Joint Discussion Paper

Series in Economics

by the Universities of

Aachen ∙ Gießen ∙ Göttingen Kassel ∙ Marburg ∙ Siegen

ISSN 1867-3678

No. 24-2014

Matthias Neuenkirch and Florian Neumeier

The Impact of UN and US Economic Sanctions on GDP Growth

This paper can be downloaded from http://www.uni-marburg.de/fb02/makro/forschung/magkspapers/index_html%28magks%29

Coordination: Bernd Hayo • Philipps-University Marburg

Faculty of Business Administration and Economics • Universitätsstraße 24, D-35032 Marburg Tel: +49-6421-2823091, Fax: +49-6421-2823088, e-mail: hayo@wiwi.uni-marburg.de

mailto:hayo@wiwi.uni-marburg.de

TheImpactofUNandUSEconomicSanctionsonGDPGrowth

MatthiasNeuenkirchaandFlorianNeumeierb

aUniversityofTrier

bPhilipps‐UniversityMarburg

Firstdraft:28March2014Thisversion:20May2014

Correspondingauthor:MatthiasNeuenkirchDepartmentofEconomicsUniversityofTrierD‐54286TrierGermanyTel.:+49–651–2012629Fax:+49–651–2013934Email:neuenkirch@uni‐trier.de*Thanks toMartinBresslein andGeorgin Isaev for their support indata collectionaswellastoMartinBresslein,JergGutmann,MatthiasUhl,andseminarparticipantsattheIAAEU Trier and Philipps‐University Marburg for their helpful comments on earlierversionsofthepaper.Theusualdisclaimerapplies.

2

TheImpactofUNandUSEconomicSanctionsonGDPGrowth

Abstract

In thispaper,weempiricallyassesshoweconomicsanctions imposedby theUNandtheUSaffectthetargetstates’GDPgrowth.Oursampleincludes68countriesandcoversthe period 1976–2012. We find, first, that sanctions imposed by the UN have astatisticallyandeconomicallysignificantinfluenceoneconomicgrowth.Onaverage,theimpositionofUNsanctionsdecreasesthetargetstate’srealpercapitaGDPgrowthrateby2.3–3.5percentagepoints (pp).Theseadverseeffects last for aperiodof10years.ComprehensiveUNeconomicsanctions,thatis,embargoesaffectingnearlyalleconomicactivity,triggerareductioninGDPgrowthbymorethan5pp.Second,theeffectofUSsanctions ismuch smaller and less distinct. The imposition ofUS sanctions decreasesGDPgrowthinthetargetstateoveraperiodof7yearsand,onaverage,by0.5–0.9pp.Keywords:Economicgrowth,economicsanctions,UnitedNations,UnitedStates.JEL:F43,F51,F52,F53.

3

1. Introduction

Economic sanctions have become one of the most important tools of statecraft ininternational politics (Cortright andLopez, 2000).Designed as ameans of compellinggovernments to comply with the imposing state’s interests, these measures aim atchangingthetargetnation’spoliciesbyinflictingeconomicdamage.Theyareviewedasanonviolent,morehumanealternativetomilitaryintervention.However,theimpositionofeconomicsanctionsisoftenmetwithharshcriticism,whichisbasedintheunpleasantrealitythateventhoughthesemeasuresaredirectedagainstgovernments,moreoftenthannot,itisthetargetstate’spublicthatbearsthecosts.Thisresultcanbeparticularlyunfair when the regime against which sanctions are directed lacks democraticlegitimation.Thereisahugeandvibrantliteratureontheadverseeffectsofeconomicsanctionsonthe target states’ humanitarian situation. Sanctions are argued to have devastatingconsequencesforthecivilianpopulationastheycannegativelyaffecttheavailabilityoffood and clean water (Cortright and Lopez, 2000; Weiss et al., 1997) and access tomedicineandhealth‐careservices(e.g.,Garfield,2002;GibbonsandGarfield,1999),aswell as have a detrimental impact on life expectancy and infant mortality (e.g., AliMohamed and Shah, 2000; Daponte and Garfield, 2000). Most of this research isqualitative, however, and based on single‐country case studies. Quantitativeassessmentsof sanctioneffects typically focuson their impactonvariousmeasuresofthehumanrightssituation(e.g.,Peksen,2009;Wood,2008),politicalstabilitywithinthetargetstate(Allen,2008;Marinov,2005),levelofdemocracy(PeksenandDrury,2010),andtheirsuccessintermsofmeetingthedesiredobjectives(e.g.,Hufbaueretal.,2009;Drury, 1998; Dashti‐Gibson et al., 1997).1 The findings are dispiriting. For example,Peksen (2009) reports that economic sanctions worsen the targeted government’srespectforhumanrights;PeksenandDrury(2010)findthateconomicsanctionshaveadetrimental impact on the level of democracy. Moreover, economic sanctions fail toachievetheiraimsin65–95%ofthecasesinwhichtheyareimposed(e.g.,Hufbaueretal.,2009;Pape,1997,1998).Empirical research on the economic consequences of economic sanctions is scarce.Evenett(2002)estimatestheimpactofeightindustrializedcountries’sanctionsagainst1 There are also theoretical public choice and game‐theoretical analyses on conditions under which

