Chapter

Appendix 2. The Macroeconomics of Aid

Author(s):
Yongzheng Yang, Robert Powell, and Sanjeev Gupta
Published Date:
March 2006
Share
  • ShareShare
Show Summary Details
Appendix Table A.2.1Official Net Transfers to SSA Countries(In percent of GDP)
1980–841985–891990–941995–992000–031980–2003
(More than 15 percent)
Equatorial Guinea47São Tomé and Príncipe38São Tomé and Príncipe78São Tomé and Príncipe52Eritrea34São Tomé and Príncipe46
Cape Verde39Guinea-Bissau35Mozambique36Guinea-Bissau24São Tomé and Príncipe33Guinea-Bissau30
Guinea-Bissau36Cape Verde30Guinea-Bissau29Mozambique23Mozambique29Mozambique27
Comoros29Equatorial Guinea29Eritrea26Rwanda23Sierra Leone28Eritrea25
São Tomé and Príncipe26Gambia, The26Cape Verde21Liberia20Congo, Dem. Rep. of25Cape Verde23
Mozambique22Mozambique22Equatorial Guinea20Malawi18Guinea-Bissau25Equatorial Guinea21
Gambia, The18Comoros20Rwanda20Eritrea18Burundi22Comoros16
Lesotho16Lesotho17Burundi20Cape Verde15Malawi17Gambia, The16
Ethiopia15Malawi19Malawi15
Gambia, The17
Uganda17
Tanzania16
Zambia15
(Between 5 and 15 percent)
Mali12Malawi14Chad13Sierra Leone12Ethiopia14Burundi15
Senegal11Chad14Niger13Burundi11Rwanda13Liberia14
Burundi11Mali13Lesotho13Niger10Uganda11Rwanda14
Seychelles10Burundi12Comoros13Chad10Niger11Ethiopia12
Central African Rep.9Niger12Sierra Leone12Central African Rep.10Zambia11Sierra Leone12
Tanzania9Zambia12Ethiopia12Burkina Faso10Tanzania11Lesotho11
Malawi9Ethiopia11Central African Rep.12Zambia10Liberia10Mali11
Burkina Faso9Tanzania10Mali11Comoros10Burkina Faso10Tanzania11
Chad8Central African Rep.10Benin10Madagascar9Gambia, The9Niger11
Niger7Burkina Faso9Madagascar10Mali9Mali9Zambia11
Botswana6Benin8Burkina Faso10Ethiopia8Cape Verde9Chad11
Benin6Senegal8Guinea9Uganda8Ghana8Burkina Faso9
Madagascar6Guinea8Ghana9Tanzania8Chad7Central African Rep.9
Rwanda6Madagascar7Senegal9Benin7Comoros7Uganda8
Zambia6Togo7Togo7Senegal7Benin6Madagascar8
Congo, Rep. of6Rwanda6Kenya6Gambia, The7Madagascar6Senegal8
Kenya5Seychelles6Côte d’Ivoire6Ghana7Benin8
Ghana6Zimbabwe5Togo6Congo, Dem. Rep.6
Sierra Leone6Guinea5Ghana6
Kenya5Lesotho5Togo6
Guinea6
(Less than 5 percent)
Swaziland4Congo, Rep. of5Cameroon4Equatorial Guinea5Central African Rep.5Seychelles5
Togo4Uganda5Congo, Dem. Rep. of4Congo, Rep. of3Senegal4Kenya4
Zimbabwe4Botswana5Angola4Angola2Lesotho4Congo, Rep. of4
