The next step in African development: Aid, investment, or another round of debt? Michael Nicholson * Sarah Lane ** U.S. Agency for International Development November 2012 Abstract Amidst intense debt relief, and alongside dramatically improved governance, investment and growth increased substantially across Africa during the past decade. This paper interprets the timing of the Heavily Indebted Poor Countries (HIPC) Initiative, launched by the IMF and World Bank in the late 1990s, as a natural experiment to see whether these positive trends were specific to Africa, or specific to HIPC countries, as well as whether debt relief itself manifests deeper structural shifts in economic governance. As many HIPC countries are presently raising their external public debt levels, we question whether these loans would be a “good kind of debt” that leads to investment and development or the beginning of a new debt cycle leading to future needs of another round of debt relief programs. Data on external debt and capital development for 46 countries of sub‐Saharan Africa and six other HIPC countries outside of Africa is used to evaluate structural breaks and parameter stability in a longitudinal panel analysis. Incorporating an identification strategy that isolates the debt relief initiatives from endogenous improvements to economic governance, we find that they had a statistically significant impact on foreign investment flows to Africa. We conclude that despite currently escalating debt levels, the data suggests investment will likely be the next step in African development. Keywords: debt relief, foreign direct investment, panel‐data econometrics, foreign aid JEL Classification: F34, E22, C23 * Office of Economic Growth, USAID/Liberia, 502 Benson Street, Mamba Point, Monrovia, Liberia, [email protected]. ** Office of Economic Growth, USAID/Washington, 1300 Pennsylvania Ave NW, Washington DC, [email protected]. The views expressed in here are those of the authors and do not necessarily reflect those of the U.S. Agency for International Development. We thank participants at the 2012 Western Economic Association International in San Francisco, CA, the Analytical Skills Session in Monrovia, Liberia, and at the Office of Competition and Economic Analysis and the Overseas Private Investment Corporation in Washington, DC. Yonatan Woldu of the USAID Knowledge Services Center provided valuable research assistance. Narek Sahakyan, Sheila Desai, Charles Richter, Olesia Kozlova, and Joseph Flynn offered helpful comments.
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
ThenextstepinAfricandevelopment:
Aid,investment,oranotherroundofdebt?
MichaelNicholson*SarahLane**
U.S.AgencyforInternationalDevelopment
November2012
Abstract
Amidst intense debt relief, and alongside dramatically improved governance, investment andgrowthincreasedsubstantiallyacrossAfricaduringthepastdecade.ThispaperinterpretsthetimingoftheHeavilyIndebtedPoorCountries(HIPC)Initiative,launchedbytheIMFandWorldBankinthelate1990s,asanatural experiment to seewhether thesepositive trendswere specific toAfrica,orspecific toHIPC countries,aswellaswhetherdebt relief itselfmanifestsdeeper structural shifts ineconomicgovernance.AsmanyHIPCcountriesarepresentlyraisingtheirexternalpublicdebtlevels,we question whether these loans would be a “good kind of debt” that leads to investment anddevelopmentor thebeginningofanewdebtcycle leading to futureneedsofanotherroundofdebtreliefprograms.Dataonexternaldebtandcapitaldevelopmentfor46countriesofsub‐SaharanAfricaand sixotherHIPC countriesoutsideofAfrica isused toevaluate structuralbreaksandparameterstabilityinalongitudinalpanelanalysis.Incorporatinganidentificationstrategythatisolatesthedebtrelief initiatives from endogenous improvements to economic governance,we find that they had astatistically significant impact on foreign investment flows to Africa. We conclude that despitecurrentlyescalatingdebt levels, thedata suggests investmentwill likelybe thenext step inAfricandevelopment.
Keywords:debtrelief,foreigndirectinvestment,panel‐dataeconometrics,foreignaidJELClassification:F34,E22,C23 *OfficeofEconomicGrowth,USAID/Liberia,502BensonStreet,MambaPoint,Monrovia,Liberia,[email protected].**OfficeofEconomicGrowth,USAID/Washington,1300PennsylvaniaAveNW,WashingtonDC,[email protected]. The views expressed in here are those of the authors and do not necessarily reflect those of the U.S. Agency forInternational Development. We thank participants at the 2012 Western Economic Association International in SanFrancisco,CA,theAnalyticalSkillsSessioninMonrovia,Liberia,andattheOfficeofCompetitionandEconomicAnalysisandtheOverseasPrivateInvestmentCorporationinWashington,DC.YonatanWolduoftheUSAIDKnowledgeServicesCenterprovidedvaluableresearchassistance.NarekSahakyan,SheilaDesai,CharlesRichter,OlesiaKozlova,andJosephFlynnofferedhelpfulcomments.
1
“You control the debt, you control everything. You find this upsetting, yes? But this is the very essence of the banking industry, to make us all,
whether we be nations or individuals, slaves to debt.”
-Umberto Calvini, The International
1.IntroductionThecycleofexternaldebtforAfricangovernmentshasburdenedtheircountriessincethe
waveof independence in the1950sand1960s. Continuingefforts atdebt relief through the20thcentury proved ineffectual, as governments, tending to optimize on a short time horizon,wouldcontinuetoborrowintoindebtedness.1Times,however,maybechanging,assuccessstoriesaboutinvestmentopportunitiesontheAfricancontinentabound.2
Thispaper investigates theapparent transition from impoverishingdebt tomarket‐basedinvestmentinAfrica,specificallyfocusingontheHeavilyIndebtedPoorCountries(HIPC)InitiativelaunchedbytheInternationalMonetaryFund(IMF)andWorldBankinthelate1990s.Ourresultssuggest that a structuralbreak in thenatureofAfricandebt and investmentoccurred in the lastdecade. Debt relief may have eliminated a disequilibrium of instability to create incentives forprofit‐oriented commercial loans.3 These loans would be a “good kind of debt” leading toinvestmentanddevelopment.
Weincorporateanidentificationstrategythatisolatesdebtrelieffromaccompanyingpolicyimprovements of economic governance. Endogeneity arises because HIPC initiatives occurredsimultaneously with adoption of the exact types of policies that improve a country’scompetitivenessasadestinationforforeigndirectinvestment(FDI).Moreover,thesepolicieswereoftenadoptedpreciselytoreachthe“decisionpoint”ofthedebtrelief,makingitdifficulttoteaseoutwhetherthepolicies,thedebtrelief,orbothleadtoastructuralshift.Thisinteractioncausesaninherentchallengefortheanalysis,adilemmasimilartoisolatingtheimpactofWTOaccessiononacountry’s trade levels since the accession process deliberately focuses on the policies andcommercialenvironmentthatfacilitatesinternationaltradeasarun‐uptoWTOmembership.4
OureconometricanalysistreatstheHIPCdebtreliefprogramasaneventstudybycreatingcontrolgroupsforHIPCcountriesandAfricancountries.Thesetwocomparisonsetsoflongitudinalpanel data allow us to identify the specific impacts of HIPC while isolating pan‐African geo‐economicand–politicaltrends.Theresultsshowthatfollowingdebtrelief,determinantsofFDIinHIPC‐designatedcountriesreflectthoseofothercountriesintheworld.Wealsoshowthatthatsub‐SaharanAfricaasageo‐politicaldesignationforeconomicactivitydoesnotmatter:ifAfrica’srecenteconomichistorydiffersfromtherestoftheworld,itappearsduetothefeaturesofbeingpoorandindebted rather than something inherent to the continent. The HIPC initiative, which includedpolicyimprovementsanddebtrelief,alteredtheeconomiclandscapeofAfrica,and,asaresult,FDIhasflowedintothecontinent. 1SeeEasterly(2002)foradescriptionofdebtreliefeffortsinfromthe1970stothenewmillennium.2Unlessotherwiseindicated,by“Africa”or“thecontinent”wereferhereintothe49countriesofsub‐SaharanAfrica.3SeeAddison(2006).4See,forexample,Rose(2004).
