Aid Volatility in Africa

8/12/2019 Aid Volatility in Africa

1/10

The Impact of Aid and Public Investment Volatility on Economic

Growth in Sub-Saharan Africa

MALIMU MUSERU, FRANCOIS TOERIEN and SEAN GOSSEL *

University of Cape Town, Rondebosch, South Africa

Summary. This study investigates the effects of aid inflows and the volatility of public investment on economic growth in 26 Sub-Sah-aran African countries over the period from 1992 to 2011. Three volatility variables comprising aid, government revenue, and publicinvestment are incorporated into an aid-growth model to test for their effect on economic growth. Using the Generalized Method ofMoments (GMM) technique and averaged data for five four-year sub-periods, we show that although foreign aid has a positive impacton growth once potential endogeneity has been accounted for, aid effectiveness may have been eroded by volatility in public investment.2013 Elsevier Ltd. All rights reserved.

Key words aid, economic growth, public investment, Sub-Saharan Africa, volatility

1. INTRODUCTION

One of the most debated issues in development economics iswhether Official Development Assistance (foreign aid) pro-motes GDP per capita growth in aid recipient countries.Two factors currently make this debate especially relevant toboth aid donors and its recipients in Sub-Saharan Africa.Firstly, indications are that the principal objective of theUnited Nations Millennium Developments Goals, which isto reduce poverty to half the 1990 level by 2015, will not bemet in Sub-Saharan Africa despite continued aid inflows(Addison, Mavrotas, & McGillivray, 2005). Secondly, thereare increasing signs of donor fatigue, especially in the context

of the current global economic crisis that has had a particu-larly negative effect on many of the traditional aid donors toSub-Saharan African countries (OECD. Development: Aidto developing countries falls because of global recession.,2012).

In recent decades there has been much research focused onthe effects of aid inflows on growth rates, as well as on deter-mining which socio-political, climatic, institutional, and eco-nomic factors undermine or enhance the effectiveness offoreign aid with respect to growth. However, overall, there isno consensus regarding aid effectiveness and disagreement re-mains as to what constitutes sufficient conditions for aid tohave a positive impact on the GDP per capita growth of aidrecipient countries. Nevertheless, there is emerging consensusthat the impact of aid on economic growth works through

public investment.Recent literature shows that volatility in taxes in

Sub-Saharan Africa reduces public investment levels 1 (Ebeke& Ehrhart, 2011). Since aid is, like taxes, a source of publicinvestment, heightened aid volatility may therefore in similarfashion negatively impact economic growth.

Although the impact of aid on growth has been extensivelyresearched, there are to our knowledge very few papers thatexamine the effect of aid volatility on growth (for examplesof such papers, see Bulr & Hamann, 2008; Fielding,Mavrotas, & discussion paper, 2005; Lensink & Morrissey,2000; Hudson & Mosley, 2008). In contrast to these papers,our study is not only concerned with the effect of aid volatilityon growth, but in addition also introduces public investment

volatility and government revenue volatility into the aid-

growth framework. As public investment is a function ofaid, central government revenue, and internal and externalborrowing, there may be other stabilizing determinants ofpublic investment even in the presence of aid volatility. Tothe best of our knowledge, this is the first study that incorpo-rates public investment volatility in the aid-growth relation-ship. In addition, unlike most previous aid-growth studieson Sub-Saharan Africa, our study uses a sample period(19922011) that includes the recent period of global economicuncertainty.

The remainder of this paper is structured as follows.Section2provides a brief overview of the previous aid-growthliterature. Section 3 discusses the methodology employed in

this study. Section 4 examines the data. Section 5 discussesthe descriptive statistics. Section 6 presents a discussion ofthe results, and Section7 concludes.

2. LITERATURE REVIEW

Academic research on the impact of aid on economic growthcan be divided into three distinct phases. The first generationliterature studied the aid-growth question within the contextof so-called gap models. These focused on the effects of financ-ing constraints on economic growth (termedgaps) in low-in-come countries, and how aid interacted with, and couldalleviate, these gaps. In the original gap model, 2 Domar(1947)saw aid as a way of alleviating the growth restriction

resulting from low savings. Thus, if a government wishes to in-crease the growth rate, it has to increase savings. In the casewhere national savings (private and public savings) are toolow (constrained), aid permits increased investment andgrowth (Morrissey, 2001). Early literature made no effort tomeasure the aid-growth nexus directly. These studies insteadtested the aid-savings link empirically by making savings thedependent variable and aid the independent variable in theirregressions.

The original saving-gap model did not survive closer scrutinyand in the later two-gap model Chenery and Strout (1966)introduced import capacity (the foreign exchange gap) as aseparate potential constraint on growth. The authors

* Final revision accepted: December 17, 2013.

World DevelopmentVol. 57, pp. 138147, 2014 2013 Elsevier Ltd. All rights reserved.

0305-750X/$ - see front matter

www.elsevier.com/locate/worlddevhttp://dx.doi.org/10.1016/j.worlddev.2013.12.001

138
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://dx.doi.org/10.1016/j.worlddev.2013.12.001http://dx.doi.org/10.1016/j.worlddev.2013.12.001http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


2/10

contended that a country will transform its economy success-fully via a simultaneous increase in skills, domestic saving,and export earnings. Despite this extension, the spirit of thetwo-gap model remained similar to the Domar-model, i.e.,investment spurs growth. Since national savings and importsare constrained in low-income countries, aid has the potentialof filling these gaps and hence fueling growth. However, the im-

pact of additional foreign finance will vary depending on whichgap is binding. The foreign finance gap, rather than the savingsgap, is deemed to have a greater impact on investment whenforeign exchange is the constraint.

Finally, using a theoretical argument, Bacha (1990) intro-duced a third gapthe fiscal gap. This stemmed from the ideathat growth prospects of highly indebted developing countriesare severely hampered by the resulting fiscal constraints. Asdebt lingers on, the primary source of growth problems isderived from budget limitations rather than foreign exchangeor savings constraints. As the majority of aid is extended di-rectly to governments, the fiscal constraint should be relaxed,provided that aid is used for productive investments.

The second generation literature attempted to directly inves-tigate the aid-growth relationship, instead of addressing this

aspect only indirectly through the aid-savings link. However,results are often contradictory, and depend heavily on thesample and period covered. 3 Levy (1988)is one of the few sec-ond generation studies specifically devoted to low incomecountries within Sub-Saharan Africa (with populations of overa million people). Using two samples, one with annual data (toavoid eliminating substantial random annual effects by averag-ing), and the other with averaged data (to mitigate the effectsof lagged responses), the results of his study over a 15-yearperiod shows that there is a positive and significant effect ofaid on growth, and that fixed capital formation is a contribu-tor to the rate of growth.

The third generation literature developed in the late 1990s asa means to bring two principle innovations to the existing liter-

ature. Firstly, data sets now covered an increasing number ofdeveloping countries as well as longer sample periods; and sec-ondly, in line with the new growth theory, studies includedexplanatory variables to control for the economic and institu-tional environment. Thus, these variables are incorporated di-rectly into reduced form growth regressions alongside the moretraditional macroeconomic variables. In addition, the aid-growth relationship is explicitly seen as nonlinear, as it is ar-gued that aid has diminishing returns (Hansen & Tarp, 2000).

