Oya CelasunOya Celasun

International Monetary FundInternational Monetary Fund

Jan WalliserJan Walliser

The World BankThe World Bank

Predictability of Aid: Do Predictability of Aid: Do Fickle Donors Undermine Aid Fickle Donors Undermine Aid Effectiveness?Effectiveness?

22

Would you like to be the Finance Minister of Mali?

Mali

-3

-2

-1

0

1

2

3

19

93

19

94

19

95

19

96

19

97

19

98

19

99

20

00

20

01

20

02

20

03

20

04

20

05

Actu

al m

inu

s P

roje

cte

d B

ud

ge

t

Aid

/GD

P

Mali

0102030405060708090

100

19

96Q

1

19

96Q

4

19

97Q

4

19

98Q

4

19

99Q

4

20

00Q

4

20

01Q

4

20

02Q

4

20

03Q

4

20

04Q

4

20

05Q

4

Pe

rce

nt

of

bu

dg

et

aid

dis

bu

rse

me

nts

33

OutlineOutline

Why does predictability matter?Why does predictability matter? Measuring PredictabilityMeasuring Predictability OECD DAC DataOECD DAC Data Predictability in IMF Program DataPredictability in IMF Program Data How do Countries Adjust?How do Countries Adjust? Summary of FindingsSummary of Findings Policy ImplicationsPolicy Implications

44

Predictability ≠ VolatilityPredictability ≠ Volatility Predictability: Difference between expected Predictability: Difference between expected

and actual aid flows for a given time periodand actual aid flows for a given time period Volatility: ex-post measure of fluctuations in Volatility: ex-post measure of fluctuations in

aid flowsaid flows Volatility in aid may not be a problem if it Volatility in aid may not be a problem if it

offsets other economic fluctuations, some offsets other economic fluctuations, some evidence for that from Chauvet and evidence for that from Chauvet and Guillaumont (2007)Guillaumont (2007)

Predictability the more appealing concept to Predictability the more appealing concept to measure “surprises” but harder to measuremeasure “surprises” but harder to measure

55

Predictability vs. Volatility in Africa

66

Effects of Low PredictabilityEffects of Low Predictability

Low-income countries cannot access Low-income countries cannot access international capital markets, no international capital markets, no incentive for large precautionary incentive for large precautionary savings (see Deaton, 1991)savings (see Deaton, 1991)

Lower-than expected aid flows may Lower-than expected aid flows may result in sharp expenditure result in sharp expenditure adjustments (e.g., lower spending on adjustments (e.g., lower spending on investments)investments)

Larger-than-expected aid flows may Larger-than-expected aid flows may not be spent productively (e.g., not be spent productively (e.g., higher consumption)higher consumption)

77

Aid is less predictable in the Aid is less predictable in the poorest countriespoorest countries

02

46

81

0|D

isbu

rse

me

nts

-Com

mitm

ents

|(p

erc

en

t of G

DP

)

0 20 40 60 80Poverty head count (% population) of people living on less than $1 a day

88

When Are Donors Fickle?When Are Donors Fickle?Reason for lack of predictability

Fickle donor problem?

Budget aid Project aid

1) Major shift in policy or country circumstances, including emergencies

No No

2) Slow project implementation speed

N/a No

3) Specific conditionality not met

Possible Possible

4) Difficulties meeting donor-specific project disbursement procedures

N/a Possible

5) Administrative delays and slow response by donors

Yes Yes

6) Aid re-allocation or additions to aid envelopes for political or donor-related reasons

Yes Yes

99

Measuring PredictabilityMeasuring Predictability Two sources for this paper: OECD-DAC and Two sources for this paper: OECD-DAC and

database constructed from IMF programsdatabase constructed from IMF programs Neither data can give reasons for low Neither data can give reasons for low

predictability predictability need statistical inference need statistical inference OECD-DAC is more comprehensive (time, OECD-DAC is more comprehensive (time,

countries) but only countries) but only donor-reported donor-reported commitments commitments and disbursements, no split budget vs. project aidand disbursements, no split budget vs. project aid