2
Theimplicationsarethatfollowingthedebtreliefinitiatives,wecananticipatethatFDIwillflowtoAfricabysimilarmechanismsasinotherregions.Ifthecontinentis“openforbusiness”,theability for countries to again engage in public financing to improvemarket conditions should beanticipatedasapositiveturnofevents.
The next section provides background on debt relief measures from the decade of the2000s. Section 3 discusses trends in private investment in Africa. Section 4 provides aneconometricanalysisofthedataondebtandinvestmentinAfrica,andsection5offersconcludingremarks.
2.LiteratureReviewandContext2.1AfricaRising
Theresultsof thispapersuggest that foreignaid in the formofdebtreliefconditionalonimprovedeconomicgovernancehasleadtoincreasedforeigninvestmentandcapitalformationonthe continent.5 These effects are evident in how private markets now view the continent. Forexample,McKinsey(2010)identifiedseveralAfrican“lions”thathaverecentlyexperiencedgrowthaccelerationandincreasedeconomicmomentum.Thereporthighlightsagroupofcountriesinsub‐Saharan Africa with specific commercial opportunities due to a combination of export‐orientedeconomiesandthediversificationofgrossdomesticproduct(GDP).AsAfricabecomesaprofitableinvestmentdestination,governmentsinformerHIPCcountrieshavebeguntoincreasedebtlevels,ostensibly for investment to support growth opportunities. This new borrowing may notnecessarily lead to a new debt cycle, if a wave of fundamental changes in economic policy anddemocraticgovernancehastakenplaceacrossthecontinent,aswepostulate.
Radelet (2010) describes 17 countries in Africa as “success stories” due to sustainedeconomic growth over the past decade.He identifies five specific reasonswhy emergingAfricancountrieshavebeengrowing:(1)democracy;(2)economicpolicies;(3)debtrelief;(4)technology;and(5)entrepreneurship.Radeletstatesthatthecombinationofthesefivefactorshelpedtobringthesecountriesoutofa low‐growthequilibrium,andthatprogresshasbeenmadeinavarietyofsocial, governanceandeconomic factors thathavehelpedemergingAfricancountriesgrowmorerapidly.Manyofthesefactorsreinforcedeachother;forexample,eliminatingadebt‐to‐GDPratioofnearly 3,000 percent, as Liberia did in the mid‐2000s, necessitates better economic policies ingeneralandbenefitssubstantiallyfromanendtoconflictandwelfare.RadeletcitesthecauseofthedebtcrisisinAfricaastheresultof“pooreconomicmanagement,unaccountableandhighlycorruptgovernments,largeamountsofborrowingatgovernmentrates,andadeepglobaleconomicshock.”He states that the fundamental shift for emerging Africa came from reforms in democratic andeconomicpolicies,butthegrowthhasbeensustainableasaresultofdebtrelief,newtechnologiesandentrepreneurship.
Clements, Bhattacharyua, andNguyen (2005) suggest thatWestern aid agencies initiatedtheHIPCprogramonthebasisthatdebtburdenintheheavilyindebtednationswasstiflinggrowthand creating impossible conditions for the countries to eradicatepoverty.The literatureondebt
5Overthelastseveralyearsmanyskeptics,likeEasterly(2001,2004)andMoyo(2010),havecriticizedaidinAfricaasineffective.
3
and economic growth holds that unsustainable debt‐to‐GDP stock creates disincentives forinvestmentinthedomesticeconomybecausethegovernmentmustdivertrevenuetoservicedebtratherthaninvest inthedomesticeconomy;risingdebt levelsthenincreaseinvestoruncertainty.Unsustainabledebt levels are thus linkedwith inflation‐generatingmonetarypolicy andoutrightexpropriation of private firms. Under such uncertainties investments tend to be in projects of ashorter duration and quicker returns and thus countries are less apt to invest in long termsustainable foreign direct investment. Furthermore, debt overhang has the tendency to delaynecessarymacro‐fiscalpolicyreform.
Dorsey (2008) demonstrates the extent that capital flowed to debt‐forgiven countriesfollowing debt relief. The aggregate current account deficit of low income countries fell from 3percent in the mid‐1990s to a near balance in 2006. Following the debt forgiveness plans,developmentaidhastakentheformofgrantsinsteadoflending.Dorsey(2008)alsopointsoutthattheAfricancountries thathavemostbenefited from this infloware thosewithextractive sectorssuchasoilandmineralwealth.FDIflowstonon‐extractive‐basedeconomieshaveincreased,dueinparttoconcurrentpoliciessuchastheliberalizationoftheeconomyandageneralopeningofthemarkettoforeigninvestors.
AsEasterly(2002)suggested,inadvanceoftheHIPCInitiative,debtreliefshouldbeofferedonly to governments that have demonstrated a clear shift in their orientation towardsmacroeconomic policy; otherwise, the program would essentially reward poor economicgovernance. To be eligible for the program, countries had to meet a set of conditions, both toimprove their economies andgovernance to be eligible for debt relief. If this hasheld true, debtreliefmaybea“proxy‐plus” for fundamentalchanges ineconomicgovernanceundertakenbythedebtreliefrecipients.
2.2DebtandDebtRelief
InAugustof1982,theyearthatMexicodefaultedonitssovereigndebttriggeringaglobaldebt crisis, the average external debt‐to‐export ratio for sub‐Saharan Africa was 325 percent,substantiallyhigherthanthe“heavilyindebted”thresholdof150percent.Africandebtwasstiflingdevelopmentefforts. Inthemid‐1980s,theParisClubbegantoaddressthesedebt issues, firstbyrestructuring loans and then through debt forgiveness.6While the Paris Clubwasmaking someprogress in debt relief, debt in Africa continued to soar. By 1992, the average external debt‐to‐export ratio for sub‐Saharan Africa countries reached 760 percent, with many countriesexperiencing external debt‐to‐export ratios far higher. The most extreme case of indebtednessoccurredinLiberiawhere,in1997,thedebt‐to‐exportratioreachedanastonishing9,212percent.However, Liberia was not alone in Africa in having an extremely high debt burden. Since thebeginningof thedebtcrisis in1982more thanadozenAfricancountrieshadpersistentexternaldebt‐to‐exportratioofmorethan1,000percent.7SeeTableA‐1intheAppendix.
6http://www.clubdeparis.org/sections/composition/historique‐50‐ans7Burundi,DRC,Eritrea,Ethiopia,Guinea‐Bissau,Liberia,Mozambique,Rwanda,SierraLeone,Somalia,Sudan,Tanzania,andUgandaallhaddebt‐to‐exportratiosofmorethan1,000percentforatleast3years.
4
In1996, theWorldBankand IMF launched theHIPC initiativewith thegoalof “ensuringthatnopoorcountryfacesadebtburdenitcannotmanage.”8Ofthe39countrieseligibleforHIPC,33wereinAfrica,asshowninFigure1.9TheHIPCinitiativewasanewefforttoprovidedebtrelieftocountries thatweremakingstrongefforts toaddresspooreconomicandpoliticalpolicies.TheinitiativewasatwostageprocessconsistingofaDecisionPointstageandaCompletionPointstage.To reach a Decision Point, countries must have established a track record of macroeconomicstability and completed a Poverty Reduction Strategy Paper (PRSP). In this phase, the countryreceivedprovisionaldebtrelief.TobeeligiblefortheCompletionPointstage,acountryneededtohavemaintainedmacroeconomicstability,begunimplementingreformsagreedtoat theDecisionPointandhaveimplementedthePRSPforatleastoneyear.AttheCompletionPointstage,countriesreceived the full debt relief agreed to at the Decision Point. In order to receive relief, the HIPCinitiativerequiredcountriestoaddresstheproblemsthatledtodebt.