A key debate in the third generation literature is whether theimpact of aid on growth is (on average) unconditionally posi-tive, or only conditionally positive, and if the latter, the ques-tion becomes: what are the necessary conditions that make aideffective in different subsets of countries in varying periods?Burnside and Dollar (2000) test the hypothesis that the effec-

tiveness of aid is dependent on a countrys economic policiesand institutional framework. These authors use a sample offormer Eastern Bloc countries over the period of 19702003,and find that aid only effectively promotes economic growthin a good policy environment. 4 Consequently, additionalstudies examined whether the aid-growth relationship is a con-ditional or an unconditional one. However, the results ofempirical studies are mixed.Collier and Dollar (2002), Collierand Hoeffler (2004)and Dalgaard, Hansen, and Tarp (2004)report a conditional relationship, while Hadjimichael,Dhaneshwar, Muhleisen, Nord, and Ucer (1995), Durbarry,Gemmell, and Greenaway (1998)andRajan and Subramanian(2008) find an unconditional relationship. According toClemens, Radelet, and Bhavnani (2012), the main objectionof the proponents of the unconditional view is that they argue

that Boones (1996) assumption of a linear impact of aid ongrowth is less realistic than a nonlinear relationship, becauseaid is subject to diminishing returns partly due to aid receivingcountries having a fixed absorptive capacity.

In tandem with the development of the third generation liter-ature, another strand of research has focused on the effect of aidvolatility on economic growth.Hudson and Mosley (2008)use

a sample of 131 developing countries over the period 19772001to test the effects of positive and negative aid volatility, thesebeing respectively defined as sudden surges and declines inaid. The argument made is that positive volatility has the poten-tial of reducing aid effectiveness due to absorptive capacity con-straints, while negative aid volatility can disrupt governmentbudgetary planning, and may result in projects being aban-doned, thus reducing aid effectiveness. These researchers ulti-mately find that surges in aid significantly reduce the sharesof government expenditure, investment, and consumers expen-diture. However, Lensink and Morrissey (2000) hypothesizethat it is aid uncertainty rather than the overall instability ofaid that affects growth. Since aid commitments are generallypredetermined, and due to the expected continuity of donor-re-cipient relations, knowing past values of aid inflows enables

recipients to anticipate some variability in aid. In order toexamine this hypothesis, Lensink and Morrissey use the stan-dard deviation of the residuals of aid forecasting equations asa proxy for aid volatility in a sample of developing countriesover a 25 year period. The results show that there is a stronglysignificant, positive, linear relationship between aid andgrowth, but only after controlling for aid uncertainty.

In summary, it is widely accepted that aid influences growththrough public investment.Ebeke and Ehrhart (2011)deviateslightly from aid literature by examining the hypothesis thattax revenue instability has had a negative impact on the levelof public investment among 37 Sub-Saharan Africa countriesover the period 19802005. They find that the instability of taxrevenues in these Sub-Saharan African countries led to public

investment volatility which reduced public investment levels.Hence, the authors conclude that the instability of tax reve-nues was detrimental to long term economic growth. However,public investment is in part financed by government revenueand aid itself. Therefore, like revenue volatility, it is envisagedthat aid volatility may also lead to public investment volatilitywhich reduces levels of public investment, and may hence bedetrimental to growth. However, not all sources of financingpublic investment are as volatile. Furthermore, because publicinvestment is financed by aid, central government revenue,and internal and external borrowing, there may be other stabi-lizing sources of public investment even in the presence of aidand revenue volatilities, and therefore it is the overall volatilityof public investment that matters and not that of its individualfinancing sources in isolation.

Incorporating public investment volatility into the standardaid-growth framework within the Sub-Saharan African con-text will provide an indication of the degree to which publicinvestment volatility may have eroded per capita growth overthe covered period, where SSA has received significant aid in-flows. Our research tends to provide an alternative way ofthinking about the aid-growth conundrum in SSA. Wheremost recent aid-growth research has focused on trying to ex-plain conditions that make aid work (e.g., good policy orgeography), the present research shows that the problem withaids impact on growth is at its main transmission mechanism,and the volatility thereof. To the best of our knowledge, this isthe first study to incorporate public investment volatility intothe aid-growth framework. In doing so, we address the ques-tion of whether public investment volatility, which has been

THE IMPACT OF AID AND PUBLIC INVESTMENT VOLATILITY ON ECONOMIC GROWTH IN SUB-SAHARAN AFRICA 139
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


3/10


4/10

control for nonlinear effects of aid on growth (the diminishingreturns to aid), attributed to limited absorptive capacity andDutch disease (Durbarry et al., 1998; Rajan & Subramanian,2011), or alternatively to aid dependence whereby governmentsrelax their revenue collection efforts due to substantial and sus-tained aid inflows (Hepp, 2008). 14 The third control variable isprivate credit extended by deposit money banks as a percentage

of GDP, which is included as a proxy for financial development(Levine & Zervos, 1998). 15

In studying the relationship between institutions andgrowth, researchers have used a number of different measuresof institutional quality. Our fourth control variable is a proxyfor institutional quality, and is derived based on Knack andKeefer (1995), who used a composite of three equally weightedmeasures: bureaucratic quality, law and order, and corruption.We replicate this measure using similar variables from thePolitical Risk Services International Country Risk Guide(PRS ICRG), 16 which has the advantage of having data avail-able for many developing countries dating back to the 1980s.As a result, we are able to determine this measure on an an-nual basis over the entire period.

The fifth control variable is a measure of restriction on

freedom,used as a proxy for the political system and the de-gree of a democratic process, measured as the combined aver-age of political rights and civil liberties on a one-to-sevenscale, with one representing the highest level of freedom andseven the lowest.

Recent cross country analyses of economic growth have incor-porated physical geography in an attemptto explainwhether geo-graphical endowments influence development. 17 Of the worldsmajor regions, Sub-Saharan Africa has the highest proportionof land and population in the tropics, andtropical regions in gen-eral lag far behind temperate regions in economic development(Bloom, Sachs, Collier, & Udry, 1998, p. 211). Hence, in accor-dance withDalgaard et al. (2004), the sixth control variable island area in tropics, measured as the percentage of the land area

in a tropical climate zone.

18

With regard to data sources, data on real per capita GDP, thegrowth rate of real per capita GDP, central government reve-nue, and public investment (as a percentage of GDP) and pop-ulation are taken from the World Bank African DevelopmentIndicators (ADI) (2013). Data on private credit (as percentageof GDP) are taken from Beck and Demirguc-Kunt (2010).The aid variable used in this study is net official developmentassistance (net ODA) provided by the OECD. The equallyweighted components of the institutional quality measure(bureaucratic quality, law and order, and corruption), were alltaken from the Political Risk Services International CountryRisk Guide (PRS ICRG) (2013) and recomputed by theauthors. The restriction on freedom variable is the average ofthemeasures for politicalrights and civil libertiesfrom the Free-

dom House World Country Ratings 2012. Finally, the variabledescribing the percentage of land area in the tropics is sourcedfrom the geography datasets of the Center for InternationalDevelopment (CID) at Harvard University.

Econometric studies involving African countries often sufferfrom missing data limitations. Hence, this study employs themultiple imputation technique ofHonaker and King (2010)to adjust the data sample for missing data. 19 This techniqueestimates or imputes missing data using a predictive modelthat incorporates both the available information and any priorknowledge of the data and relationships between the variables.Multiple possible complete data sets are generated and the ex-pected value for any missing observation is the mean value ofthese multiple imputations. Each of these data sets can beanalyzed individually, or the results can be combined using

Table1.D

escriptivestatistics:1992

2011

Variable

Sub-Sample1(199295)

Sub-Samp

le2(199699)

Sub-Sample3(200003

)

Sub-Sample4(200407)

Sub-Sample5(200811)

Mean

r

Min.

Max.

Mean

r

Min.

Max.

Mean

r

Min.

Max.

Mean

r

Min.

Max.

Mean

r

Min.

Max.