IMF-based data IMF-based data measures government measures government disbursement expectations and allows tracing disbursement expectations and allows tracing adjustmentsadjustments to “aid surprises” in a macro- to “aid surprises” in a macro-consistent framework, has information on budget consistent framework, has information on budget and project aidand project aid

1010

Commitments and Commitments and Disbursements in OECD-DAC Disbursements in OECD-DAC

datadata Data are adjusted by excluding those donors Data are adjusted by excluding those donors

who never report any commitmentswho never report any commitments On an annual basis, and averaged over time, On an annual basis, and averaged over time,

African countries receive 1 percent of GDP African countries receive 1 percent of GDP moremore in disbursements than were committed; in disbursements than were committed; countries in other regions receive between countries in other regions receive between 0.2 and 1.1 percent of GDP 0.2 and 1.1 percent of GDP lessless than than committedcommitted

both disbursement shortfalls and excess both disbursement shortfalls and excess disbursements are important to gauge disbursements are important to gauge predictability issuespredictability issues

1111

Predictability in OECD-DAC Predictability in OECD-DAC datadata

Predictability is low: On average, the absolute Predictability is low: On average, the absolute deviation of annual disbursements from deviation of annual disbursements from commitments ranges from 1.7 percent (MNA) to commitments ranges from 1.7 percent (MNA) to 3.4 percent of GDP (Sub-Saharan Africa), albeit 3.4 percent of GDP (Sub-Saharan Africa), albeit with some improvement over time with some improvement over time

These numbers are particularly staggering for post-These numbers are particularly staggering for post-conflict cases such as Sierra Leone (9 percent), conflict cases such as Sierra Leone (9 percent), Mozambique (4.7 percent), but also very high in Mozambique (4.7 percent), but also very high in Zambia (6.5 percent), a country with difficult donor Zambia (6.5 percent), a country with difficult donor relations/uneven policy implementationrelations/uneven policy implementation

Results robust to changes in assumptions – for Results robust to changes in assumptions – for example if we assume committed resources are example if we assume committed resources are disbursed over 3 years rather than 1 yeardisbursed over 3 years rather than 1 year

1212

Low Predictability: Justified Low Predictability: Justified Caution or Fickle Donors? Caution or Fickle Donors?

Length under IMF programs mattersLength under IMF programs matters

-10

01

02

03

0

Com

mitm

ents

-Dis

bu

rsem

ents

(%

GD

P)

0 5 10 15Duration of IMF Program (years)

1313

Simple Regression of Absolute Deviations, Simple Regression of Absolute Deviations, Disbursement Shortfalls, Excess Disbursements (% Disbursement Shortfalls, Excess Disbursements (%

of GDP) of GDP) Years in IMF -0.201** -0.173 -0.242**Program [0.082] [0.118] [0.115]

IMF Program 0.66 0.18 1.151Dummy [0.621] [0.645] [0.961]

Governance (-1) -0.164 0.395 -0.601[0.397] [0.408] [0.480]

Net Aid (%GDP) (-1) 0.139*** 0.079* 0.168***[0.029] [0.042] [0.025]

Emergency Aid 0.547*** 0.782*** 0.475***(% GDP) [0.124] [0.256] [0.111]

Negative TOT 0.012 0.03 0.006 Shocks [0.021] [0.024] [0.032]

Positive TOT -0.022 -0.011 -0.015* Shocks [0.014] [0.027] [0.009]

Constant 1.558 -0.596 2.98[1.373] [1.845] [1.771]

R-Squared 0.23 0.22 0.29444 190 254

1414

Some lack of predictability is associated Some lack of predictability is associated with factors related to aid effectiveness with factors related to aid effectiveness considerationsconsiderations

Less predictable when less stable Less predictable when less stable relationship with donors as approximated relationship with donors as approximated by length of IMF program (but not when by length of IMF program (but not when entering or existing IMF program)entering or existing IMF program)

Less predictable when large emergency aid Less predictable when large emergency aid disbursementsdisbursements

Both factors seem reasonably close to Both factors seem reasonably close to cases where donors may not/cannot be cases where donors may not/cannot be predictable if they care about effectivenesspredictable if they care about effectiveness

But: leaves a large part of lack of But: leaves a large part of lack of predictability unexplained that could come predictability unexplained that could come from technical factors (project aid) and from technical factors (project aid) and conditionality (budget aid)conditionality (budget aid)

1515

RobustnessRobustness

Instrumentation of governance and net aid Instrumentation of governance and net aid with standard instruments (settler with standard instruments (settler mortality, colonial history, UN votes).mortality, colonial history, UN votes).