Figure1:HIPCcountriesinAfrica
In2005,theIMFandWorldBankinitiatedtheMultilateralDebtReliefInitiative(MDRI)tofurther increase debt relief to reach theMillenniumDevelopment goals. MDRI provides for 100percent debt relief to eligible countries on debt from the IMF, World Bank and the AfricanDevelopmentBank.SeeFigure2.
8http://www.imf.org/external/np/exr/facts/hipc.htm9TheotherHIPCcountrieswereAfghanistan,Bolivia,Guyana,Haiti,HondurasandNicaragua.
5
Figure2:TotalDebtRelief/GDP(2009),allcountries
Debt relief for theWestAfricancountryofLiberia is, literally,off thecharts.Toallow for
visibility in a comparison to other recipients of debt relief inAfrica, Figure 3 provides the sameinformationasFigure2withLiberiaexcludedtoadjustfortheoutlier.
Figure3:TotalDebtRelief/GDP(2009),Liberiaexcluded
100%
50%
25%
0%
50%
100%
150%
200%
250%
300%
350%
400%
LIB
STP
GBS
DRC
BRD
SLE
CAR
MAU
MLW
MOZ
GAM
ZAM
MAD
NIG
COM
TAN
TOG
UGD
RWA
MLI
GHA
CGO
BKF
GUI
SEN
BEN
ETH
CDI
CAM
CHD
HIPC/GDP MDRI/GDP
100%
50%
25%
0%
20%
40%
60%
80%
100%
120%
STP
GBS
DRC
BRD
SLE
CAR
MAU
MLW
MOZ
GAM
ZAM
MAD
NIG
COM
TAN
TOG
UGD
RWA
MLI
GHA
CGO
BKF
GUI
SEN
BEN
ETH
CDI
CAM
CHD
6
ThirtyAfricannationsreceivednearly$50billionindebtreliefaspartoftheHIPCinitiativeandanother$24billionaspartofMDRIbetween2000and2009.10TheimpactofthisdebtreliefisillustratedinFigure4,whichshowsadramaticdeclineinexternaldebtaspercentageofGDPfrommore than 75 percent in 1994 to less than 20 percent by 2000. As could be expected, Africanexternaldebt fellsubstantially immediately followingtherelief initiatives.Figure4showsacleardownwardtrendfrom1995to2006,butsincethen,theamountofexternaldebtisagainincreasingtomorethanitwasintheheightofthedebtcrisis.Thedebt‐to‐GDPratiohasbeenquiteflatforthelastfiveyears,andin2010,theaverageexternaldebt‐to‐exportratioforAfricancountrieswas136percent,belowthe150percentthresholdtobeconsideredheavilyindebted.However,debtinrealdollars increased from$154billion in2006 to$189billion in2010. Liberia, theoverachiever indebtrelief,hasrecentlycraftedamedium‐termexpenditureframeworktoincreaseitspublicdebtto45percentofGDPover thenext threeyears inpart through the issueofnew treasurybills.11Shouldthisbeaconcerntothedonorswhorecentlypaidoff$75billionofdebt?
Figure4:Africanexternaldebt(2005US$andasapercentofGDP)
AccordingtoReisenandNdoye(2008)theoptimaldebt‐to‐exportratioisnomorethan150
percent.However,mostAfricancountriesinthe1980sand1990sreachedapeakofmorethan700percent for theaverageemerging country,nearly700percent for threshold countries (excludingLiberia), andnearly900percent foroilproducersandother sub‐SaharanAfricancountries.Only
10Thisisinend‐2009netpresentvalueterms.Source:UnitedNations,MillenniumDevelopmentGoalsIndicators11InternationalMonetaryFund(2012)
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
$0
$50
$100
$150
$200
$250
$300
$350
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
% of G
DPU
S $
Billions
constant 2005 $
% GDP
7
recently, has the average debt‐to‐export ratio fallen below the 150 percent ratio level asrecommendedbyReisenandNdoye.
As described earlier, each recipient country under debt relief faced two key dates: a
“DecisionPoint”,underwhich theyweredeemedtobeeligible fordebt relief; anda “CompletionPoint”, when they essentially graduated from the debt relief program. Figures A1 to A4 in theappendix provide information about certain debt flows with highlights of these points for eachcountry. At times, they illustrate the dramatic increase in FDI for particular countries. Thisrelationship appears particularly strong in countries that have experienced consistent, robustgrowth in recent years. In the early 2000s, African countries began to take on new and bettereconomicanddemocraticpolicies and theyhad a lesseningdebtburden.Around this timemanyAfricancountriesemerged fromthe lowgrowthequilibriummanyof themhadbeenstuck in fordecades. In this context,weconsider theHIPCdecisionpointas thestartofa structuralbreak;akindof“proxyplus”,wherebothdebtreliefandpolicychangesmadeitpossibleformanyAfricancountriestobreakthedebtcircleandbegintogrowtheireconomiesandreducepoverty.FurtherevidenceofthisstructuralbreakisshownintheWorldBankpovertynumberswhere,in2008,theabsolutenumberofpeoplelivinginpovertyinAfricafellforthefirsttimeinhistory.12
3.TrendsinAfricanFDIDebtreliefrepresentsoneformofofficialdevelopmentassistance(ODA)intendedtospur
long‐termmarket‐based economic growth.Moss (2011) refers to the concepts of “diggingholes”and “capital flows”, of which the former represents service provision and infrastructuredevelopment,whilethelatterrepresentsassistanceintendedtospuranenablingenvironmentforgrowth.Trends inAfricanFDI indicatewhetherODA in the formofdebt reliefhas appropriatelytriggeredcapitalflowsintoarelativelyassistance‐dependentregion.