economic sanctions may trigger policy changes. Examples are Kaempfer et al. (2004), Kaempfer andLowenberg(1988,1999),andEatonandEngers(1992).

4

the South African Apartheid regime on these countries’ bilateral trade relations withSouthAfricabetween1978and1999.His findings suggest that theUSAnti‐ApartheidActhadthestrongestinfluenceonSouthAfricanexports.Hufbaueretal.(2009)relyonalargesampleofbi‐andmultilateraleconomicsanctionepisodesandestimategravitymodels. They find that the imposition of economic sanctions significantly reduces thevolumeofbilateraltradebetweentheimposingandthetargetstate.Thispaperisthefirsteconometricassessmentoftheimpacteconomicsanctionshaveon the target’s overall economic development.2 More precisely, we analyze the effecteconomicsanctionshaveonthetargetcountries’GDPgrowthrate,therebyfocusingon(i) multilateral sanctions imposed by the United Nations as well as (ii) unilateralsanctionsimposedbytheUnitedStates.TheUNSecurityCouncil(UNSC)cancallonitsmemberstatestopartiallyorcompletelyinterrupteconomicrelationswithastatethatthreatensorbreachesinternationalpeaceandsecurity.Firstemployedin1965againstRhodesia,theuseofthismeasurehasbecomeincreasinglypopularduringthepasttwodecades(seealsoFigure1ainSection3.2).AllUNmemberstatesareobligedtoadoptthesanctionmeasuresdeterminedbytheUNSC,whichiswhytheseareexpectedtobeparticularlyeffective.WithregardtotheUS,noothercountryintheworldhasimposedeconomic sanctions more often (Hufbauer, 1998; Hufbauer et al., 2009). Althoughunilateral,theimportanceoftheUnitedStatestotheglobaleconomymaymakethemaninfluentialpolicyinstrument.We compiled a unique dataset comprised of UN and US sanction episodes between1976 and 2012. Our results suggest, first, that sanctions imposed by the UN have asignificant influence on economic growth.On average, the imposition ofUN sanctionsdecreasesthetargetstate’srealpercapitaGDPgrowthrateby2.3–3.5percentagepoints(pp). An investigation of the dynamics of the sanction effects reveals that thedetrimental influence decreases over time and becomes insignificant after 10 years.Differentiatingbetweencategoriesofeconomicsanctions,we find that comprehensiveUNeconomicsanctions—thatis,embargoesonnearlyalleconomicactivitybetweenUN2Hufbaueretal.(2009:211ff)provideroughapproximationsoftheeffectofeconomicsanctionsonthe

targetcountries’grossnationalproduct.However,theauthorsthemselvesadmitthattheirassessmentisrather rudimentary. They simply consider the initial reduction in net exports and foreign grantsassociatedwiththeimpositionofeconomicsanctions,weighthisfigurewitha“sanctionmultiplier,”whichis based on the authors’ subjective judgment of the substitution elasticities of domestic demand andinternationalsupplyoftheembargoedgoods,andputthismeasureinrelationtothetargetstate’sgrossnationalproduct.However, economic sanctionsmayaffect the target country’sGNP in severalways, asoutlinedinmoredetailinSection2ofthispaper.