GDPpercapitagrowth(%)

0.8

6

2.0

0

6.3

1

4.0

2

1.5

4

2.40

4.1

4

6.2

5

1.0

6

2.6

6

7.2

84.47

2.4

6

2.6

1

4.4

1

9.3

4

2.3

0

2.0

3

1.6

2

6.8

7

GDPpercapita(US$)

694

1040

109

4516

730

1092

121

4626

747

1090

127

4054

797

1159

151

4119

8491

207

176

4237

Aid(%ofGDP)

16.1

5

14.5

0

0.2

2

58.7

1

10.3

7

9.55

0.3

4

45.7

3

10.3

8

9.4

0

0.3

9

32.3

2

10.0

2

6.8

6

0.2

8

23.3

6

8.5

7

5.5

4

0.3

5

19.5

5

Centralgovernmentrevenue(%ofGDP)

20.4

3

14.4

2

8.0

3

73.9

0

19.8

2

14.23

6.4

9

73.4

5

21.0

9

13.9

6

9.8

4

69.5

7

21.1

2

11.9

0

8.9

1

58.2

5

20.5

6

9.1

4

9.1

4

46.1

5

Publicinvestment(%ofGDP)

7.4

1

4.8

6

0.7

5

23.5

5

6.6

4

3.49

0.7

2

13.1

4

6.3

9

3.3

6

1.7

4

13.0

5

7.0

8

4.0

5

2.1

7

18.8

2

7.5

9

3.3

2

2.5

8

16.3

0

Aidvolatility

0.8

5

1.3

9

3.8

1

2.8

0

0.6

8

1.30

3.0

4

2.6

6

0.2

8

1.2

7

3.3

81.93

0.2

9

1.2

0

3.2

9

2.1

6

0.3

0

0.9

6

2.7

3

1.3

3

Revenuevolatility

0.6

5

0.9

5

0.8

4

3.7

6

0.4

3

0.66

0.6

6

1.8

9

0.4

3

0.9

2

0.8

32.47

0.3

2

0.8

4

1.4

2

2.1

0

0.4

8

0.9

1

1.5

6

2.1

6

Publicspendingvolatility

0.0

7

0.9

8

1.7

6

1.9

8

0.0

2

0.95

2.1

3

1.4

7

0.1

9

0.7

3

1.9

01.07

0.2

1

0.8

6

1.9

5

1.0

3

0.1

8

0.5

4

0.7

0

1.1

6

Privatecredit(%ofGDP)

11.7

9

9.9

3

0.1

3

47.4

4

11.2

3

11.67

0.4

0

61.4

3

12.6

6

12.1

5

2.2

3

64.4

0

15.9

0

15.6

1

1.4

9

67.2

1

17.5

51

4.5

0

3.2

5

77.8

7

Institutionalquality

2.5

1

0.7

0

1.0

1

3.9

4

2.4

4

0.53

1.3

3

3.4

0

2.1

3

0.5

2

1.1

9

3.22

2.0

7

0.5

0

1.3

8

3.0

6

2.1

2

0.4

6

1.5

0

3.0

0

Freedom

4.6

5

1.1

7

2.1

3

7.0

0

4.4

2

1.35

1.5

0

7.0

0

4.2

5

1.3

3

1.5

0

7.00

4.1

0

1.4

7

1.6

3

7.0

0

4.2

1

1.4

2

1.5

0

7.0

0

N

otethatthevolatilitymetricsaremeasured

asthelogarithmof4-yearrollingstandard

deviation.

http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


5/10

the approach ofRubin (1987). The use of multiple rather thansingle imputations for missing observations is deemed moreappropriate in providing more robust standard error acrossthe estimated values (Honaker & King, 2010).

5. DESCRIPTIVE STATISTICS

A casual comparison of the descriptive statistics for the fivefour-year period averages employed in this study reveals signsof economic development in the sample over time. Averagegross domestic product (GDP) per capita increased from$694 over 199295 to an average of $849 over 200811, sup-ported by improving average per capita GDP growth from0.86% over 199295 to 2.30% over 200811. In contrast,average aid to GDP declined for most of the sample periods(being 16.15%, 10.37%, 10.38%, 10.02%, and 8.57% respec-tively for the five sub-periods). According to Addison et al.(2005), the decline in aid volumes from aid donors (both bilat-eral and multilateral) to Sub-Saharan Africa in the first threeperiods was a result of a re-allocation away from Africa toEastern Europe and East Asia. The decline in aid over the last

period is probably a combined effect of declining aid flowsresulting from the global financial crisis during most of thisperiod, and robust economic growth (i.e., a greater averageGDP numbers), as indicated above. Not surprisingly, Africaslargest economy (South Africa) received the smallest amountof aid as a percentage of GDP among the sample countries.South Africa and Botswana (both having very low to aid to

GDP ratios) provide a good argument for those who believethat growth can be achieved in Africa by weaning Africancountries off foreign aid.

Over the sample period public investment initially trendeddownward (from 7.41% of GDP during 19925, to 6.39% in200003), before turning upward again to end at 7.59% for200811. A fundamental enabler of public investment, and

hence growth in any country, is its ability to tax its citizens.The challenge to the Sub-Saharan Africa countries in the sam-ple (most of which are major recipients of aid) is that signifi-cant volumes of aid create disincentives for the efficientcollection and administration of tax revenue (Moss, Petters-son, & Van de Walle, 2006). It is of interest to note that onaverage central government revenue as a ratio of GDP wasstable at around 18.3% over the first three periods, but some-what higher at approximately 19.4% over the latter two peri-ods. There is thus in broad terms an inverse relationshipbetween the aid trend discussed above and central governmentrevenue to GDP, which lends some support to the hypothesisthat aid may reduce the efficiency of tax collection.

A recent study by Ebeke and Ehrhart (2011) shows thattax revenue volatility causes public investment instability,

which reduces public investment levels and ultimatelygrowth. Looking at the average numbers for the 26 sampledSub-Saharan Africa countries, it is noted that revenue vola-tility nearly halved over the first four periods, beforeincreasing again over the most recent period, possibly as aresult of the global economic instability over this period(Table 1).

Table 2. The impact of aid on growth in Sub-Saharan Africa (19922011)

19922011

Dependent variable: GDP per capita growth Formulation

1 2 3 4 5 6 7 8 9

Aid 0.111*** 0.137*** 0.118** 0.148*** 0.098* 0.086 0.086 0.100 0.0900.045 0.047 0.050 0.060 0.058 0.125 0.308 0.150 0.334

Aid2 0.006 0.003 0.001 0.005* 0.006 0.002 0.002 0.002 0.0030.009 0.012 0.003 0.002 0.003 0.004 0.010 0.003 0.011

Public investment 0.098 0.128 0.106 0.1830.082 0.117 0.182 0.411

Per capita GDP t1 6.944*** 8.947*** 7.124*** 7.554*** 6.836** 10.108*** 8.139*** 8.609*** 11.224***

2.002 2.834 2.167 2.125 3.172 3.176 2.254 2.426 3.670

Private credit 0.535*** 0.510** 0.500** 0.632* 0.421* 0.403*** 0.667*** 0.425 0.372*

0.158 0.246 0.234 0.340 0.260 0.159 0.224 0.416 0.224

Institutional quality 4.332*** 4.665*** 4.387*** 3.357*** 4.018*** 4.182*** 3.157*** 3.122*** 3.941***

1.040 0.418 1.078 0.996 0.740 0.522 0.982 1.047 1.053

Freedom 0.913* 0.912* 0.910* 0.660 0.199 1.254* 0.383 0.592 0.7900.530 0.470 0.483 0.431 0.618 0.732 0.466 0.431 0.442

Aid volatility 0.786*** 0.665** 1.196** 0.816**

0.272 0.294 0.586 0.380Revenue volatility 1.255*** 0.577* 1.105** 0.680**

0.464 0.343 0.468 0.325

Public investment volatility 1.235** 1.291**

0.631 0.597

# of countries 26 26 26 26 26 26 26 26 26# of observations 130 130 130 130 130 130 130 130 130PvaluesSargan statistic 0.39 0.32 0.32 0.38 0.45 0.53 0.53 0.60 0.31AR (1) (test for serial correlation) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00AR (2) (test for serial correlation) 0.70 0.60 0.68 0.59 0.64 0.56 0.63 0.77 0.51

Values in italics are robust standard errors with Whites correction for autocorrelation.* Significance at the 10% levels.** Significance at the 5% levels.***

Significance at the 1% level.