Throughoutthelatterhalfofthetwentiethcentury,Africalaggedtherestoftheworldwith
regardtothecreationorreceiptofprivateinvestment,capitalformation,andglobalFDIflows.Forits primary source of external capital, Africa relied on development aid, usually in the form ofconcessionaryloans.Inanempiricalanalysis,Nicholson(2012)identifiessub‐SaharanAfricaasan“AID‐oriented region” in which aid flows continue to dominate private investment flows.13 AsFigure5shows,theAfricanshareofFDIindevelopingcountrieswasmorethan50percentinthe1970’s, reaching a peak of 59 percent in 1972, but slipped below 10 percent by the 1990s;meanwhile,Africa’sshareofdevelopmentaidhasreachedaround40percent.14
12http://iresearch.worldbank.org/PovcalNet/index.htm?113Thisdesignationcontrastswith“FDI‐orientedregions”inwhichinvestmenthasovertakendevelopmentastheprimarysourceofexternalcapital.FDI‐orientedregionsincludeEastAsia,LatinAmerica,andtheformerSovietUnion,wheredevelopmentaidhasbeendrawingdownheavilyinthepastdecadeandprivateinvestmenthasbeenincreasingsubstantially.14RegardingtotalglobalFDI(not just those todevelopingcountries), in1970,Africareceived6percentoftotalglobalFDI,but these fell toaround1percent formostof the1980sand1990s.Source:UNCTADStat.http://unctad.org/en/Pages/Statistics.aspx
8
Figure5:TrendsindevelopmentaidandFDIinAfrica
Source:WorldDevelopmentIndicators
Asiedu (2002) investigates the “Africa effect” in which the explanatory factors for FDI in
developing countriesmay not have similar force in sub‐Saharan Africa. Using data for the years1988to1997,shefindsthatdeterminantsofFDIsuchasreturnoninvestmentandinfrastructurehavepositiveimpactsindevelopingcountriesoutsideofAfricabuthavenostatisticallyvalidimpactinsub‐SaharanAfricancountries.15Inaddition,shefindsthat“opennesstotrade”,asdefinedbytheratiooftotaltradetoGDP,promotesFDIinalldevelopingcountriesbuthasasignificantlysmallermarginal impact in sub‐SaharanAfrica. She concludes that, on thismetric,Africa is different andsuggests that effective policies elsewheremay not be effective here. Asiedu explains the lack ofexplanatory power of return on investment as due to a reputation effect across the continent,includingriskofinvestment, lackofinformation,andtheriskofpolicyreversal.Theopenness‐to‐tradeindicatormaybelesseffectiveinAfricaduetothedebtcyclestory,andinfrastructurecouldbe related to resource‐based FDI in Africa. Collier (2007) suggests that a potential issue forinvestmentinAfricamayhavebeenthe“time‐consistency”problem, inthatpresentgovernmentsareunabletobindthemselves fromconfiscating investments inthefuturewhichthendiminishesinvestment.Thisdynamic,however,mayhavechangedbasedonastructuralpolicyreformofthe2000s.
Anyanwu (2012) finds in an analysis of FDI from 1996 to 1998 that FDI flows to African
countries are positively associatedwithmarket size, openness, rule of law, clusters, and naturalresources. He also found that FDI is negatively correlated with higher financial development inAfrica, and that “higherFDI goeswhere foreignaidgoes.” Heargues that foreignaidaffectsFDI
15NotethatAsiedu(2002)’smeasuresinfrastructurebythenumberoftelephonesper1,000people,whichmayno longer be a reliable indicator given changes in telecommunication technology. Shepoints out thatinfrastructuremightnotbeasrelevantforresource‐basedFDI,whichiscommontoSub‐SaharanAfrica.
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
70.0%
$0
$5,000
$10,000
$15,000
$20,000
$25,000
$30,000
$35,000
$40,000
$45,000
$50,000
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
Share of flows to developing world
Millions of current USD
FDI ODA FDI Share ODA Share
9
througha “positivevanguardeffect”,by loweringperceptionsof investment risksandcultivatingdonor‐specificnorms,whilealsoimprovingsocialandphysicalinfrastructure.
Insub‐SaharanAfrica,incomingFDIincreasedfromUS$6.7billionin2000toUS$25.7billionin
2010.Withthesetrends,foreigninvestmenthasbecomeamuchmoresignificantsourceofcapitalforinvestmentinAfricancountriesandin2009accountedfor18percentoftheirgrossfixedcapitalformation.16Thecatalystforthechangeininvestmentflowscouldbeanumberoffactors,includingdebtrelieforothertypesofdevelopmentaid.Wereassessthesestatisticsunderthehypothesisthatthedebtreliefprogramsrepresentapotentialstructuralbreak in theAfrican investmentclimate.Theprogramsthemselvesmayberepresentativeofabroadergeo‐politicalshiftinAfrica:post‐warandpost‐debt.
4.IdentificationStrategyandEconometricAnalysisOur identification strategy is based on the natural experiment presented by the HIPC debt
reliefprogramofthe2000s,bycreatingcontrolgroupsof“non‐HIPCcountriesinAfrica”and“HIPCcountries outside of Africa.” These two comparison sets of longitudinal panel data allow us toidentifythespecificimpactsofHIPCwhileisolatingpan‐Africangeo‐economicand–politicaltrends.TheidentifyingassumptionissimplythatHIPCdebtreliefinitiativesaffectedonlythosecountriesthatreceiveddebtrelief,withallotherdifferencesamongcountriescapturedbyeithergeography,fixedeffects,orotherrelevantcontrols.By includingbothnon‐HIPCcountries inAfricaandHIPCcountriesoutsideofAfricaweobtainsufficientheterogeneitytocapturetheimpactofthepolicies.
Tothispoint,weengagedinthreedifferenttests,outlinedinTable1.
Table1:DifferentTests
ResearchQuestion EconometricMethod Result
DidsomethinghappeninAfrica?
Clemente,Montañes,andReyes(1998)
Solvesforglobalmaxint‐stattoidentify“optimal”structuralbreakovertime
DiditaffectthedeterminantsofFDI?
ElliotandMüller(2006) Analysisofparameterstabilityofcoefficients
Wasitaresultofdebtrelief?
NaturalExperimentaroundanEventStudy
Controlvs.HIPCshowswhethertheeffectswereduetoHIPCornot
Controlvs.AfricashowswhethertheeffectswereAfrican‐centricornot
Controlvs.“Event”showswhethertheeffectswereduetodebtrelief,ortoa“proxyplus”
16DataaccordingtoWDI.Thesefiguresdippedslightlyfollowingtheglobalrecession.
10
We collected data on debt, gross capital formation, and FDI from the World Development
Indicatorson42sub‐SaharanAfricancountriesfrom1970to2010totestthisproposition.17Table2describesthedatausedintheanalysiswithsummarystatisticspresentedintheappendixusingdeterminantssuggestedbyAnwanyu(2012).Table2:DataDescription
Variable DefinitionFDIflows Foreigndirectinvestment,netinflowsUrbPop UrbanPopulation( percent oftotal)GDPpc GDPpercapita(constant2000$)Trade Trade(percent ofGDP)Credit Domesticcredittoprivatesector( percent ofGDP)ExRate Officialexchangerate(localcurrency perUS$,periodaverage)Telephone Telephonelines(per1,000 people)ODAflows NetODAreceivedGDPgrowth GDPgrowth(annualpercent)Source:WDITables4.1Evidenceofashift:structuralbreaksindebtandFDI
As shown in Figure 4 above, debt in constant dollars tracks very closely with debt as apercentageofGDPuntilaround2004or2005.Inrecentyears,debtinconstantdollarshasslightlyrisenwhiledebtasapercentageofGDPhasfallen.AlthoughAfricandebt levels innominaltermsarebeginning to return topre‐relief totals, could this recent increase represents investment thatleadstogrowth?IftherehasbeenafundamentalchangeinthenatureofpublicandprivateloansinAfrica,itcouldmanifestitselfinmacroeconomicdatathroughashiftintheestimatedrelationshipsbetweendifferentvariables.Forexample,ifdebtpriortothebreakfundedactivitieslessconducivetoeconomicgrowth,suchaspersonalconsumptionorwarfare,theestimatedrelationshipbetweendebt and growth would be different. In statistical terms, there would be a difference in theestimatedcoefficientsatsomemeasurablesignificance:thatis,astructuralbreak.
Clemente, Montañes, and Reyes (1998) develop tests that allow for the data to revealstructuralbreaksinatimeseries,whichinvolvesaglobalsearchforthemaximumabsolutevalueoftheteststatistic;asBaum(2005)describestheprocess,ityields“thestrongestrejectionoftheunitroot null hypothesis.” We employ this test to identify specific shifts in the data to provide anindicationofwhetherastatisticallysignificantshiftoccurredforindividualcountriesofAfricaoverthepastdecade;thatis,whethersomethingfundamentallychangedintherelationshipofdebtandinvestment.