5

memberstatesandthesanctionedcountry—exertthestrongestinfluence;theytriggerareductioninrealGDPgrowthofmorethan5pp.Ourfindingsarerobusttomodificationsof the empirical specification that control for potential changes in a country’sinstitutional,political,andsocialenvironment.Moreover,whencomparingtheeffectofUNeconomicsanctionswhichwereactuallyimposedtothosewhichwereblockedbyavetointheUNSCwefindthatonlytheformeronesexertanadverseimpactoneconomicgrowth,indicatingthatourresultsarenotdrivenbyomittedfactorsthatcoincidewithsanctionperiods.Second,theadverseeffectofUSsanctionsonrealGDPgrowthismuchsmaller and less lasting than that of UN sanctions. The imposition of US sanctionsdecreasesGDPgrowth in the targetstateoveraperiodof7yearsand,onaverage,by0.5–0.9pp.The remainder of this paper is organized as follows. Section 2 provides sometheoreticalargumentsforwhysanctionsmayhaveadversegrowtheffectsinthetargetcountries and outlines the research hypotheses. Section 3 introduces the empiricalmethodology and the dataset. Section 4 presents the results. Section 5 explores therobustness of our findings with respect to changes in the control sample. Section 6examinestheimpactofsanctionsonGDPgrowthovertime.Section7concludes.2. TheoreticalConsiderationsandHypotheses

Economic sanctions are intended to be coercive measures that fall between merediplomatic pressure and the extreme action of military intervention. According toformer UN Secretary‐General Kofi Annan, sanctions “represent more than just verbalcondemnation and less than the use of armed force.”3 Or, as the formerUS PresidentWoodrowWilson put it: “A nation boycotted is a nation that is in sight of surrender.Applythiseconomic,peaceful,silent,deadlyremedyandtherewillbenoneedforforce”(quotedinHeine‐Ellison,2001:83).Theoretically,economicsanctionsarepowerfuldueto their potential to inflict economic damage. Thus, one should expect UN and USeconomic sanction episodes to have a detrimental impact on the target nation’seconomicdevelopment.Yet, there ishardlyanyempiricalassessmentof theeconomiccostsincurredbysanctions.

3UNPressReleaseSG/SM/7360.

http://www.un.org/News/Press/docs/2000/20000417.sgsm7360.doc.html(accessedinMarch2014).

6

Thereareseveralchannelsthroughwhichsanctionsmayadverselyaffecttheeconomicperformanceof the targetstate.Themostobviousof these includeaslump inexportsand imports, the related loss of bargaining power on international markets, and thecontraction of international capital flows, that is, withdrawal of foreign directinvestment, foreign aid, and financial grants (Hufbauer et al., 2009; Evenett, 2002).However,suchadverseeffectsmayoccurevenwhentradeembargoesorsuspensionsofinternational aid and capital flows are not explicitly imposed. Economic sanctions areoftenusedasasymbolicinstrumenttostigmatizepoliticalregimes(Whang,2011).Theassociated loss of reputation may very well isolate the target state within theinternationalcommunityanddeterdonorsfromfurtherprovidingaidandinvestments.Economic sanctions aim at triggering political reforms or even overthrowing thetarget’s political regime. Moreover, economic agents may view sanctions as a sort ofearly‐warning signal that political or societal conflicts in the target state have thepotentialtoescalate.Sanctionsthusrepresentorindicateaseriousthreattothetargetstate’spoliticalstabilityandcan invokeagreatdealofuncertaintyabout the futureofthepoliticalandlegalsystem.Thisoughttohaveaharmfulimpactonthetargetstate’strade and financial relations aswell as on its domestic and foreigndirect investment.Indeed,empiricalevidencesuggeststhatsanctionepisodesareassociatedwithpoliticalturmoilandtransition(PeksenandDrury,2010;Allen,2008;Marinov,2005).Politicalinstability, in turn, is found to have detrimental effects on investment and savings aswellasoneconomicgrowth(Alesinaetal.,1996;AlesinaandPerotti,1996;Aizenmanand Marion, 1993). In a similar vein, sanctions may affect the target’s access tointernational credit markets as investors might be concerned about the sanctionedstate’ssolvencyorthepaymentpracticesofasuccessorregime.Finally, the imposition of economic sanctions often results in an expansion of theshadow economy as economic agents try to evade sanction measures, involving amarginalization of licit commerce as well as public acceptance of illegal economicactivity(e.g.,Andreas,2005;Heine‐Ellison,2001;CrawfordandKlotz,1999).Moreover,governmentsagainstwhich sanctionsaredirectedoften fail to foster compliancewithlaws as economic sanctions undermine their authority and legitimacy.What is more,targetgovernmentsmayevenpromote illegaleconomicactivities inorder togeneratefunds, secure supplies, and strengthen their power. Also, the decline in governmentauthority as well as the rise of political instability frequently involve an increase in