142 WORLD DEVELOPMENT


6/10

6. RESULTS AND DISCUSSION

The GMM results for the full sample period of 19922011are presented in Table 2, while the results for the period of19922007 (in order to exclude possible anomalies as a resultof the global financial crisis) are presented in Table 3.

The analysis commences with a discussion of the control

variables. The first control variable is conditional convergencederived from the Solow (1956) model, which predicts thateconomies tend to converge toward a steady-state path. Con-ditional convergence effects are captured by including the ini-tial (lagged) level of real GDP per capita. Our data for boththe full and reduced sample periods yield the expected negativesign (6.944 and10.790, respectively) at a significance levelof one percent, thus supporting the conditional convergencehypothesis.

In studying the relationship between institutions andgrowth, researchers have used a number of different measuresof institutional quality. The present paper followsKnack andKeefer (1995)by employing the sum of three equally weightedmeasures bureaucratic quality, law and order, and corrup-tion. Most previous studies, such as that of Burnside and

Dollar (2000), assumed these variables to be relatively stableover time for any given country, and therefore used fixedmeasures of institutional quality over the entire sample period.We depart from this approach by updating the institutionalquality variable on an annual basis for each of the countriesin our sample. The coefficients for institutional quality for

the full and reduced sample periods are 4.332 and 3.602,respectively, and are higher than those found by both Hepp(2008)and Burnside and Dollar (2000). 20 This could suggestthat the institutional quality within Sub-Saharan Africa isimproving and therefore having an increasingly positive im-pact on GDP growth, especially as in both cases this variablein significant at the one percent level.

As hypothesized, we find the coefficients of private credit tobe positive, with generally a high level of statistical significanceacross all formulations for both sample periods. Restriction onfreedom mostly have negative coefficients, thus supporting theliterature that states that less democratic countries grow at aslower pace compared to more democratic economies. How-ever, the finding that this control variable is only significantin five of the 19922007 formulations, suggests that either itmay not play a significant role in GDP growth in Sub-SaharanAfrica over the sample period or the more plausible alternativethat the effect of this variable is already captured by the insti-tutional quality variable.

The discussion now shifts to the variables of interest, startingwith foreign aid. Several influential studies 21 have argued thatthe impact of aid on growth is conditional on other variables

such as policyor geography, and that aid by itself has no im-pact on growth. 22 However, the results for both of our samplescontradict such findings and reveal that aid has a positive im-pact on growth. During 19922011, aid contributed between0.09% and 0.14% per year to real GDP per capita growth forevery additional one percent of aid (expressed as a percentage

Table 3. The impact of aid on growth in Sub-Saharan Africa (19922007)

19922007

Dependent variable: GDPper capita growth

Formulation

1 2 3 4 5 6 7 8 9

Aid 0.223*** 0.173*** 0.198*** 0.170*** 0.167*** 0.144 0.174 0.191 0.164

0.024 0.026 0.019 0.051 0.021 0.150 0.193 0.173 0.119Aid2 0.004** 0.005 0.003 0.002 0.003 0.006 0.002 0.003 0.002

0.002 0.003 0.002 0.004 0.012 0.007 0.005 0.004 0.005

Public investment 0.182*** 0.194*** 0.164*** 0.368**

0.058 0.041 0.043 0.142

Per capita GDP t1 10.790*** 9.990*** 11.959*** 10.487*** 17.048*** 11.639*** 12.585*** 10.093*** 13.752***

1.097 1.455 0.733 2.868 1.624 2.748 2.477 2.286 1.610

Private credit 0.763*** 0.747*** 0.910*** 0.609 0.835*** 0.736*** 0.709** 0.661 0.840***

0.120 0.165 0.242 0.318 0.190 0.233 0.212 0.517 0.220

Institutional quality 3.602*** 3.289*** 3.268*** 2.734*** 4.023*** 2.424*** 2.799*** 2.683*** 2.771***

0.475 0.531 0.585 1.085 0.651 1.297 1.028 1.013 0.894

Freedom 0.729** 0.943*** 0.685** 0.573** 0.762** 0.485 0.644 0.825 0.1350.322 0.337 0.270 0.553 0.363 0.499 0.509 0.545 0.320

Aid volatility 0.896** 0.796** 1.172*** 0.961*

0.425 0.407 0.468 0.586

Revenue volatility 0.908*** 0.454* 0.855** 0.645**

0.190 0.263 0.447 0.256

Public investment volatility 1.204*** 1.284***

0.309 0.215

# of countries 26 26 26 26 26 26 26 26 26# of observations 106 106 106 106 106 106 106 106 106PvaluesSargan statistic 0.53 0.61 0.38 0.37 0.37 0.30 0.37 0.45 0.37AR (1) (test for serial correlation) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00AR (2) (test for serial correlation) 0.40 0.59 0.50 0.52 0.50 0.50 0.56 0.59 0.50

Values in italics are robust standard errors with Whites correction for autocorrelation.* Significance at the 10% levels.** Significance at the 5% levels.*** Significance at the 1% level.

http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


7/10

of GDP) received. In addition, when the years of 200811 areexcluded, aids contribution increases to between 0.14% and0.22%. Furthermore, aids impact is at a high significance level(mostly one percentage) for the period 19922007, althoughthis decreases to between one and five percentages for the entiresample period, perhaps suggesting a slight weakening of theimpact of aid on GDP growth as a result of the recent financial

crisis. However, as discussed below it should be noted that theaid variable is no longer significant once the public investmentvariable is included (formulations six to nine).

The GMM results of the extent to which the volatility inpublic investment potentially counteracts the impact of aidon growth are shown in formulations five to nine ofTables2 and 3. We now investigate the extent to which the volatilityin public investment (whether resulting from volatility in aidflows, government revenues, or even the aggregate publicinvestment measure) potentially counteracts the impact ofaid on growth. In this regard, we expand the basic regression(formulation one) by testing various formulations that in-clude volatility variables related to aid, government revenue,and public investment. In accordance with Lensink andMorrissey (2000), volatility is measured as a four-year rolling

standard deviation of the change in the variable, andexpressed in logarithmic terms.

The three volatility variables included in the various formu-lations (government revenue, aid, and public investment), allhave the expected negative coefficients with respect to GDPgrowth in both samples. Revenue and aid volatilities appearto be significant at the one and five percentage levels while vol-atility of public investment is mostly only statistically signifi-cant at the five and ten percentage levels. Overall, theseresults suggest that volatilities in public investment, aid, andgovernment revenues are all significant obstacles to growth,which could explain why Sub-Saharan Africa countries haveexperienced low growth rates despite significant aid flows overthe covered period.

Formulations six to nine report the results of including pub-lic investment as an additional explanatory variable. Theobjective is to test whether aids impact on growth is throughvolume or through efficiency. When public investment is ex-cluded (formulations one to five), one plausible interpretationof the aid variable is that any impact aid has on growth isthrough volume. 23 Once public investment is controlled for(formulations six to nine) any significant impact of aid ongrowth will be through efficiency. The results show that whenpublic investment is included, there is a loss of significance forthe aid coefficient, as well as a general reduction in coefficientsize. This suggests that in Sub-Saharan African countries, aiddoes not have an efficiency effect on growth, but rather, anyeffect of aid is through volume instead (which accords withLensink & Morrissey, 2000).

The reader is also referred to Appendix B, which containsthe correlation matrixes for the variables of the regressionspecifications for, respectively, the periods 19922011, and19922007.