Mauritius4Congo, Dem. Rep. of3Congo, Rep. of4Congo, Dem. Rep.2Togo3Zimbabwe3
Sierra Leone3Zimbabwe3Seychelles3Seychelles2Seychelles3Angola2
Côte d’Ivoire2Mauritius2Gabon1Zimbabwe2Guinea3Botswana2
Uganda2Angola2Botswana1Swaziland1Cameroon2Côte d’Ivoire2
Cameroon2Gabon2Swaziland0Kenya1Kenya2Cameroon2
Guinea2Cameroon2South Africa0Côte d’Ivoire1Zimbabwe1Swaziland1
Congo, Dem. Rep. of1Côte d’Ivoire1Mauritius0Cameroon0Equatorial Guinea1Mauritius1
Ghana1Swaziland0Nigeria–2South Africa0Angola1South Africa0
Gabon1Nigeria0Liberian.a.Botswana0Swaziland1Gabon–1
Nigeria0Liberian.a.Namibian.a.Mauritius–1Côte d’Ivoire1Nigeria–1
South African.a.South African.a.Nigeria–2Congo, Rep. of1Namibian.a.
Eritrean.a.Namibian.a.Gabon–3South Africa0
Namibian.a.Eritrean.a.Namibian.a.Mauritius0
Angolan.a.Botswana0
Liberian.a.Nigeria–2
Gabon–5
Namibian.a.
Source: World Bank (2005).Note: n.a. denotes not available.
Source: World Bank (2005).Note: n.a. denotes not available.
Table A.2.2.Aid Inflows and Real Exchange Rate Appreciation: Empirical Evidence
StudySampleAid and Real Exchange Rate Appreciation RelationshipEffect of 1 Percent Real Increase in Aid
Cross-Sectional Studies
Van Wijnbergen (1985)6 African countries, 1969–83MixedAppreciation of 0.2–0.9 percent over two years for some countries; no significant change for others
Adenauer and Vagassky (1998)4 CFA franc zone countries, 1980–92PositiveAppreciation of 0.13 percent over 2 years
Bulir and Lane (2002)9 developing countriesPositiven.a.
Prati, Sahay, and Tressel (2003)87 developing countries, 1960–98PositiveAppreciation of 0.04 percent in the first year
Elbadawi (1999)62 developing countries (28 from Africa), 1984–85, 1989–90, 1994–95PositiveAppreciation of 0.09 percent in the first year1
Yano and Nugent (1999)44 developing countries, 1970–90Mixedn.a.
Country-Specific Studies
Younger (1992)Ghana, 1960–88Positiven.a.
Kasekende and Atingi-Ego (1999)Uganda, 1970–96PositiveAppreciation of 0.03 percent in the first year
Nyoni (1998)Tanzania, 1967–93NegativeDepreciation of 0.13 percent in the first year
Sackey (2002)Ghana, 1962–96NegativeDepreciation of 0.03 percent in the first year
IMF (2005e)Ethiopia, Ghana, Mozambique, Tanzania, UgandaNegativeDepreciation of 1.5–6.5 percent the year following an aid surge2
Sources: As citedn.a. indicates that the study does not yield an elasticity measure.