Use of country fixed effectsUse of country fixed effects

Findings robust to IV, but IMF effect Findings robust to IV, but IMF effect absorbed into country fixed effectsabsorbed into country fixed effects

Even with these additions, large Even with these additions, large unexplained part of aid predictabilityunexplained part of aid predictability

1616

Predictability in IMF-based dataPredictability in IMF-based data IMF reports allow to identify expected project IMF reports allow to identify expected project

and budget aid disbursements and outturnsand budget aid disbursements and outturns Projections used are typically the last available Projections used are typically the last available

before start of the budget yearbefore start of the budget year Need long-term IMF program engagement for Need long-term IMF program engagement for

datadata performing similar regression analysis performing similar regression analysis on this more limited data yields no longer on this more limited data yields no longer significant results regarding IMF, emergency significant results regarding IMF, emergency aid matters for project aid disbursementsaid matters for project aid disbursements

Projections and outturns are embedded in an Projections and outturns are embedded in an internally consistent set of macro variables internally consistent set of macro variables can trace where adjustments are being madecan trace where adjustments are being made

1717

Budget Aid and Taxes: How well are they projected?Budget Aid and Taxes: How well are they projected?Average

Budget AidAverage

Deviation

Mean Absolute Deviation

Average Tax Revenue

Average Deviation

Mean Absolute Deviation

Burkina Faso 1993-1999 2.95 -1.08 1.40 10.25 -0.02 0.912000-2005 2.88 0.06 0.44 11.00 -0.66 0.701993-2005 2.92 -0.55 0.96 10.60 -0.31 0.81

Ghana 1998-1999 1.85 -0.28 0.28 15.31 -0.71 0.712000-2005 3.44 0.35 0.84 18.93 0.78 1.261998-2005 3.04 0.19 0.70 18.02 0.41 1.12

Mali 1993-1999 3.52 0.12 1.06 12.30 -0.01 0.972000-2005 2.38 0.53 1.15 14.25 -0.67 0.721993-2005 2.99 0.31 1.10 13.20 -0.31 0.86

Rwanda 1997-1999 3.07 -2.21 2.21 9.68 -0.48 1.542000-2005 7.20 1.10 1.22 11.71 0.49 0.911997-2005 5.82 0.00 1.55 11.03 0.17 1.12

Sierra Leone 2001-2005 5.97 -1.46 2.66 11.19 0.44 0.70

Tanzania 1993-1999 3.08 -0.51 0.58 12.67 -1.06 1.382000-2005 3.92 -0.19 0.52 11.39 0.31 0.461993-2005 3.46 -0.36 0.55 12.08 -0.43 0.96

Whole Sample 1993-1999 3.16 -0.42 1.21 11.98 -0.13 0.892000-2005 3.42 -0.04 0.97 13.46 -0.18 0.891993-2005 3.31 -0.20 1.07 12.82 -0.16 0.89

1818

Where and How Do Countries Adjust to Budget Aid Where and How Do Countries Adjust to Budget Aid Shocks?Shocks?

-1.2

-0.7

-0.2

0.3

0.8

Budget Aid

Shortfall

Tax Revenue

Shortfall

Excess Dom.

Bank Financing

Other Excess Current

Expenditure

Dom. Fin.

Investment

Shortfall

Net Debt Service

and Arrears

Clearance

Revenue and financing adjustments Expenditure adjustments

-1.1

-0.6

-0.1

0.4

0.9

Excess Budget

Aid

Excess Tax

Revenue

Dom. Bank

Financing

Shortfall

Other Excess Current

Expenditure

Excess Dom.