7. CONCLUSION

Foreign aid effectiveness has been a topic of debate in re-cent decades in academic discourse as well as the policy are-na. Renewed interest in the subject emerged as a parallel

debate following the proclamation of the Millennium Devel-opment Goals (MDGs). In this study, the effectiveness of aidas a contributor to economic growth was investigated for apanel of 26 Sub-Saharan Africa countries over the period19922011.

Our results indicate that aid has contributed to growth inSub-Saharan Africa over the time period covered, especially

in the study period preceding the financial crisis (19922007). This may, however, be linked to the fact that for oursample countries over this period average aid made up a lar-ger percentage of per capita GDP than for the period after2007.

However, we also find strong evidence to suggest that aidsimpact on growth may have been eroded by the volatility ofpublic investment (aids main transmission mechanism)which in turn may be linked to the volatility in governmentrevenues and aid itself. While this may support Moss et al.(2006), who argue that aid dis-incentivises the efficient collec-tion and administration of tax revenue, it also indicates thatin accordance withEbeke and Ehrhart (2011), a move awayfrom volatile trade taxes to the more stable domestic indirecttaxes could facilitate the reduction of some of this volatility

in public investment that is due to revenue volatility. Alter-natively SSA countries raise bond financing backed by secu-ritization of future-flows, such as remittances, tourismreceipts, and export receivables (Ratha, Mohapatra, & Plaza,2008).

Our finding that aid volatility has a negative impact on eco-nomic growth in Sub-Saharan Africa contradicts earlier re-search byLensink and Morrissey (2000) who argued that aidvolatility captures expected changes in aid and had no impacton economic growth. The policy implication of this is that at-tempts should be made to reduce the volatility of aid flows andthe instability that this brings into government revenues andplanning. One option could be multi-year aid agreements thatbring more stability in aid flows to specific countries. A spe-

cific innovation proposed byRatha et al. (2008)is a form ofsecuritization(or effective front-loading) of future aid com-mitments through the sale of bonds backed by these commit-mentsa system used by the International FinancingFacilities for Immunization (IFFIm) to make aid flows in sup-port of global immunization initiatives more reliable and pre-dictable.

Unsurprisingly, we find that institutional quality (composedin our methodology of measures of bureaucratic quality, lawand order, and corruption) is both a strong and significantvariable that affects GDP growth in our sample. This empha-sizes the continued importance of Sub-Saharan African coun-tries continuing to improve both their institutionalframeworks and capacity. Lastly, we find that, in accordancewithLensink and Morrissey (2000), aid does not have an effi-

ciency effect on growth, but rather, any effect of aid is throughvolume instead.

Although our research examined the link between economicgrowth, aid, and volatilities in aid, government revenues, andpublic investment as a whole, an investigation into the contri-bution of the volatilities of each source of public investment tothe overall volatility of public investment was beyond thescope of this study. This could be an interesting avenue forfurther research. In addition, future research could be dedi-cated to investigating the causes of the volatilities of varioussources of financing public investment.

http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


8/10

NOTES

1. Note that these authors found that greater volatility in taxes leads to areduction in actual level of public spending than would otherwise be thecase, most probably because of the given the uncertainty that it introducesinto the levels of government revenue available for such long-terminvestment.

2. The model is often referred to as the Harrod-Domar model.

3. For example, Mosley (1980) finds that aid was positively andsignificantly correlated with growth in his poorest country sample, butnegatively and insignificantly correlated in his middle income countrysample.

4. Their policy index comprised the budget surplus, the inflation rate,and a measure for trade openness.

5. See for exampleBoone (1996), Burnside and Dollar (1997), Burnsideand Dollar (2000), Tarp (2006), Roodman (2007), Doucouliagos andPaldam (2009), Hauk and Wacziarg (2009), Rajan and Subramanian(2011).

6. Note that most of the third generation studies with the exceptionofHansen and Tarp (2000), Hansen and Tarp (2001), Dalgaard et al.(2004), Roodman (2007)and Rajan and Subramanian (2005), Rajan andSubramanian (2008), employs either OLS or two-stage least squaresestimations (2SLS) procedures without fixed effects.

7. The full list of countries used in our study can be found inAppendixA. Note that the number of Sub-Saharan African countries in our samplewas determined by data constraints, but that the sample size comparesfavorably with comparative studies such as those by-Gomanee, Girma,and Morrissey (2005): 25 countries, Lensink and Morrissey (2000) : 36countries, and Clemens, Radelet & Bhavnani (2004) : 22 countries.

8. Measured as disbursements minus amortizations.

9. The EDA is the sum of the grant equivalents of all development flowsdisbursed in a given period.

10. SeeRenard and Cassimon (2001) for a critique of aid measures.

11. For the purposes of this paper, instability and volatility are usedinterchangeably.

12. This variable was discarded due to concerns over collinearity.

13. This variable is measured as a natural logarithm of the initial GDPfigure.

14. Aid squared was initially included in all formulations and the resultsof the variable were insignificant with an unexpected positive sign. For this

reason we dropped the variable from our analysis. Our results were notsensitive to this choice.

15. Private credit by deposit money banks will be referred to as privatecredit.

16. The ICRG model allows users to make their own risk assessmentsbased on the ICRG model or to customize the model to meet their uniquespecifications. If particular risk factors are of greater importance to theuser, composite risk ratings can be recalculated by giving greater weight tothe preferred factors.

17. See Sachs (2003) or Easterly and Levine (2003) for more on theinfluence of geography on economic development.

18. This variable was discarded due to concerns over collinearity.

19. Other methods considered were listwise deletion, which reduces thesample size and can introduce biases where representative cases areremoved, and mean imputation, which can reduce variances and impactrelationships between variables (Anderson, Basilevsky, & Hum, 1983).

20. Burnside and Dollar (2000)report coefficients ranging between 0.7and -0.84 in their sample of lower income countries, while Hepp (2008)reports coefficients of 0.8 - 1.5 of sample of Highly Indebted PoorCountries (HIPC).

21. For example, Burnside and Dollar (2000), or Dalgaard and Tarp(2001).

22. To justify aids conditional impact on growth, researchers interactedaid with several variables and found the interacted aid terms to besignificant while aid alone was not. See for example Burnside and Dollar(2000), Lensink and Morrissey (2000), Collier and Dollar (2002) andRajan and Subramanian (2008).

23. Admittedly, aids main transmission mechanism is public investment;and this indirect effect on growth is captured by the significant aidcoefficient in formulations 1-5. However, once we include publicinvestment, in formulations 6-9, then not surprisingly aid effectivenessis less significant since its main transmission mechanism is accounted for.

REFERENCES

Addison, T., Mavrotas, G., & McGillivray, M. (2005). Aid to Africa: Anunfinished agenda.Journal of International Development, 17, 9891001.

Anderson, A. B., Basilevsky, A., & Hum, D. P. J. (1983). Missing data: Areview of the literature. Handbook of survey research. San Diego:Academic Press, 415-494.

Bacha, E. (1990). A three-gap model of foreign transfers and the GDPgrowth rate in developing countries.Journal of Development Econom-ics, 279296.

Beck, T., & Demirguc-Kunt, A. (2010). Financial institutions and marketsacross countries and over time: The updated financial development andstructure database. World Bank Economic Review, 24(1), 7792.

Bloom, D., Sachs, J., Collier, P., & Udry, C. (1998). Geography,demography, and economic growth in Africa. Brookings Papers onEconomic Activity, 2, 207295.

Boone, P. (1996). Politics and the effectiveness of foreign aid. EuropeanEconomic Review, 40, 289329.

Bulr, A., & Hamann, J. (2008). Volatility of development aid: From thefrying pan into the fire. World Development, 36(10), 20482066.

Burnside, C., & Dollar, D. (1997). Aid, policies, and growth.World Bankpolicy research working paper 569252.