Indicates response to 1 percent increase in ratio of ODA to GNP.

Indicates response to an aid surge with inflows ranging between 2 percent and 11 percent of GDP.

Sources: As citedn.a. indicates that the study does not yield an elasticity measure.

Indicates response to 1 percent increase in ratio of ODA to GNP.

Indicates response to an aid surge with inflows ranging between 2 percent and 11 percent of GDP.

Table A.2.3.Empirical Studies of Kinks in the Relationship between Inflation and Growth
Inflation threshold (percent)Growth effect of higher inflation below the thresholdCountriesPeriodInflation measureRemarks
Fischer (1993)15Negative801960–89CPI
Barro (1991)10–20Not significant1171960–9010-year average CPI
Sarel (1996)8Positive871970–905-year average CPI
Bruno and Easterly (1998)40Not significant971961–92CPI
Ghosh and Phillips (1998)>5Positive1451960–96Average annual CPI
Kochhar and Coorey (1999)5Positive84 (low- and middle-income countries only)1981–95Average annual CPI
Khan and Senhadji (2000)7–11 (for developing countries only)Positive1401960–985-year average CPI1–3 percent threshold for industrial countries Controlled for investment and unemployment
Burdekin, Denzou, Keil, Sitthiyot, and Willett (2000)3 (for developing countries only)Positive511967–928 percent threshold for industrial countries
Gylfason and Herbertsson (2001)10–20Positive1701960–925-year average GDP deflator
CPI = consumer price index.
CPI = consumer price index.
Table A.2.4.Incremental Effect of Aid on Domestic Revenue
StudySampleDomestic Revenue1
Heller (1975)9 Anglophone African countries, 1960–71–0.42
Pack and Pack (1990)Indonesia, 1966–860.29
Cashel-Cordo and Craig (1990)46 Least Developed Countries, 1975–80(African countries) 10.29
(non-African countries) –4.25
Gang and Khan (1990)India, 1961–840.00
Khilji and Zampelli (1991)Pakistan, 1960–86-0.01
Leuthold (1991)8 African countries, 1973–810.00
Khan and Hoshino (1992)5 Asian countries, 1955–761.20
Gupta (1993)India, 1969–930.01
Pack and Pack (1993)Dominican Republic, 1968–86–0.39
Rubino (1997)Indonesia–1.40
Iqbal (1997)Pakistan, 1976–950.00
Franco-Rodrigues, Morrissey, and McGillivray (1998)Pakistan, 1960–95(Direct effects) –2.90
(Total effects) –3.60
MacGillivray and Ahmed (1999)The Philippines, 1960–920.10
Franco-Rodriguez (2000)Costa Rica, 1971–941.10
McGillivray (2000)Pakistan, 1956–950.00
Swaroop, Jha, and Sunil Rajkumar (2000)India, 1970–950.00
Sources: McGillivray and Morrisey (2001); Feeny and McGillivray (2003); as cited.

Figures are the total effect of a one unit increment in aid on domestic revenue collections.

Sources: McGillivray and Morrisey (2001); Feeny and McGillivray (2003); as cited.

Figures are the total effect of a one unit increment in aid on domestic revenue collections.

Table A.2.5.Tax Revenue and Trade Taxes, by Region(In percent of GDP)
International Trade Taxes
Tax RevenueImport dutiesExport duties
Country Subgroups1Early 1990sEarly 2000sEarly 1990sEarly 2000sEarly 1990sEarly 2000s
Americas14.916.32.51.90.20.0
Sub-Saharan Africa16.315.94.93.51.00.4
Central Europe and BRO227.323.41.40.90.80.4
North Africa and Middle East15.117.13.63.00.10.1
Asia and Pacific13.613.23.21.90.30.2
Small islands25.524.513.59.70.30.0
Unweighted average
Developing countries317.917.63.92.70.50.2
High-income countries26.627.52.01.30.00.2
PRGF–eligible countries15.214.84.83.50.60.3
Source: Keen and Simone (2004).

Subgroups contain only developing countries.

Baltic countries, Russia, and the other countries of the former Soviet Union.

Defined as low- and middle-income countries.

Source: Keen and Simone (2004).

Subgroups contain only developing countries.

Baltic countries, Russia, and the other countries of the former Soviet Union.

Defined as low- and middle-income countries.