17DatawasnotavailableforEquitorialGuinea,Namibia,Niger,SaoTomeandPrincipe,andSomalia.Duetothetimeperiodunderinvestigation,SouthSudanwasnotincludedintheanalysis.
11
Figure6shows theresults for theFDIseries for theexamplecountryofLiberia.TheclemaotestidentifiestheoptimalstructuralbreakforLiberiain2009,theyearbetweenitsdecisionpointandcompletionpoint.NotealsothesharpmovementintheFDItimeseriesin1989,theyearthathostilities erupted in the country. The Stata command clemao2, incorporating the double‐breakmodel of Baum, Barkoulas, and Caglayan (1999) accounts for the possibility of two breaks andrespectstheimplicationsofabothaglobalandalocalmaximum.Figure6:Clemaoanalysisofstructuralbreak
Tables A‐2 and A‐3 in the Appendix show, the structural breaks for debt and FDI,respectively,forallAfricancountries.Forexposition,Figure7showhistogramsofthedatainthosetables.Thesehistogramsshowdebtwithapeakinthelate1980sandFDIwithapeakinthemid‐2000s.This visual evidence is consistentwith a story that a debt crisis occurred inAfrica in themid‐1980sthatwasresolvedbythemid‐to‐late2000s,afterwhichtimeforeigninvestmentbegantoflowtotheeconomiesofthecontinent.
-4.00e
+08
-2.00e
+0802.00e
+08
4.00
e+08
D.fdi
1970 1980 1990 2000 2010year
D.fdi
-4.5
-4-3
.5-3
-2.5
brea
kpoint t-statis
tic
1970 1980 1990 2000 2010year
Breakpoint t-statistic: min at 2007
in series: fdiClemente-Montañés-Reyes single AO test for unit root
12
Figure7:StructuralbreaksforFDIanddebt
Table 3 shows the statistical relationship between external debt and private investment, as
measuredbygrosscapitalformationandFDI.Anegativerelationshipappearstoexistbetweendebtstockandgrosscapital formation,butthiseffectdisappearswhenaccountingforcountry‐specificfeaturesthroughtheinclusionoffixedeffects.TheflowofFDIasashareofGDP,however,isshowntobenegativelyrelatedtoexternaldebtstockswhenaccountingforcountryandyearfixedeffects.TheseresultssuggestthatloweringdebtcouldincreaseFDIinacountry.
Table3:Debt/GDPimpactonprivateinvestment
GrossCapitalFormation(n=1,473)
ForeignDirectInvestment(n=1,495)
ExternalDebt(noFixedEffects)
‐0.0071** 0.0058***(‐2.11) (3.69)
R2 0.0030 0.0090ExternalDebt
(withFixedEffects)0.0040 ‐0.0121***(0.77) (‐6.18)
R2 0.0361 0.1045T‐statisticsinparentheses.***99percentsignificant**95percentsignificant*90percentsignificant.
4.2ImpactonFDI:changesinparameterstability
Debtmanagement,however,isnotsomuchadeterminantofFDIasitisaproxyforgeneralconditionsofeconomicgovernanceandsowecheckifastructuralbreakhasoccurredamongthedeterminants of FDI in the same timeframe as debt relief. Although these results do not implycausality, they demonstrate that the changing nature of FDI flows to Africa are strongly linkedtemporallywiththedebtreliefinitiatives.
0
2
4
6
8
10
12
14
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
FDI Debt
13
While the determinants of FDI flows have been discussed extensively in the economicliterature18,researchonAfricahastendedtoemphasizeitsdifferences.Asreferencedabove,Asiedu(2002)concludesthatthedeterminantsofFDImaydifferinAfricacomparedtoelsewhere.Asherdataendedwith1997,theperiodpriortoeventsinvestigatedinthispaper,Table4replicatestheanalysis using data from 1970 to 2010. As found by Asiedu, little correlation exists betweenopenness and return on investment on FDI in Africa.Moreover, a ChowTest for the break year2004 suggests thatwith regards to the determinants of FDI, no structural break exists for thesedeterminants.
Table4:ReplicatingAsiedu(2002)
FDI(n=37)
Openness 0.048 ‐0.020
(1.33) (‐0.50)
Telephones 612.3*** 975.3***
(4.17) (5.20)
ROI ‐138.981 ‐693.870
(‐0.29) (‐1.63)
growth ‐2.068
(‐0.32)
Govt ‐0.418***
(‐3.76)
Money ‐0.000
(‐0.09)
Intercept ‐515.917 ‐957.189
R_sq 0.7218 0.8040
Chowtest(2004):
F(3,30)=0.03 Prob>F=0.9920
For individual countries, however, the relationship between FDI and African economies
appearstohavechanged.Anwanyu(2012)asks:“WhyDoesForeignDirectInvestmentGoWhereItGoes?” and provides updates on the statistically relevant determinants of FDI for sub‐Saharancountries.WeusethesepreviousanalysesofFDI inAfrica to investigate the impactofdebtreliefinitiatives on foreign capital flow to the region. Replicating the Anwanyu determinants, weincorporatetheStatatestqLL(quasi‐locallikelihood),basedonElliotandMüller(2006),totestforparameter instability. It tests whether any structural break occurred over the time period inquestion. Table 5 shows the qLL determinants for all countries, clearly suggesting a shift in thedeterminantsofFDIformanyAfricancountriesbetween1970and2010.
18See,amongothers,Caves(1982)andMarkusen(1995).
14
Table5:ParameterStabilityforDeterminantsofFDI,individualAfricancountries
Country t‐stat Country t‐stat Country t‐stat
Angola nodata Gambia,The ‐33.445 Rwanda ‐36.48
Benin ‐33.135 Ghana ‐95.697SaoTomeandPrincipe nodata
Botswana ‐190.536 Guinea ‐669.997 Senegal ‐28.831
BurkinaFaso ‐42.947 Guinea‐Bissau ‐316.18 Seychelles ‐38.58
Burundi ‐124.561 Kenya ‐42.199 SierraLeone ‐35.44
Cameroon ‐49.068 Lesotho ‐44.428 Somalia nodata
CapeVerde ‐51.412 Liberia ‐60 SouthAfrica ‐88.693
CentralAfricanRepublic ‐52.012 Madagascar ‐54.297 SouthSudan nodata
Chad ‐56.459 Malawi ‐56.24 Sudan ‐154.853
Comoros ‐349.867 Mali ‐38.901 Swaziland ‐42.155
Congo,Dem.Rep. ‐34.908 Mauritania ‐649.334 Tanzania ‐124.502
Congo,Rep. ‐45.171 Mauritius ‐29.839 Togo ‐78.739
Coted'Ivoire ‐29.716 Mayotte nodata Uganda ‐75.925
EquatorialGuinea ‐869.616 Mozambique ‐54.583 Zambia ‐31.328
Eritrea ‐52.313 Namibia ‐228.327 Zimbabwe ‐33.382
Ethiopia ‐363.75 Niger ‐38.408
Gabon ‐36.566 Nigeria nodata
ThresholdValues(forstatisticalsignificance)
1percent 5percent 10percent
‐40.24 ‐35.74 ‐33.45
4.3Aretheseresultsaresultofdebtrelief?
As the Clemente, Montañes, and Reyes (1998) test indicated structural breaks in both timeseries with peaks in expected years and the Elliot and Müller (2006) suggested statisticallysignificantdifferenceindeterminantsofFDIforparticularcountries,butnotforAfricaasawhole,thissectionincorporatesanaturalexperimentthatexploitsinformationcapturedinpaneldata.Forthenaturalexperimentaroundaneventstudy,weuse:
Control(non‐HIPCcountries)vs.HIPCshowswhethertheeffectswereduetoHIPCornot: Control(non‐Africancountries)vs.AfricashowswhethertheeffectswerespecifictoAfrica Control(non‐indebtedcountries)vs.“Event”showswhethertheeffectswereduetodebt
relief(a“proxy‐plus”).