Burnside, C., & Dollar, D. (2000). Aid, policies, and growth. TheAmerican Economic Review, 90(4), 847868.

Chang, C., Fernandez-Arias, E., & Serven, L (1999). Measuring aid flows:A new approach. World Bank policy research working paper 2050 .

Chang, C., Fernandez-Arias, E., & Serven, L. (2002). Measuring aid flows:A new approach. Global Economic Quarterly, 3(2), 197218.

Chenery, H., & Strout, A. (1966). Foreign assistance and economicdevelopment.The American Economic Review, 679733.

http://-/?-http://-/?-http://refhub.elsevier.com/S0305-750X(13)00285-4/h0005http://refhub.elsevier.com/S0305-750X(13)00285-4/h0005http://refhub.elsevier.com/S0305-750X(13)00285-4/h0005http://refhub.elsevier.com/S0305-750X(13)00285-4/h0005http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://refhub.elsevier.com/S0305-750X(13)00285-4/h0035http://refhub.elsevier.com/S0305-750X(13)00285-4/h0035http://refhub.elsevier.com/S0305-750X(13)00285-4/h0035http://refhub.elsevier.com/S0305-750X(13)00285-4/h0035http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0050http://refhub.elsevier.com/S0305-750X(13)00285-4/h0050http://refhub.elsevier.com/S0305-750X(13)00285-4/h0050http://refhub.elsevier.com/S0305-750X(13)00285-4/h0050http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0065http://refhub.elsevier.com/S0305-750X(13)00285-4/h0065http://refhub.elsevier.com/S0305-750X(13)00285-4/h0065http://refhub.elsevier.com/S0305-750X(13)00285-4/h0065http://refhub.elsevier.com/S0305-750X(13)00285-4/h0065http://refhub.elsevier.com/S0305-750X(13)00285-4/h0065http://refhub.elsevier.com/S0305-750X(13)00285-4/h0065http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0060http://refhub.elsevier.com/S0305-750X(13)00285-4/h0050http://refhub.elsevier.com/S0305-750X(13)00285-4/h0050http://refhub.elsevier.com/S0305-750X(13)00285-4/h0050http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0040http://refhub.elsevier.com/S0305-750X(13)00285-4/h0035http://refhub.elsevier.com/S0305-750X(13)00285-4/h0035http://refhub.elsevier.com/S0305-750X(13)00285-4/h0035http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://refhub.elsevier.com/S0305-750X(13)00285-4/h0030http://-/?-http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://refhub.elsevier.com/S0305-750X(13)00285-4/h0025http://-/?-http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://refhub.elsevier.com/S0305-750X(13)00285-4/h0020http://-/?-http://-/?-http://refhub.elsevier.com/S0305-750X(13)00285-4/h0005http://refhub.elsevier.com/S0305-750X(13)00285-4/h0005http://refhub.elsevier.com/S0305-750X(13)00285-4/h0005http://-/?-http://-/?-http://-/?-http://-/?-


9/10

Clemens, M., Radelet, S., & Bhavnani, R. (2012). Counting chickens whenthey hatch: The short term effect of aid on growth. The EconomicJournal, 122(561), 590617.

Clements, B., Bhattacharya, R., & Quoc Nguyen, T. (2003). External debt,public investment, and growth in low-income countries.IMF workingpaper.

Collier, P., & Dollar, D. (2002). Aid allocation and poverty reduction.European Economic Review, 46, 14751500.

Collier, P., & Hoeffler, A. (2004). Aid, policy and growth in post-conflictsocieties. European Economic Review, 48, 11251145.Dalgaard, C. J., & Hansen, H. (2001). On aid, growth and good policies.

Journal of Development Studies, 37(6), 1741.Dalgaard, C. J., Hansen, H., & Tarp, F. (2004). On the empirics of foreign

aid and growth. The Economic Journal, 114, F191F216.Domar, E. (1947). Capital expansion, rate of growth, and employment.

Econometrica, 14, 137147.Doucouliagos, H., & Paldam, M. (2009). The aid effectiveness literature:

The sad results of 40 years of research. Journal of Economic Surveys,23(3), 433461.

Durbarry, R., Gemmell, N., & Greenaway, D. (1998). New evidence onthe impact of foreign aid on economic growth. Centre for Research inEconomic Development and International Trade working paper, 98(8).

Easterly, W. (2003). Can foreign aid buy growth?. The Journal ofEconomic Perspectives, 17(3), 2348.

Easterly, W., & Levine, R. (2003). Tropics, germs, and crops: how

endowments. Journal of Monetary Economics, 50, 339.Ebeke, C., & Ehrhart, H. (2011). Tax revenue instability in Sub-Saharan

Africa: Consequences and remedies. Journal of African Economies,21(1), 127.

Fielding, D., & Mavrotas, G. The volatility of aid, discussion paper,World Institute for Development Economics Research. (2005). http://ideas.repec.org/p/ess/wpaper/id3166.html.

Gomanee, K., Girma, S., & Morrissey, O. (2005). Aid and growth in Sub-Saharan Africa: Accounting for transmission mechanisms.Journal ofInternational Development, 10551075.

Hadjimichael, M. T., Dhaneshwar, G., Muhleisen, M., Nord, R., & Ucer,E. M. (1995). Sub-Saharan Africa: Growth, savings, and investment,198693. IMF occasional paper 118, International Monetary Fund,Washington, DC.

Hansen, H., & Tarp, F. (2000). Aid effectiveness disputed. Journal ofInternational Development, 12, 375398.

Hansen, H., & Tarp, F. (2001). Aid and growth regressions. Journal ofDevelopment Economics, 64, 547570.

Hauk, W., & Wacziarg, W. (2009). A Monte Carlo study of growthregressions. Journal of Economic Growth, 103147.

Hepp, R (2008). Can debt relief buy growth? Fordham EconomicsDiscussion Paper Series dp2008-22, Fordham University, Departmentof Economics.

Honaker, J., & King, G. (2010). What to do about missing values in timeseries cross-section data. American Journal of Political Science, 54(3),561581.

Hudson, J., & Mosley, P. (2008). The macroeconomic impact of aidvolatility. Economic Letters, 99, 486489.

Knack, S., & Keefer, P. (1995). Institutions and economic performance:Cross-country tests using alternative institutional measures. Economicsand Politics, 7(3), 207227.

Lensink, R., & Morrissey, O. (2000). Aid instability as a measure ofuncertainty and the positive impact of aid on growth. Journal of

Development Studies, 36(3), 3149.Levine, R., & Zervos, S. (1998). Stock markets, banks, and economicgrowth. American Economic Review, 537558.

Levy, V. (1988). Aid and growth in Sub-Saharan Africa: The recentexperience.European Economic Review, 32(9), 17771795.

Morrissey, O. (2001). Does aid increase growth?. Progress in DevelopmentStudies, 1, 3750.

McQuinn, K., & Whelan, K. (2006). Conditional convergence revisited:Taking Solow very seriously. Central Bank and Financial ServicesAuthority of Ireland.

Mosley, P. (1980). Aid, savings and growth revisited. Oxford Bulletin ofEconomics and Statistics, 42(2), 7995.

Moss, T., Pettersson, G., & Van de Walle, N (2006). An aid-institutions

paradox? A review essay on aid dependency and state building in Sub-Saharan Africa. Center for Global Development working paper no. 74 .OECD. Development: Aid to developing countries falls because of global

recession. (2012). http://www.oecd.org/newsroom/developmentaidto-developingcountriesfallsbecauseofglobalrecession.htm.

Pattillo, C., Poirson, H., & Ricci, L. (2002). External debt and growth.IMF working paper.

Rajan, R., & Subramanian, A. (2005). What undermines aids impact ongrowth? National Bureau of Economic Research working paper no.11657.