7

corruption.Asaconsequence,transactioncostsincreaseandmoreresourcestendtobeusedunproductively.Thestrengthoftheeffecteconomicsanctionshaveonthetargetstate’seconomymaybe related to various factors. For example, it could be that the impact of economicsanctionsdependsontheirseverity.PrevioussanctionsemployedbytheUNandtheUSrangefromfreezingprivateandpublicfundsandassetstobanninggrantsandcreditstoimposingembargoesoncertainoralleconomicactivities(foranoverview,seeTable1inSection3.2).MultilateralUNsanctionsoughttohaveastrongeradverseeffectonthetargetcountry’sGDPgrowth thanunilateralUSsanctionssimplybecauseof the largernumberofcountriesinvolvedintheimpositionoftheformer.Accordingly,weformulatethreehypothesesthatweputtoanempiricaltestinthispaper.H1:UNandUSeconomic sanctionshave anegative effect on the target country’s realGDPpercapitagrowth.H2:Thenegativeeffectincreaseswithincreasingseverityofthesanctions.H3:ThenegativeeffectisstrongerforUNsanctionsthanforUSsanctions.3. EmpiricalMethodologyandData

3.1EmpiricalMethodology

To assess the impact of UN and US economic sanctions on the sanctioned state’seconomicperformance,weestimatedifferentversionsofthefollowingmodel:

(Eq.1) , ′ , , , The dependent variable , represents the growth rate of country i’s real GDP percapitain2005USdollarscomparedtothepreviousyear. isacountry‐specificeffectthataccountsforindividualheterogeneityduetounobservedtime‐invariantfactors. isatime‐fixedeffectand , anerrorterm.Oursampleincludesall68countriesagainstwhich UN and US economic sanctions were imposed (see Table A1 in the Appendix)duringoursampleperiodof1976to2012.WefirsttestH1andevaluatetheeffectofUNandUSeconomicsanctionsbyincludingdummy variables that take the value 1 during years in which UN or US sanctions,respectively,wereimposed.Inasecondstep,wetestH2anddiscriminatebetweenthreecategoriesofsanctions(seeSection3.2)thatdifferwithrespecttotheirseverity.Tothisend,weemployseparatedummyvariablesforeachsanctioncategory.

8

To date, the UN and the US have imposed economic sanctions for primarily threereasons: (i) to coerce states (ormilitant groups within states) to terminate acts thatthreatenorinfringethesovereigntyofanotherstate,i.e.,byresortingtoviolenceagainstanother state or destabilizing the incumbent government; (ii) to foster democraticchange in a country, protect democracy, or destabilize an autocratic regime; (iii) toprotectthecitizensofastatefrompoliticalrepressionandenforcehumanrights.4Allthreereasonsforimposingeconomicsanctions—engagementininterstateconflict,autocratic tendencies,andpolitical repression—might in themselvesaffectacountry’seconomic development. Todisentangle their effects onGDP growth from the effect ofeconomicsanctionsandthuscircumventanomittedvariablebias,itiscrucialtoincludeappropriatecontrolvariablesinourempiricalspecification.The vector , includes, first, the Political Terror Scale indicator, which measuresphysicalintegrityrightsviolationsonafive‐pointscale(1:lowestdegreeofviolation;5:highest degree of violation). Second, we control for the degree of democracy orautocracy in a country using a policy variable that is scaled from +10 (stronglydemocratic) to –10 (strongly autocratic). Third, we take into account (i) interstatearmedconflicts,(ii)internalarmedconflictswithoutinterventionfromotherstates,and(iii) internationalizedinternalarmedconflictswith interventionfromotherstates.Forallthreetypesofconflictweincludeseparatedummyvariablesforminorconflictsandwars,respectively.Finally,weconsidercontrolvariablesthatarestandardineconomicgrowthequations:thelogofrealpercapitaGDPin2005USdollars,5tradeopenness(importsplusexportsdividedbyGDP),andthelogofpopulation.Weemploythefirstlagofthesevariablestocircumventproblemsofreversecausality.AlistofthecontrolvariablesalongwiththeirdefinitionsandsourcescanbefoundinTableA2intheAppendix.3.2DataonUNandUSSanctions

We compiled a unique dataset comprised of all UN and US sanction episodes thatoccurredbetween1976and2012.UNsanctionswerecollectedfromtheUNwebsiteandcross‐checkedwithWood’s(2008)dataset,which,unfortunately,ends in2001.ForUS4InformationontheobjectivesofUNandUSeconomicsanctionsisobtainedfromHufbaueretal.(2009)

aswellasfromthewebsitesoftheUNandtheUSCongress.5Notethatthisvariablealsoservesasproxyforacountry’scapitalstocksincereliabledataforthelatter

aredifficulttocollectforthecountriesandperiodunderinvestigation.