Table A.2.6.Relationship between Public and Private Investment in Sub-Saharan Africa
StudyData and CoverageResults
Greene and Villanueva (1991)23 developing countries (Asia, SSA, and Latin America); pooled samplePublic investment increases private investment
Hadjimichael and Ghura (1995)41 SSA countries; panel dataPublic investment increases private investment; also studied other policy determinants of private investment
Odedokun (1997)48 developing countries (SSA, Asia, Europe, and North Africa); panel dataPublic infrastructure investment increases private investment; non-infrastructure crowds out private investment
Ghura and Goodwin (2000)31 developing countries (Asia, SSA, and Latin America); panel dataPublic investment increases private investment in SSA, but crowds out in Asia and Latin America
Belloc and Vertova (2004)7 HIPC countries, 5 in SSA; country-level vector autoregressionsPublic investment increases private investment in 6 of 7 countries
Table A.2.7.The Effect of Public Investment on Output
StudyData and CoverageResults
Canning and Bennathan (2000)Penn World Tables and specialized infrastructure data; annual; all country income levelsOutput elasticity with respect to roads and electricity in range of 0.05–1.0; highest for MICs; strong input complementarity
Calderon and Serven (2004)World Bank data with infrastructure indices; 5-year averages; all country income levelsLarge output gains to closing infrastructure gaps
Barro (1991)Penn World Tables, World Bank and UN; averages, 1960–85; all country income levelsNo impact of public investment on growth
Easterly and Rebelo (1993)Specialized data capturing broad public investment; 10-year averages; all country income levelsNo impact of public investment aggregate on per capita growth, but big impact (0.6) for transport and communication and direct public investment of the government (0.4)
Devarajan, Swaroop, and Zou (1996)Government Financial Statistics; 5-year averages; LICs and MICsStudies impact of changes in budget composition; capital spending has negative effect
Khan and Kumar (1997)Penn World Tables, World Bank, World Economic Outlook; 10-year averages; LICs and MICsPublic investment has positive impact on per capita growth (0.3), but weakening in recent samples
Clements, Bhattacharya, and Nguyen (2003)World Bank; 3-year averages; LICsPublic investment has positive impact (0.2) when not deficitfinanced
Gupta, Clements, Baldacci, and Mulas-Granados (2004)IMF; annual; LICsPositive impact of public capital outlays (0.7) on per capita growth
Source: Adapted from Table 1 in IMF (2004a) to those studies that include evidence from LICs.Note: Growth regression studies use country average growth rates typically calculated over 5- or 10-year periods. The numbers in parentheses in the third column indicate the estimated response of this average to a permanent increase in the expenditure share of 1 percent of GDP. They should thus be interpreted as the long-run response to a small permanent change in the expenditure share. While Gupta Clements, Baldacci, and Mulas-Granados (2004) use annual growth data, their specification allows them to distinguish short- and long-run impacts.LIC = low-income country; MIC = middle-income country.
Source: Adapted from Table 1 in IMF (2004a) to those studies that include evidence from LICs.Note: Growth regression studies use country average growth rates typically calculated over 5- or 10-year periods. The numbers in parentheses in the third column indicate the estimated response of this average to a permanent increase in the expenditure share of 1 percent of GDP. They should thus be interpreted as the long-run response to a small permanent change in the expenditure share. While Gupta Clements, Baldacci, and Mulas-Granados (2004) use annual growth data, their specification allows them to distinguish short- and long-run impacts.LIC = low-income country; MIC = middle-income country.
Table A.2.8.Impact of Health and Education Sectors on Output
StudyData and CoverageResults
Barro (1991)Penn World Tables, World Bank and UN; averages 1960–85; all country income levelsSignificant role for human capital (especially initial stock of secondary education)
Barro and Sala-i-Martin (1995)Data similar to Barro’s (1991); 10-year averages; all country income levelsPositive impact of education spending on per capita growth (0.2)
Devarajan, Swaroop, and Zou (1996)Government Financial Statistics; 5-year averages; LICs and MICsImpact of changes in budget composition; current spending has positive impact, mixed results for functional breakdown