Rajan, R., & Subramanian, A. (2008). Aid and growth: What does thecross-country evidence really show?. The Review of Economics andStatistics, 90(4), 643665.

Rajan, R., & Subramanian, A. (2011). Aid, Dutch disease, and manufac-turing growth. Journal of Development Economics, 94, 106118.

Ratha, D., Mohapatra, S., & Plaza, S. (2008). New sources and innovativemechanisms for financing development in Sub-Saharan Africa.World

Bank policy research working paper WPS4609.Renard, R., & Cassimon, D. (2001). On the pitfalls of measuring aid.

United Nations University World Institute for Development EconomicResearch Discussion paper no. 2001/69.

Roodman, D. (2007). The anarchy of numbers: Aid, development, andcross-country empirics. The World Bank Economic Review, 21(2),255277.

Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. NewYork: Wiley.

Sachs, J. (2003). Institutions dont rule: Direct effects of geography on percapita income. National Bureau of Economic Research working paperno. 9490.

Solow, R. (1956). A contribution to the theory of economic growth.Quarterly Journal of Economics, 70, 6594.

Tarp, F. (2006). Aid and development. Swedish Economic Policy Review,13, 961.

APPENDIX A. LIST OF SAMPLE COUNTRIES

The following countries are included in this study:Botswana, Burkina Faso, Cameroon, Republic of the Con-

go, Cote dIvoire, Ethiopia, Gabon, The Gambia, Ghana,Guinea, Guinea-Bissau, Kenya, Madagascar, Malawi, Mali,Mozambique, Nigeria, Senegal, Sierra Leone, South Africa,Sudan, Tanzania, Togo, Uganda, Zambia, Zimbabwe.

APPENDIX B. CORRELATION MATRICES

Table 4Table 5

http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070http://refhub.elsevier.com/S0305-750X(13)00285-4/h0080http://refhub.elsevier.com/S0305-750X(13)00285-4/h0080http://refhub.elsevier.com/S0305-750X(13)00285-4/h0080http://refhub.elsevier.com/S0305-750X(13)00285-4/h0085http://refhub.elsevier.com/S0305-750X(13)00285-4/h0085http://refhub.elsevier.com/S0305-750X(13)00285-4/h0085http://refhub.elsevier.com/S0305-750X(13)00285-4/h0085http://refhub.elsevier.com/S0305-750X(13)00285-4/h0090http://refhub.elsevier.com/S0305-750X(13)00285-4/h0090http://refhub.elsevier.com/S0305-750X(13)00285-4/h0090http://refhub.elsevier.com/S0305-750X(13)00285-4/h0095http://refhub.elsevier.com/S0305-750X(13)00285-4/h0095http://refhub.elsevier.com/S0305-750X(13)00285-4/h0095http://refhub.elsevier.com/S0305-750X(13)00285-4/h0095http://refhub.elsevier.com/S0305-750X(13)00285-4/h0100http://refhub.elsevier.com/S0305-750X(13)00285-4/h0100http://refhub.elsevier.com/S0305-750X(13)00285-4/h0100http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0115http://refhub.elsevier.com/S0305-750X(13)00285-4/h0115http://refhub.elsevier.com/S0305-750X(13)00285-4/h0115http://refhub.elsevier.com/S0305-750X(13)00285-4/h0115http://refhub.elsevier.com/S0305-750X(13)00285-4/h0120http://refhub.elsevier.com/S0305-750X(13)00285-4/h0120http://refhub.elsevier.com/S0305-750X(13)00285-4/h0120http://refhub.elsevier.com/S0305-750X(13)00285-4/h0120http://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://ideas.repec.org/p/ess/wpaper/id3166.htmlhttp://ideas.repec.org/p/ess/wpaper/id3166.htmlhttp://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://refhub.elsevier.com/S0305-750X(13)00285-4/h0145http://refhub.elsevier.com/S0305-750X(13)00285-4/h0145http://refhub.elsevier.com/S0305-750X(13)00285-4/h0145http://refhub.elsevier.com/S0305-750X(13)00285-4/h0145http://refhub.elsevier.com/S0305-750X(13)00285-4/h0150http://refhub.elsevier.com/S0305-750X(13)00285-4/h0150http://refhub.elsevier.com/S0305-750X(13)00285-4/h0150http://refhub.elsevier.com/S0305-750X(13)00285-4/h0150http://refhub.elsevier.com/S0305-750X(13)00285-4/h0155http://refhub.elsevier.com/S0305-750X(13)00285-4/h0155http://refhub.elsevier.com/S0305-750X(13)00285-4/h0155http://refhub.elsevier.com/S0305-750X(13)00285-4/h0155http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h9000http://refhub.elsevier.com/S0305-750X(13)00285-4/h9000http://refhub.elsevier.com/S0305-750X(13)00285-4/h9000http://refhub.elsevier.com/S0305-750X(13)00285-4/h9000http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0180http://refhub.elsevier.com/S0305-750X(13)00285-4/h0180http://refhub.elsevier.com/S0305-750X(13)00285-4/h0180http://refhub.elsevier.com/S0305-750X(13)00285-4/h0180http://refhub.elsevier.com/S0305-750X(13)00285-4/h0185http://refhub.elsevier.com/S0305-750X(13)00285-4/h0185http://refhub.elsevier.com/S0305-750X(13)00285-4/h0185http://refhub.elsevier.com/S0305-750X(13)00285-4/h0185http://refhub.elsevier.com/S0305-750X(13)00285-4/h9005http://refhub.elsevier.com/S0305-750X(13)00285-4/h9005http://refhub.elsevier.com/S0305-750X(13)00285-4/h9005http://refhub.elsevier.com/S0305-750X(13)00285-4/h9005http://refhub.elsevier.com/S0305-750X(13)00285-4/h0195http://refhub.elsevier.com/S0305-750X(13)00285-4/h0195http://refhub.elsevier.com/S0305-750X(13)00285-4/h0195http://refhub.elsevier.com/S0305-750X(13)00285-4/h0195http://www.oecd.org/newsroom/developmentaidtodevelopingcountriesfallsbecauseofglobalrecession.htmhttp://www.oecd.org/newsroom/developmentaidtodevelopingcountriesfallsbecauseofglobalrecession.htmhttp://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://refhub.elsevier.com/S0305-750X(13)00285-4/h0230http://refhub.elsevier.com/S0305-750X(13)00285-4/h0230http://refhub.elsevier.com/S0305-750X(13)00285-4/h0230http://refhub.elsevier.com/S0305-750X(13)00285-4/h0230http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0250http://refhub.elsevier.com/S0305-750X(13)00285-4/h0250http://refhub.elsevier.com/S0305-750X(13)00285-4/h0250http://refhub.elsevier.com/S0305-750X(13)00285-4/h0250http://refhub.elsevier.com/S0305-750X(13)00285-4/h0260http://refhub.elsevier.com/S0305-750X(13)00285-4/h0260http://refhub.elsevier.com/S0305-750X(13)00285-4/h0260http://refhub.elsevier.com/S0305-750X(13)00285-4/h0265http://refhub.elsevier.com/S0305-750X(13)00285-4/h0265http://refhub.elsevier.com/S0305-750X(13)00285-4/h0265http://refhub.elsevier.com/S0305-750X(13)00285-4/h0265http://refhub.elsevier.com/S0305-750X(13)00285-4/h0265http://refhub.elsevier.com/S0305-750X(13)00285-4/h0265http://refhub.elsevier.com/S0305-750X(13)00285-4/h0265http://refhub.elsevier.com/S0305-750X(13)00285-4/h0260http://refhub.elsevier.com/S0305-750X(13)00285-4/h0260http://refhub.elsevier.com/S0305-750X(13)00285-4/h0260http://refhub.elsevier.com/S0305-750X(13)00285-4/h0250http://refhub.elsevier.com/S0305-750X(13)00285-4/h0250http://refhub.elsevier.com/S0305-750X(13)00285-4/h0250http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0245http://refhub.elsevier.com/S0305-750X(13)00285-4/h0230http://refhub.elsevier.com/S0305-750X(13)00285-4/h0230http://refhub.elsevier.com/S0305-750X(13)00285-4/h0230http://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://refhub.elsevier.com/S0305-750X(13)00285-4/h0225http://www.oecd.org/newsroom/developmentaidtodevelopingcountriesfallsbecauseofglobalrecession.htmhttp://www.oecd.org/newsroom/developmentaidtodevelopingcountriesfallsbecauseofglobalrecession.htmhttp://refhub.elsevier.com/S0305-750X(13)00285-4/h0195http://refhub.elsevier.com/S0305-750X(13)00285-4/h0195http://refhub.elsevier.com/S0305-750X(13)00285-4/h0195http://refhub.elsevier.com/S0305-750X(13)00285-4/h9005http://refhub.elsevier.com/S0305-750X(13)00285-4/h9005http://refhub.elsevier.com/S0305-750X(13)00285-4/h9005http://refhub.elsevier.com/S0305-750X(13)00285-4/h0185http://refhub.elsevier.com/S0305-750X(13)00285-4/h0185http://refhub.elsevier.com/S0305-750X(13)00285-4/h0185http://refhub.elsevier.com/S0305-750X(13)00285-4/h0180http://refhub.elsevier.com/S0305-750X(13)00285-4/h0180http://refhub.elsevier.com/S0305-750X(13)00285-4/h0180http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0175http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h0170http://refhub.elsevier.com/S0305-750X(13)00285-4/h9000http://refhub.elsevier.com/S0305-750X(13)00285-4/h9000http://refhub.elsevier.com/S0305-750X(13)00285-4/h9000http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h0165http://refhub.elsevier.com/S0305-750X(13)00285-4/h0155http://refhub.elsevier.com/S0305-750X(13)00285-4/h0155http://refhub.elsevier.com/S0305-750X(13)00285-4/h0155http://refhub.elsevier.com/S0305-750X(13)00285-4/h0150http://refhub.elsevier.com/S0305-750X(13)00285-4/h0150http://refhub.elsevier.com/S0305-750X(13)00285-4/h0150http://refhub.elsevier.com/S0305-750X(13)00285-4/h0145http://refhub.elsevier.com/S0305-750X(13)00285-4/h0145http://refhub.elsevier.com/S0305-750X(13)00285-4/h0145http://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://refhub.elsevier.com/S0305-750X(13)00285-4/h0135http://ideas.repec.org/p/ess/wpaper/id3166.htmlhttp://ideas.repec.org/p/ess/wpaper/id3166.htmlhttp://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://refhub.elsevier.com/S0305-750X(13)00285-4/h0125http://refhub.elsevier.com/S0305-750X(13)00285-4/h0120http://refhub.elsevier.com/S0305-750X(13)00285-4/h0120http://refhub.elsevier.com/S0305-750X(13)00285-4/h0120http://refhub.elsevier.com/S0305-750X(13)00285-4/h0115http://refhub.elsevier.com/S0305-750X(13)00285-4/h0115http://refhub.elsevier.com/S0305-750X(13)00285-4/h0115http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0105http://refhub.elsevier.com/S0305-750X(13)00285-4/h0100http://refhub.elsevier.com/S0305-750X(13)00285-4/h0100http://refhub.elsevier.com/S0305-750X(13)00285-4/h0100http://refhub.elsevier.com/S0305-750X(13)00285-4/h0095http://refhub.elsevier.com/S0305-750X(13)00285-4/h0095http://refhub.elsevier.com/S0305-750X(13)00285-4/h0095http://refhub.elsevier.com/S0305-750X(13)00285-4/h0090http://refhub.elsevier.com/S0305-750X(13)00285-4/h0090http://refhub.elsevier.com/S0305-750X(13)00285-4/h0090http://refhub.elsevier.com/S0305-750X(13)00285-4/h0085http://refhub.elsevier.com/S0305-750X(13)00285-4/h0085http://refhub.elsevier.com/S0305-750X(13)00285-4/h0085http://refhub.elsevier.com/S0305-750X(13)00285-4/h0080http://refhub.elsevier.com/S0305-750X(13)00285-4/h0080http://refhub.elsevier.com/S0305-750X(13)00285-4/h0080http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070http://refhub.elsevier.com/S0305-750X(13)00285-4/h0070