9

sanctions, we relied on Hufbauer et al.’s (2009) dataset and augmented it withinformation from the US Congress websites. Each sanction was categorized as either“mild,” “moderate,” or “severe,” based on the definitions found inWood (2008) (seeTable1).Table1:DefinitionofsanctioncategoriesLevel UNsanctions USsanctions1:Mild Restrictions on arms and other

military hardware; typically includetravel restrictions on a nation’sleadership or other diplomaticsanctionsaswell

Retractionsof foreignaid,bansongrants, loans, or credits, orrestrictions on the sale of specificproducts or technologies; notincluding primary commoditiesembargoes

2:Moderate Moderate sanctions such as fuelembargoes, restrictions on trade inprimary commodities, or thefreezing of public and/or privateassets

Importorexportrestrictions,banson US investment, and othermoderate restrictions on trade,finance,andinvestmentbetweentheUSandtargetnation

3:Severe Comprehensive economicsanctions such as embargoes on allor most economic activity betweenUNmemberstatesandthetarget

Comprehensive economicsanctions such as embargoes on allor most economic activity betweentheUSandthetargetnation

Source:Wood(2008:500).Figures 1a and1b illustrate the frequency of sanctions and their severity over time.The overall number of country/year observations inwhichUN sanctions are in place(200;9.3%oftheobservations)ismuchlowerthanthatforUSsanctions(618;28.6%).Similarly,UNsanctionshavebeenimposedagainstonly23countries,whereasatotalof64countrieshaveatleastonenon‐zeroobservationforUSsanctions.Inaddition,theUSsanctionsareonaverageharsherthanthoseoftheUNas21.8%ofUSsanctionsfallintocategory3(comparedto12%fortheUN).Thesefindingsarenotsurprising,ofcourse,sinceUNsanctionshavetobeenactedbytheUNSC,whichconsistsoffivevetopowers,whereasUSsanctionsonlyhave topass theUS legislative.Also interesting is thehugeincreaseinthefrequencyofUNsanctionsaftertheendoftheColdWar.Thefrequencyofsanctionsishighestduringthe1990sduetotheFirstGulfWar,theYugoslavWars,andseveralcivilwarsinAfrica.

10

Figure1a:UNsanctionsandtheirseverityovertime

Figure1b:USsanctionsandtheirseverityovertime

4. EmpiricalResults

4.1BinarySanctionVariable

First, we put H1 to the test. The results are shown in Table 2. We estimate threedifferent specificationsofEquation (1): one includingadummy forUN sanctionsonly(Column(1)ofTable2);onewithadummyforUSsanctionsonly(Column(2));andonewithseparatedummiesforbothUNandUSsanctions(Column(3)).6 6Notethatwerelyonrestrictedsamplesinthecaseof(1)and(2)sincenotallsamplecountrieswere

subjecttoUNsanctionsortoUSsanctions.

0

3

6

9

12

151976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

mild moderate severe

05101520253035

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

mild moderate severe

11

Table2:TheimpactofsanctionsonGDPgrowth:binarysanctionvariable (1) (2) (3)log(realGDP/capita)t‐1 –0.19 *** –0.07 *** –0.08 ***opennesst‐1 0.03 ** 0.02 *** 0.02 ***log(population)t‐1 0.05 –0.06 *** –0.06 ***politicalterrort –2.20 *** –0.72 *** –0.68 ***polityscoret –0.23 * –0.11 ** –0.11 ***interstateconflictt minor –13.98 * –2.03 * –2.22 **war –10.34 *** –7.18 *** –7.70 ***internalconflictw/ointerventiont minor 1.12 –0.56 –0.48war –3.85 * –4.00 *** –3.90 ***internalconflictw/interventiont minor –2.18 0.79 –1.32war –5.05 ** –4.99 *** –5.93 ***UNsanctions(yes/no)t –2.77 ** ––– –2.30 ***USsanctions(yes/no)t ––– –1.07 ** –0.85 *R2 0.29 0.17 0.18 observations 616 2079 2160countries 23 64 68 Notes:DependentvariableistheannualgrowthrateofGDPpercapitain2005USdollars.Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.In confirmation of H1, both UN and US economic sanctions reveal a negative andsignificantinfluenceonthetargetcountry’srealGDPgrowth.Theadverseeffectis–2.77ppwhen onlyUN sanctions are considered and –1.07ppwhen onlyUS sanctions areconsidered.Thisharmful impact is slightlysmallerwhenemployingboth indicators inonespecification,indicatingcollinearitybetweenthevariables.StatisticaltestingrejectsthenullhypothesisthattheadverseeffectofUNsanctions(–2.30pp)andUSsanctions(–0.85pp)isequalatthe10%level(F(1,2043)=2.73*).Therefore,wecanaffirmH3aswellsincetheadverseeffectofeconomicsanctionsonrealGDPgrowthisstrongerforUNsanctionsthanforUSsanctions.4.2DifferentLevelsofSanctions