Wedeveloptheeconometricapproachforaneventstudybyisolatingbothapolicyevent(suchastheHIPCdecisionpoint)andacontrolgroup(suchasnon‐HIPCcountry).19Thecontrolgroupisimpactedbyallotherinfluencesexceptforthepolicyevent.Thisapproachdirectlyaccountsfortheimplications that the “something” that happened in Africa over the past decade resulted fromdeeper shifts in economic governance; that is, we isolate the “proxy‐plus.” The control groupsaddress whether HIPC occurred due to better economic governance, leading to a shift in the
19Asanexampleofthisapproach,seeBuraimo,Migali,andSimmons(2012).
15
impactsofFDI,orwhetherbettereconomicgovernanceleadtobothHIPCdebtreliefandashiftintheinflowsofforeigndirectinvestment.Figure8diagramstheeventstudyintoquadrants.Figure8:EventStudy
Non‐HIPCCountry(J=0) HIPCCountry(J=1)
Pre‐DecisionPoint(D=0) Event=0 Event=0
Post‐DecisionPoint(D=1) Event=0 Event=1
TheestimatingequationisbasedonAnwanyu’s(2012)determinantsofFDIinAfrica:1 , , , , , ,
, , , , ɛ , Weconductestimationof(1)forthethreenaturalexperimentsusingtheStatacommandxtregartoaccountforautoregressivityinthetimeseries.Table6showstheresults.
16
Table6:ResultsofFDIflows
Baseline Control HIPC Control Event Control Africa
L.FDIflows 0.467*** ‐0.0443 0.397*** 0.403*** ‐17.12 0.671* ‐0.206
(17.59) (‐0.63) (7.42) (14.40) (‐0.11) (2.17) (‐0.66)
urbpop 1.481*** 1.811* ‐0.614 1.586*** 0.386*** 0.242 1.292
(4.73) (1.97) (‐1.08) (4.46) (4.27) (0.11) (0.59)
GDPpc 0.0680* 0.234 ‐0.145 0.0757* 0.0356 ‐0.0714 0.138
(2.14) (0.98) (‐0.63) (2.32) (0.11) (‐0.17) (0.33)
Trade 1.693* 1.443 0.119 1.432 0.143 1.752 ‐0.0886
(2.21) (0.59) (0.08) (1.73) (0.86) (0.68) (‐0.03)
Credit 8.701*** 34.50*** ‐16.02*** 10.35*** 1.810 2.781 7.284
(4.32) (5.37) (‐4.11) (4.82) (1.41) (0.53) (1.27)
ExRate 0.0537 1.749 ‐0.853 0.0522 ‐10.10 ‐1.681 1.733
(1.53) (1.72) (‐1.67) (0.63) (‐1.83) (‐0.57) (0.58)
Telephone ‐24.00** ‐73.87** 34.16 ‐29.17** ‐0.0106 14.29 ‐40.06
(‐3.11) (‐3.04) (1.77) (‐2.98) (‐0.12) (0.35) (‐0.96)
ODAflows 0.00245*** 0.0118*** ‐0.0057*** 0.00307*** 1.448 ‐0.00026 0.00298
(5.09) (7.48) (‐6.04) (5.82) (0.04) (‐0.13) (1.41)
GDPgrowth 2.411 5.961 ‐2.226 2.896 ‐0.00171* 3.141 ‐0.479
(1.01) (0.73) (‐0.50) (1.16) (‐2.07) (0.30) (‐0.04)
_cons ‐774.0*** ‐500.6*** ‐800.2*** ‐5.000 ‐736.6***
(‐6.52) (‐3.54) (‐6.20) (‐0.68) (‐5.21)
N 1303 1303 1303 1303
adj.R‐sq 0.303 0.358 0.304 0.307
t‐statisticsinparentheses
*p<0.05 **p<0.01 ***p<0.001"
ThefirstresulttoacknowledgeisthatnoneofthecoefficientsintheanalysisforAfricaarestatisticallysignificant.Althoughthislackofstatisticalrelevanceisdueinparttothesmallsamplesizeofnon‐Africancountries,itunderscoresastrongerimplication:thevarianceofdeterminantsofFDIarenotaffectedasmuchbythepresenceofcountriesinsub‐SaharanAfricaasmuchasbytheHIPCdesignationandthe“event”oftheDecisionPointfordebtrelief. DebtreliefisastoryaboutAfricaonlybecausemostdebtrelieftookplaceinAfrica.
Thebaseline results are consistentwithAnyanwu (2012)with the statistically significantvariablesUrbPop,Trade, andODAcarrying signs consistent theconclusions that:1)FDI flows tocountries that receivemore foreign aid; 2) largemarket size (representedbyUrbanPopulation)attracts FDI; and 3) export‐oriented economies facilitate foreign direct investment. Also in thebaseline,thecoefficientonTelephoneisnegative,whichiscounterintuitiveandcallsintoquestionthe continued relevance of using kilometers of telephone lines as ameasure of infrastructure indevelopingcountries.
17
The impact of debt relief initiatives is manifest in the variables Credit and ODA. ThecoefficientonCreditispositiveinthebaselinemodel,whichsuggeststhathigherlevelsofdomesticcredit to theprivatesector (asapercentageofGDP)hasapositive impactonFDI flows. For theHIPCanalysis, theCreditcoefficient ispositive for thecontrolgroups inboththeHIPCandEventanalysesbutstatisticallyhalfthesizeforHIPCcountries.20Theseresultssuggestthattheimpactofdomesticmarkets on FDI is a dominant factor in financially secure (non‐HIPC) economies. TheEvent study provides further supporting evidence in that the diminished impact on the Creditcoefficient is statistically insignificant. The conclusion is that by the standards of statisticalrelevance, Credit has a negative influence on FDI for heavily indebted poor countries thatdisappeared following the debt relief initiatives. That is, prior to the initiatives, a heavy dose ofdomesticcreditnegativelyimpactedFDI.Followingdebtrelief,thelevelsofdomesticcreditdidnotaffectFDI.Bythesestandards,recentlyincreasingdebtlevelsinformerlyindebtedcountriesare“agoodkindofdebt.”
ODAfollowsasimilarpattern,althoughwithamuchsmallerimpactonthemagnitudeofthecoefficient. While FDI may follow foreign aid, the effect is diminished in heavily indebted, poorcountries.These results are consistent forboth theHIPCandEvent control groups, although theHIPCcountrieshaveasmalleraggregatecoefficientforODAflows.21ThisresultsuggeststhatwhileFDI may follow foreign aid, the effect is diminished in heavily‐indebted, poor countries. Thediminished effect on ODA does not extend to the Event analysis, suggesting that following theDecisionPoint the relationshipbetween foreign aid and foreigndirect investment is the same inHIPC countries as elsewhere,which supports thewidely‐expressed idea that “aidworks in goodenvironments.”22
2034.50‐16.02=18.48210.0118‐0.0057=0.006122See,amongothers,BurnsideandDollar(2000).
18
5.ConclusionsOurresultssuggestthatthenextstepinAfricandevelopment,followingdebtrelief,should
be investment.We conclude that a fundamental shift in economic governance occurred over thepastdecadecreatingmarket‐basedincentivesforprivateinvestment.Althoughtheclimatecanshiftquickly, through such events as another global recession or a resurgence of civil conflict, theevidenceimpliesthatthedebtreliefinitiativesweresuccessfulandAfricaneconomiesarecurrentlyintheprocessofsustained,market‐basedgrowth.
