Hopkins University Press. Press.
van de Walle, Nicholas, and Timothy Johnston. 1996. . 1996a. "Resettlement and Development: The
Improving Aid to Africa. Washington, D.C.: Overseas Bankwide Review of Projects Involving Involuntary
Development Council. Resettlement, 1986-1993." Environment Department
Paper 032. Washington, D.C.
Wade, Robert. 1995. "The Ecological Basis of Irrigation
Institutions: East and South Asia." WorldDevelopment . 1996b. "Technical Assistance." Lessons and
23. Practices 7. World Bank, Operations Evaluation
Department, Washington, D.C.
White, H. 1992. "The Macroeconomic Analysis of Aid
Impact." Journal ofDevelopment Studies 28. . 1997a. "Annual Review of Development
Effectiveness." Report 17196. World Bank,
Williamson, John, ed. 1994. The Political Economy of Operations Evalution Department, Washington, D.C.
Policy Reform. Washington, D.C.: Institute for
International Economics. . 1997b. World Development Report 1997: The
State in a Changing World New York: Oxford
Worldl Bank. 1992. "Effective Implementation: Key to University Press.
Development Impact." Report of the World Bank's
Management Task Force. Washington, D.C. . 1998a. Global Development Finance 1998.
Washington, D.C.
-. 1993. World Development Report 1993: Investing
in Health. New York: Oxford University Press. . 1998b. World Development Indicators 1998.
Washington, D.C.
-. 1994. World Development Report 1994:
Inftrastructure for Development. New York: Oxford
University Press.
148