Fin. Investment

Net Debt Service

and Arrears

Clearance

Revenue and financing adjustments Expenditure adjustments

s

1919

Common FeaturesCommon Features Need not only to absorb aid shock, but Need not only to absorb aid shock, but

often tax shock goes in the same often tax shock goes in the same direction, particularly when aid shortfall direction, particularly when aid shortfall (Tanzania and Ghana have practically (Tanzania and Ghana have practically identical tax revenue and aid shortfalls)identical tax revenue and aid shortfalls)

Current expenditure overruns are Current expenditure overruns are present whether in case of excess aid present whether in case of excess aid or aid shortfalls, but are twice as large or aid shortfalls, but are twice as large when in cases of excess aid when in cases of excess aid disbursementsdisbursements

2020

Key PatternsKey Patterns Budget aid shortfalls are absorbed Budget aid shortfalls are absorbed

through higher domestic financing or through higher domestic financing or new arrears and cuts in domestically new arrears and cuts in domestically financed investment spendingfinanced investment spending

Investment spending does not Investment spending does not accelerate in times of excess budget accelerate in times of excess budget aid, instead recurrent expenditure aid, instead recurrent expenditure increasesincreases

Excess budget aid disbursements Excess budget aid disbursements results in reduction of domestic debt results in reduction of domestic debt and higher current spendingand higher current spending

2121

-50

510

Dom

estic

Ban

k F

inan

cing

-6 -4 -2 0 2 4Budget Aid Deviations

Budget Aid and Domestic Bank Financing Deviations

2222

-2-1

01

2P

ublic

Inv

estm

ent

Dev

iatio

ns

-8 -6 -4 -2 0Budget Aid Shortfalls

Investment Deviations and Budget Aid Shortfalls

2323

-20

24

6C

urre

nt E

xpen

ditu

re D

evia

tions

0 1 2 3 4Budget Aid Excesses

Current Expenditure Deviations and Aid Excesses

2424

-4-2

02

4C

urre

nt E

xpen

ditu

re D

evia

tions

-8 -6 -4 -2 0Budget Aid Shortfalls

Current Expenditure Deviations and Aid Shortfalls

2525

Summary of Messages (1)Summary of Messages (1) Aid is not very predictable Aid is not very predictable but low but low

predictability results by both shortfalls and predictability results by both shortfalls and excess disbursementsexcess disbursements

Some of the lack of predictability can be Some of the lack of predictability can be explained with country factors/stability of explained with country factors/stability of policy implementation and/or the policy implementation and/or the occurrence of emergencies, which occurrence of emergencies, which generally could be seen as justifiable generally could be seen as justifiable reason to be unpredictablereason to be unpredictable

However, according to our regressions, However, according to our regressions, important elements cannot be explained important elements cannot be explained by fundamentalsby fundamentals

2626

Summary of Messages (2)Summary of Messages (2) Even where the country environment Even where the country environment

is stable, budget aid is unpredictable is stable, budget aid is unpredictable (with about 1 percent of GDP or 1/3 of (with about 1 percent of GDP or 1/3 of average budget aid at risk each year)average budget aid at risk each year)

Budget aid is less predictable than Budget aid is less predictable than tax revenuetax revenue

Low predictability of aid, particularly Low predictability of aid, particularly budget aid, is disruptive for budget budget aid, is disruptive for budget management and leads to permanent management and leads to permanent losses of domestic investment losses of domestic investment spendingspending

2727

Policy ImplicationsPolicy Implications Identify explicitly under which circumstances Identify explicitly under which circumstances

lack of predictability is acceptable for aid lack of predictability is acceptable for aid effectiveness reasonseffectiveness reasons

Focus on those types of aid for which Focus on those types of aid for which predictability is an essential ingredient predictability is an essential ingredient (budget aid, aid that finances recurrent costs)(budget aid, aid that finances recurrent costs)

Improve the measurement of predictability in Improve the measurement of predictability in particular adapt the OECD survey on the particular adapt the OECD survey on the Paris declaration to identify aid types and Paris declaration to identify aid types and draw on recipient projectionsdraw on recipient projections

Move to longer-term mechanisms of aid Move to longer-term mechanisms of aid allocation à la Eifert-Gelb (2006) linked to allocation à la Eifert-Gelb (2006) linked to country characteristics country characteristics longer-term longer-term commitments of aid beyond 1-3 year commitments of aid beyond 1-3 year windowswindows