WhilerecognizingthatHIPCandMRDImaybeaproxyfordeepereconomicgovernance,thedebt relief itself appears to have played a substantive role regarding foreign direct investment.Sincemuchofthe$75billionindebtreliefgrantedtoHIPCcountrieswasintheformofoverseasdevelopmentassistance,theseresultsposeaninterestingquestionabouttherelationshipbetweenforeignaidandFDI,whetheraidisa“signal”forconfidenceinthemarketsofdevelopingcountriesorwhetheraidoffersadirectchanneltoimprovemarketconditions.Thispaperhasdemonstratedadirectimpact,totheextentthatdebtreliefitselfwasmanifestedinassistanceandhaspositivelyaffectedtheflowsofforeigndirectinvestmenttoformerlyindebtedcountries.
AnotherconclusiontobedrawnisthatthereisnostatisticaldifferenceforHIPCcountriesinAfricacomparedtoHIPCcountriesoutsideofAfrica. TotheextentthatAfricancountrieshavehistoricallyoffereddifferentincentivesforFDIandothermarket‐basedactivities,thesedifferencesappeartohavearisenbecausecountriesinAfricahaveskewedtowardsbeingheavily‐indebtedandpoor. One story is that when it comes to debt and FDI, Guinea is more like Guyana than likeBotswana.DebtreliefisastoryaboutAfricaonlybecausemostdebtrelieftookplaceinAfrica.Animplicationisthatfollowingthedebtreliefinitiatives,wecananticipatethatFDIwillflowtoAfricaby similarmechanismsas inother regions. If the continent is “open forbusiness”, the ability forcountriestoagainengageinpublicfinancingtoimprovemarketconditionsshouldbeanticipatedasapositiveturnofevents.
One future direction will analyze FDI by sector, specifically disaggregating the flows toaccountforextractiveindustries.AsforeigninvestmentinAfricahastraditionallybeenfocusedonnaturalresources,furtherindicationsthatthemarketpotentialoftheregionhasshiftedinthepastdecadewillbeindicatedbythemixofinvestmentopportunities. Thislineofresearchwouldalsobenefitfromastrongermeasureofinfrastructurethataccountsforroads,electricity,andinternetaccess.
19
WorksCited
Addison,Tony.2006.Debtrelief:Thedevelopmentandpovertyimpact.SwedishEconomicPolicyReview13:205‐230.
Anyanwu,JohnC.2012.WhyDoesForeignDirectInvestmentGoWhereItGoes?:NewEvidence
fromAfricanCountries.AnnalsofEconomicsandFinance13(2):433‐470.Asiedu,Elizabeth.2002.OntheDeterminantsofForeignDirectInvestmenttoDeveloping
Countries:IsAfricaDifferent?WorldDevelopment30(1):107‐119.Baum,ChristopherF.2005.Stata:Thelanguageofchoicefortime‐seriesanalysis?StataJournal,
5(1):46‐63.Baum,ChristopherF.,JohnBarkoulas,andMustafaCaglayan.1999.Longmemoryorstructural
breaks:caneitherexplainnonstationaryrealexchangeratesunderthecurrentfloat?JournalofInternationalFinancialMarkets,9(4):359‐376.
Buraimo,Babatunde,GiuseppeMigali,andRobSimmons.2012.Corruptiondoesnotpay:An
analysisofconsumerresponsetoItaly’sCalciopoliscandal.LancasterUniversityManagementSchool.Paperpresentedat2012WesternEconomicAssociationInternationalinSanFrancisco,CA.
Burnside,CraigandDavidDollar.2000.Aid,Policies,andGrowth.AmericanEconomicReview.
90(4):847‐868.Clements,Benedict,RinaBhattacharyua,andToanQuocNguyen.2005.CanDebtReliefBoost
GrowthinPoorCountries?InternationalMonetaryFundEconomicIssuesNo.34.Dorsey,Thomas.2008.ACapitalStory.InternationalMonetaryFundFinanceandDevelopment
Volume45,Number2.Easterly,William.2002.HowDidHeavilyIndebtedPoorCountriesBecomeHeavilyIndebted?
ReviewingTwoDecadesofDebtRelief.WorldDevelopment30(10):1677‐1696.InternationalMonetaryFund.May2012.Liberia:EighthReviewUndertheThree‐Year
ArrangementUndertheExtendedCreditFacility.Levine,Ross.1992.ASensitivityAnalysisofCross‐CountryGrowthRegressions.AmericanEconomic
Review82(4):942‐63.McKinseyGlobalInstitute.2010.Lionsonthemove:TheprogressandpotentialofAfrican
economies.McKinsey&Company.Moss,ToddJ.2011.AfricanDevelopment:MakingSenseoftheIssuesandActors,SecondEdition.
Boulder,CO:LynneRiennerPublishers.Nicholson,MichaelW.2012.Regionalvariationsintheimpactofdevelopmentaidonforeigndirect
investment.SSRNWorkingPaperNo.2140872http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2140872.
20
Radelet,Steve.2007.RevivingEconomicGrowthinLiberia.CenterforGlobalDevelopmentWorking
PaperNumber133.Radelet,Steve.2010.EmergingAfrica:How17CountriesareLeadingtheWay.Washington,DC:
CenterforGlobalDevelopment.Reisen,HelmutandSokhnaNdoye.2008.PrudentversusImprudentLendingtoAfrica:FromDebt
RelieftoEmergingLeaders.OECDDevelopmentCentreWorkingPaperNo.268.Rose,AndrewK.2004.DoWeReallyKnowThatTheWTOIncreasesTrade?AmericanEconomic
Review94(1):98‐114.
21
AppendixTableA‐1:Debt‐to‐ExportRatiosCountry 1970 1980 1985 1990 1995 2000 2005 2010Angola 215 % 115% 49% 38%Benin 86% 214 % 365% 425% 345% 405% 266% 130 %Botswana 73% 24 % 47% 26% 29% 15% 9% 35%BurkinaFaso 84% 191 % 333% 245% 378% 600% 368%Burundi 58% 205 % 359% 1018 % 899% 2014% 1457%Cameroon 48% 137 % 116% 292% 532% 441% 214% 46%CapeVerde 311 % 228% 219% 130% 133 %CentralAfricanRepublic 40% 97 % 193% 318% 414% 454% 599%Chad 46% 162 % 176% 219% 266% 465% 49% 52%Comoros 409 % 690% 527% 461% 669% 509%Congo(Brazzaville) 125% 147 % 243% 326% 430% 185% 120% 37%Congo(Kinshasa) 45% 201 % 312% 372% 824% 1213% 433% 169 %Coted'Ivoire 72% 210 % 296% 504% 411% 288% 143% 123 %Djibouti 64 % 140% 134% 141%Eritrea 28 % 313% 1070%Ethiopia 1287 % 1401 % 558% 334% 211 %Gabon 65% 55 % 58% 145% 148% 112% 69% 29%Gambia 26% 133 % 248% 194% 228% 239% 358% 199 %Ghana 121% 372 % 467% 376% 347% 252% 174% 88%Guinea 299 % 416% 417% 292% 185 %Guinea‐Bissau 0% 978 % 2275 % 2865 % 3021 % 1384%Kenya 100% 158 % 269% 320% 248% 224% 120% 95%Lesotho 97% 79 % 430% 404% 355% 263% 99% 76%Liberia 68% 112 % 268% 2319% 1940%Madagascar 228% 230% 720% 721% 564% 394% 246%Malawi 192% 270 % 373% 348% 527% 606% 480% 60%Mali 555% 277 % 661% 594% 568% 456% 235%Mauritania 31% 322 % 355% 454% 292% 613% 344% 110 %Mauritius 73 % 100% 54% 60% 34% 21% 21%Mozambique 2214 % 2285 % 2114 % 1029% 199% 170 %Niger 45% 140 % 410% 472% 497% 534% 394%Nigeria 79% 47 % 408% 270% 274% 126% 42% 11%Rwanda 19% 113 % 196% 488% 1545 % 841% 512%Senegal 61% 176 % 310% 258% 260% 276% 164% 115 %Seychelles 84 % 81% 80% 71% 63% 94%SierraLeone 46% 187 % 557% 807% 752% 1035% 599% 239 %Somalia 212% 329 % 2955 % 2641 %SouthAfrica 74 % 67% 46% 45%Sudan 112% 642 % 1251 % 2958 % 2561 % 845% 350% 178 %Swaziland 47% 51 % 114% 39% 24% 23% 20% 30%Tanzania 1199 % 582% 525% 284% 145 %Togo 32% 193 % 253% 235% 347% 350% 197%Uganda 52% 285 % 256% 836% 532% 530% 336% 73%Zambia 85% 202 % 547% 585% 555% 652% 216% 52%Zimbabwe 50 % 193% 163% 183% 149% 213% 139 %Source:World Bank, World Development Indicators and Author Calculations. Countries with no data areremovedfromthetable.