Krueger and Lindahl (2001)Similar to Barro (1991) augmented with World Values Survey; data averages of various lengths; all country income levelsBig impact of change in schooling on growth, but only detectable in long-period averages (10-20 years); size of effect varies with econometric specification
Baldacci and others (2004)World Bank and IMF; 5-year averages; LICs and MICsPositive impact of spending on education (0.5) and health (0.4) after 5 years; education impact rises to 1.4 after 15 years; diminishing returns to level of education and health spending
Canning, and Sevilla Bloom (2003)Penn World Tables, UN, and World Bank; 10-year averages; all country income levelsStudies impact of health indicators and not expenditure; 1-year increase in average life expectancy raises output by 4 percent
Note: These studies use country average growth rates typically calculated over 5- or 10-year periods. The numbers in parentheses in the third column indicate the estimated response of this average to a permanent increase in the expenditure share of 1 percent of GDP. They should thus be interpreted as the long-run response to a small permanent change in the expenditure share.LIC = low-income country; MIC = middle-income country.
Note: These studies use country average growth rates typically calculated over 5- or 10-year periods. The numbers in parentheses in the third column indicate the estimated response of this average to a permanent increase in the expenditure share of 1 percent of GDP. They should thus be interpreted as the long-run response to a small permanent change in the expenditure share.LIC = low-income country; MIC = middle-income country.
Table A.2.9.Impact of Increasing Corruption by One Unit1
StudyImpact onFinding
Mauro (1995)Real per capita GDP growth–0.3 to –1.8 percentage points
Leite and Weidmann (2002)Real per capita GDP growth–0.7 to –1.2 percentage points
Tanzi and Davoodi (2002)Real per capita GDP growth–0.6 percentage points
Abed and Davoodi (2002)Real per capita GDP growth–1 to –1.3 percentage points
Mauro (1995)Ratio of investment to GDP–1 to –2.8 percentage points
Mauro (1998)Ratio of public education spending to GDP–0.7 to –0.9 percentage points
Mauro (1998)Ratio of public health spending to GDP–0.6 to –0.9 percentage points
Gupta, Davoodi, and Alonso-Terme (2002)Income inequality (Gini coefficient)+3.5 to +4.25 Gini points
Gupta, Davoodi, and Alonso-Terme (2002)Income growth of the poor–2 to –10 percentage points
Ghura (2002)Ratio of tax revenues to GDP–1 to –2.9 percentage points
Tanzi and Davoodi (2002)Measures of government revenues to GDP ratio–0.1 to –2.7 percentage points
Gupta, de Mello and Sharan (2002)Ratio of military spending to GDP+ 1 percentage point
Gupta, Davoodi, and Tiongson (2002)Child mortality rate+ 1.1 to 1.5 deaths per 1,000 live births
Gupta, Davoodi, and Tiongson (2002)Primary student dropout rate+ 1.1 to 1.4 percentage points
Tanzi and Davoodi (1998)Ratio of public investment to GDP+ 0.5 percentage points
Tanzi and Davoodi (1998)Percent of paved roads in good condition–2.2 to –3.9 percentage points
Sources: IMF, Fiscal Affairs Department; Transparency International (2001).

Corruption is measured on a scale of 0 (highly clean) to 10 (highly corrupt).

Most of the above studies use the Transparency International corruption measure, rescaled so that higher values in the range 1-10 correspond to higher corruption. When the International Country Risk Guide measure of corruption was used, it was rescaled in the same way. Both the Transparency International and International Country Risk Guide measures rely on expert perceptions. For Gupta, Davoodi, and Tiongson (2000), an index was constructed from National Service Delivery Surveys. This has the advantage of being based on reported service client experience.
Sources: IMF, Fiscal Affairs Department; Transparency International (2001).

Corruption is measured on a scale of 0 (highly clean) to 10 (highly corrupt).

Most of the above studies use the Transparency International corruption measure, rescaled so that higher values in the range 1-10 correspond to higher corruption. When the International Country Risk Guide measure of corruption was used, it was rescaled in the same way. Both the Transparency International and International Country Risk Guide measures rely on expert perceptions. For Gupta, Davoodi, and Tiongson (2000), an index was constructed from National Service Delivery Surveys. This has the advantage of being based on reported service client experience.

    Other Resources Citing This Publication