TheresultsforthetestofH2areshowninTable3.Weestimatethreeversionsofourempirical model: the first includes three indicator variables for UN sanctions only(Column (4)); the second includes the same set of variables for US sanctions only

12

(Column (5)); and the third includesa totalof six sanction indicators forboth theUNandUS(Column(6)).Table3:TheimpactofsanctionsonGDPgrowth:differentsanctionlevels (4) (5) (6)log(realGDP/capita)t‐1 –0.19 *** –0.07 *** –0.08 ***opennesst‐1 0.03 * 0.02 *** 0.02 ***log(population)t‐1 0.06 –0.06 *** –0.05 ***politicalterrort –2.20 *** –0.74 *** –0.70 ***polityscoret –0.28 ** –0.11 ** –0.12 ***interstateconflictt minor –13.85 * –2.02 * –2.21 **war –10.62 *** –7.14 *** –7.97 ***internalconflictw/ointerventiont minor 0.90 –0.54 –0.49war –3.96 * –3.96 *** –3.85 ***internalconflictw/interventiont minor –2.46 0.79 –1.36war –5.21 ** –4.95 *** –5.93 ***UNsanctionst mild –1.69 ––– –1.68 *moderate –3.89 ** ––– –3.43 ***severe –6.03 * ––– –5.30 ***USsanctionst mild ––– –1.25 ** –1.34 ***moderate ––– –0.74 –0.05severe ––– –0.72 0.04 R2 0.30 0.17 0.18observations 616 2079 2160countries 23 64 68 Notes:DependentvariableistheannualgrowthrateofGDPpercapitain2005USdollars.Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.Again, we find some degree of collinearity since most of the coefficients for thesanction variables are slightly smaller in Column (6) compared to the results for UNsanctionsonlyorUSsanctionsonly.Toconservespace,thefollowingdiscussionfocusesontheresultsinColumn(6).In short, the empirical evidence concerning H2 is mixed. The adverse effect of UNsanctions clearly increases over the three categories. Mild sanctions, which includerestrictionsonarmsortravel,leadtoadeclineinthetargetcountry’srealGDPgrowthrateof1.68pp.Moderate sanctions, suchas fuel embargoes, trade restrictions, or the

13

freezing of assets, have a larger adverse effect of –3.43 pp. Severe sanctions, such asembargoesonmostoralleconomicactivity,arethemostharmfultogrowth(–5.30pp).Incontrast,wefindnoevidencethattheadverseeffectofUSsanctionsincreaseswiththeir severity. The only significant coefficient is found for mild sanctions (–1.34 pp),which include retractions of foreign aid. Moderate or severe sanctions, such as traderestrictions or complete embargoes, do not have a significantly negative impact ongrowth.Onepossibleexplanationisthatretractionof foreignaidmightactuallyhurtacountry, whereas unilateral trade restrictions can possibly be circumvented by thetargetstate.That is, the target isstillable to tradewithothercountries,perhapsevenwithsomeoftheothervetopowersontheUNSC—thosenotagreeingtothesanctions—ormightnotevenhavehadastrongtraderelationshipwiththeUSinthefirstplace.7We have at least some evidence in support of H3. The adverse effect of economicsanctionsonthetargetcountry’srealGDPgrowthisstrongerformoderateandsevereUNsanctionsthanforUSsanctionsof thesametype.Statistical testingrejects thenullhypothesisat the5% level (moderate sanctions:F(1,2039)=6.19**; severesanctions:F(1,2039) = 5.67**). In case of mild sanctions, however, we cannot reject the nullhypothesis(F(1,2039)=0.11).Toputourfindingsintoperspective,wecomparetheadverseeffectsofsanctionswiththenegativeconsequencessomeofthecontrolvariablesinColumn(6)ofTable3haveon economic growth. The effect of severe UN sanctions (–5.30 pp) is statisticallyindistinguishable from the consequences of (i) intrastatewars (–3.85pp; F(1,2039) =0.50),(ii) internationalizedintrastatewars(–5.93pp;F(1,2039)=0.08),andeven(iii)interstatewars(–7.97pp;F(1,2039)=1.46).5.RobustnessTests