22
Appendix:DataSummaryStatisticsVariable Obs. Mean Std.Dev. Min Max
FDI 1742 3.07 8.17 (82.89) 145.20UrbPop 2080 0.31 0.15 0.03 0.86GDPpc 1885 891.18 1,322.63 54.51 9,279.11Trade 1855 72.83 40.31 ‐ 280.36Credit 1802 18.84 17.84 0.68 161.98ExRate 2125 339.05 1,122.41 0.00 18,498.60Telephone 1776 1.96 4.07 0.01 30.30ODA 1820 12.17 13.12 (0.25) 181.01GDPgrowth 1890 3.72 7.28 (51.03) 106.28
23
FigureA1:HIPCCountriesinAfrica
$0
$500
$1,000
$1,500
$2,000
$2,500
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Burkina Faso
$0
$500
$1,000
$1,500
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Burundi
$0
$500
$1,000
$1,500
$2,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Benin
$0
$500
$1,000
$1,500
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Central African Republic
$0
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Cameroon
$0
$100
$200
$300
$400
$500
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Comoros
$0
$5,000
$10,000
$15,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Congo (Brazzaville)
$0
$1,000
$2,000
$3,000
$4,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Chad
$0
$5,000
$10,000
$15,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Congo (Kinshasa)
$0
$5,000
$10,000
$15,000
$20,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Cote d'Ivoire
$0
$2,000
$4,000
$6,000
$8,000
$10,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Ethiopia
$0
$200
$400
$600
$800
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
Gambia
$0
$2,000
$4,000
$6,000
$8,000
$10,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Ghana
$0
$1,000
$2,000
$3,000
$4,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Guinea
$0
$200
$400
$600
$800
$1,000
$1,200
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Guinea‐Bissau
$0
$1,000
$2,000
$3,000
$4,000
$5,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Liberia
$0
$1,000
$2,000
$3,000
$4,000
$5,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Madagascar
Legend
External Debt
FDI
Completion Point
Decision Point
24
$0
$2,000
$4,000
$6,000
$8,000
$10,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Zambia
$0
$2,000
$4,000
$6,000
$8,000
$10,0001980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Uganda
$0
$2,000
$4,000
$6,000
$8,000
$10,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Tanzania
$0
$500
$1,000
$1,500
$2,000
$2,500
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Togo
$0
$500
$1,000
$1,500
$2,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Sierra Leone
$0
$100
$200
$300
$400
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Sao Tome and Principe
$0
$1,000
$2,000
$3,000
$4,000
$5,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Senegal
$0
$500
$1,000
$1,500
$2,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Rwanda
$0
$2,000
$4,000
$6,000
$8,000
$10,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Mozambique
$0
$500
$1,000
$1,500
$2,000
$2,5001980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Niger
$0
$500
$1,000
$1,500
$2,000
$2,500
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Mauritius
$0
$1,000
$2,000
$3,000
$4,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Mali
$0
$500
$1,000
$1,500
$2,000
$2,500
$3,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Mauritania
$0
$1,000
$2,000
$3,000
$4,000
1980
1983
1986
1989
1992
1995
1998
2001
2004
2007
2010
Millions
Malawi
Legend
External Debt
FDI
Completion Point
Decision Point
25
TableA‐2:StructuralBreaks,debt
Country Onebreak Twobreaks Country Onebreak TwobreaksAGO 2009 2000 2004 MLI 1986 1987 2007BDI 1988 1988 2005 MOZ 2002 1991 2002BEN 1983 1983 2007 MRT 1983 1979 1988BFA 1988 1984 1995 MUS 1988 1988 2002BWA 2006 1988 2002 MWI 1988 1991 2007CAF 1988 1982 1988 NER 1982 1982 2007CIV 1982 1982 1993 NGA 1984 1984 2007CMR 1996 1991 2007 RWA 1988 1989 2007COG 1982 1982 1991 SDN 1991 1983 1991COM 1988 1989 2005 SEN 1982 1983 2003CPV 2004 2000 2008 SLE 1982 1980 1988ERI 2005 2000 2005 STP 1990 1992 2004ETH 1983 1984 2000 SWZ 1998 1979 1998GAB 1988 1988 2004 SYC 2008 2002 2007GHA 1991 1983 1992 TCD 2005 1992 2004GIN 1989 1982 1991 TGO 1980 1980 1993GMB 1988 1983 1995 TZA 1978 1979 2003GNB 1989 1984 1991 UGA 1988 1988 2007KEN 1988 1981 1988 ZAF 2008 2004 2007LBR 1988 1988 2004 ZAR 1987 1979 1988LSO 1989 1984 1991 ZMB 1982 1984 2007MDG 1982 1983 2007 ZWE 1986 1983 1991TableA‐3:StructuralBreaks,FDI
Country Onebreak Twobreaks Country Onebreak TwobreaksAGO gapsindata gapsindata MLI 2006 1992 2000BDI 1998 1998 2007 MOZ gapsindataBEN 2004 1989 2004 MRT 2003 2000 2003BFA 2004 1991 2004 MUS 2004 1998 2006BWA 1999 1999 2003 MWI 2000 2000 2005CAF 2004 1979 2005 NER 2007 1978 2005CIV 1990 1990 1994 NGA 2004 1986 2004CMR 1999 1999 2005 RWA 2004 1991 2004COG 2008 2004 2007 SDN gapsindataCOM gapsindata SEN 2003 1994 2006CPV 2004 1996 2005 SLE 2001 1984 2003ERI 2000 2000 2003 STP gapsindataETH gapsindata SWZ 2003 1986 2003GAB 1999 1992 1999 SYC 2002 1994 2002GHA 2007 1995 2007 TCD 2000 2000 2004GIN 2004 1998 2004 TGO 1998 1980 1998GMB gapsindata TZA 1996 1996 2002GNB gapsindata UGA gapsindataKEN 2004 1993 2004 ZAF 1998 1994 2005LBR 2007 1987 2007 ZAR gapsindataLSO 1992 1994 2001 ZMB 2004 1986 2004MDG 2008 2004 2007 ZWE 1996 1994 1996
Related Documents