10/10

ScienceDirect

Available online at www.sciencedirect.com

Table 4. Correlation matrix (19922011)

Aid Aidvolatility

Revenue Freedom Growth Per capitaGDP t1


Publicinvestment

Publicinvestment volatility

Privatecredit

Revenuevolatility

Aid 1.00 0.66 0.36 0.07 0.05 0.47 0.12 0.52 0.41 0.31 0.23Aid volatility 0.66 1.00 0.44 0.21 0.12 0.59 0.14 0.27 0.49 0.62 0.23Revenue 0.36 0.44 1.00 0.34 0.05 0.44 0.19 0.09 0.07 0.25 0.59Freedom 0.07 0.21 0.34 1.00 0.22 0.27 0.22 0.31 0.07 0.33 0.07

Growth 0.05 0.12 0.05 0.22 1.00 0.02 0.12 0.27 0.02 0.05 0.12Per capita GDP t1 0.47 0.59 0.44 0.27 0.02 1.00 0.19 0.03 0.17 0.39 0.24Institutional quality 0.12 0.14 0.19 0.22 0.12 0.19 1.00 0.13 0.01 0.11 0.07Public investment 0.52 0.27 0.09 0.31 0.27 0.03 0.13 1.00 0.54 0.11 0.06Public investmentvolatility

0.41 0.49 0.07 0.07 0.02 0.17 0.01 0.54 1.00 0.18 0.06

Private credit 0.31 0.62 0.25 0.33 0.05 0.39 0.11 0.11 0.18 1.00 0.15Revenue volatility 0.23 0.23 0.59 0.07 0.12 0.24 0.07 0.06 0.06 0.15 1.00

Table 5. Correlation matrix (19922007)

Aid Aidvolatility

Revenue Freedom Growth Per capitaGDP t1


Publicinvestment

Publicinvestment volatility

Privatecredit

Revenuevolatility

Aid 1.00 0.68 0.35 0.08 0.05 0.46 0.15 0.58 0.46 0.30 0.24Aid volatility 0.68 1.00 0.47 0.21 0.12 0.60 0.17 0.35 0.56 0.60 0.22Revenue 0.35 0.47 1.00 0.36 0.04 0.44 0.21 0.10 0.11 0.26 0.59Freedom 0.08 0.21 0.36 1.00 0.24 0.29 0.21 0.35 0.05 0.30 0.11Growth 0.05 0.12 0.04 0.24 1.00 0.00 0.12 0.25 0.06 0.04 0.20Per capita GDP t1 0.46 0.60 0.44 0.29 0.00 1.00 0.19 0.09 0.24 0.36 0.21Institutional quality 0.15 0.17 0.21 0.21 0.12 0.19 1.00 0.12 0.02 0.12 0.08Public investment 0.58 0.35 0.10 0.35 0.25 0.09 0.12 1.00 0.57 0.16 0.12Public investmentvolatility

0.46 0.56 0.11 0.05 0.06 0.24 0.02 0.57 1.00 0.24 0.11

Private credit 0.30 0.60 0.26 0.30 0.04 0.36 0.12 0.16 0.24 1.00 0.13Revenue volatility 0.24 0.22 0.59 0.11 0.20 0.21 0.08 0.12 0.11 0.13 1.00


Aid Volatility in Africa

Documents

Transcript of Aid Volatility in Africa