5.1NarrowingtheControlWindow

Thusfar,ourestimatesare—roughlyspeaking—basedonacomparisonofconditionalmeans of growth in periods duringwhich sanctions are in places compared to timeswhen they are not. It could be argued, however, that the institutional, political, andsocialenvironmentisnotcomparableduringtheseperiods.Furthermore,theimposition7 To confirm this impression, we estimate amodification of (6): we replace the per capita real GDP

growthrateasleft‐handsidevariablebythetradeopennessindicator(i.e.,importsplusexportsdividedby GDP). The results confirm that severe UN sanctions lead to amuch stronger decline in a country’sopenness (–13.28 pp) than severe US sanctions (–3.39 pp) (F(1,2039) = 8.17***). None of the othersanctionvariablesaresignificant.Resultsareavailableonrequest.

14

ofsanctionsmightbeaconsequenceofanenvironmentthatisconsideredbadbytheUNand/or the US. To address this potential endogeneity problemwe reduce the controlsample and, first, consider awindowof five (three) years around the sanctionperiodinsteadofthefullsampleperiod.Acomparisonofconditionalmeansofgrowthduringthe sanction period and this small window of time around it should provide a morerobustestimateoftheadverseeffectsofsanctionssincethe(typicallyslowlychanging)institutional,political,andsocialenvironmentismorelikelytobestableoveranarrowwindowoftime.Thedecisiontoliftsanctions,however,mightbedrivenbyhavingachievedthedesiredchanges in the environment. As a consequence, the years immediately following asanctionperiodmightbecharacterizedbyadifferent institutional,political,andsocialregime aswell. Thus, an obvious robustness test is a further reduction of the controlsample by leaving out all observations after a period of UN and US sanctions. Theremainingsamplecomprisesthefive(three)yearsbeforesanctionswereimposedandthesanctionperioditself.In total, we explore the robustness of our findings with four modifications to thesampleperiod.Inadditiontothesanctionperiod,weconsider(i)awindowoffiveyearsaround,(ii)awindowofthreeyearsaround,(iii)thefiveyearsbefore,and(iv)thethreeyears before. Table A3 in the Appendix presents the results for the binary sanctionvariablesandTableA4for theversioninwhichwetakeaccountof theseverityof thesanctions.Ingeneral, theresults forUNsanctions in therestrictedsamplesaresimilar to thosefor the full sample in termsof sizeandsignificance. In thecaseof thebinarysanctionindicator, theadverseeffect is even slightly larger, ranging from–2.65pp to–3.54pp(TableA3)comparedto–2.30pp(Table2).Thecoefficientsformildsanctionsarealsolarger in absolute terms throughout all modifications compared to the unrestrictedsample.Inthecaseofmoderateandseveresanctions,theestimatesfortheunrestrictedsampleareinbetweenthoseofthetruncatedsamples.Themaximumadverseeffect isfound for severe sanctionswhenonly considering the threeyearsbefore the sanctionperiod(–7.80pp).The results forUS sanctions,however, arenot robust tomodifications in thesampleperiod. The binary sanction indicator is insignificant, irrespective of which samplerestrictionisimposed.Whenincludingdifferentvariablesforthedegreeofseverity,the

15

findingformildsanctionsisreplicatedonlywhenthesampleisrestrictedtofive(three)yearsbeforethesanctionperiod.Therefore, multilateral UN sanctions have a (much) stronger adverse effect on thetarget country’s GDP growth compared to unilateral US sanctions. Asmentioned, thereasonforthiscouldbeassimpleasthatthereisalargernumberofcountriesinvolvedin the imposition of UN sanctions. To summarize, the imposition of UN sanctionsdecreasesthetargetstate’sannualrealpercapitaGDPgrowthrateby2.3–3.5ppandtheimposition of comprehensiveUN economic sanctions triggers a reduction in real GDPgrowthofmorethan5pp.Oneexplanationforthenon‐robustresultsforUSsanctionsmightbethattheirimpactismoreheterogeneousacrosstargetcountries thanthatofUNsanctions.Asdiscussedabove, countries might circumvent US sanctions by increasing their trade with othercountriesortheymaynotevenhavehadarelevanttraderelationshipwiththeUSinthefirst place. Therefore, as part of our robustness tests, we extend Equation (1) byinteractingtheUSsanctionvariableswiththedistanceofthetargetcountry’scapitaltoWashington,DC.That is,wetestwhethergreaterdistance fromtheUS—capturingthebilateral trade potential—leads to a mitigation of the adverse consequences of USsanctions. However, this idea finds no support in the data: the interaction effects areinsignificantwhenemployingeitherthebinarysanctionvariableorindicatorvariablesfordifferentlevelsofsanctions.8AnotherexplanationmightbethattheimpactofUSsanctionsislesslastingthanthatofUNsanctions.Thatis,theeffectofUSsanctionsonthetargetcountry’sGDPistooshort‐lived to be significant in our empirical setup aswe compare average conditionalGDPgrowthratesacrossthesanctionperiodandthenon‐sanctionperiod.WewillreturntothisissueinSection6whereweexploretheimpactofsanctionsovertime.5.2CounterfactualAnalysisforUNSanctions

Next, we apply a control group approach and compare the growth effect of UNeconomicsanctionswhichwereactually imposedandsanctionswhichdidnotbecomeeffective. In this regard,we takeadvantageofapeculiarityof theUNdecision‐makingprocess:beforetheUNmaycalluponitsmemberstatestoimposeeconomicsanctions,theUNSChas to adopt a resolution inwhich itsmembersdeclare that thedesignated8Resultsareavailableonrequest.

16

targetstateeitherthreatens internationalpeaceandsecurityorviolateshumanrights.Theadoptionofsucharesolutionand,thus,theimpositionofeconomicsanctionscanbeprevented by any of the five permanent members of the UNSC—i.e., China, France,Russia,theUnitedKingdom,andtheUnitedStates—sincetheyareendowedwithavetoright.Fortunately,thedraftsofallvetoedresolutionsareaccessibleattheUNwebsite.Wefocusoncountriesagainstwhichresolutionsweredirectedbutfailedduetothevetoof either one or two permanent members of the UNSC.9 Arguably, the pre‐sanctionpolitical and social environment in countries which were actually exposed to UNsanctionsshouldbecomparabletothat incountrieswhichwerealmostsanctioned(atleast on average), in particular, since a majority of UNSC members supported theimposition of sanction measures against countries within the latter group. Thus,countries against which failed resolutions were directed can be considered ascounterfactuals and utilized to examine whether the adverse growth effect of UNsanctionsisdrivenbyomittedfactorsthatcoincidewithsanctionperiods.For this purpose, we extend our sample and include—in addition to the countrieswhichwereactually exposed toUNsanctions—also countrieswhichwerealmost (i.e.,theadoptionofacorrespondingresolutionfailedduetoavetointheUNSC)subjecttoUN sanctions. We then compare the GDP growth effects of realized sanctions andunsuccessful resolutions. To do so,we consecutively add three indicator variables forfailed resolutions toourbaselineempiricalmodel.Ourbinary indicatorvariables takeonthevalue1(i)inthevetoyear,(ii)inthevetoyearplusthetwofollowingyears,(iii)in thevetoyearplus the four followingyears. IfUNeconomicsanctionshavea causalinfluence on economic growth then the effect of the sanctions should be morepronouncedthanthatoffailedresolutions.Table A5 in the Appendix presents the results for the binary sanction variable andTableA6fortheversioninwhichwetakeaccountoftheseverityofthesanctions.Thefindingssuggestthatfailedresolutionsdonotaffectthedesignatedtargetcountry’sGDPgrowth.Ineachofthesixspecifications,theindicatorvariableforvetoedresolutionsisnot statistically different from zero. What is more, the adverse effect of economicsanctions—regardless of their severity—is notably larger than that of vetoed

9 Failed resolutions were directed against 13 countries: Argentina, Bosnia and Herzegovina, China,

Cyprus,France,Guatemala,Israel,Macedonia,Myanmar,Syria,UK,Vietnam,andZimbabwe.

17

resolutions. Thus, our previous results are unlikely affected by omitted factors thatcoincidewithsanctionperiods.6.ImpactofSanctionsoverTime

So far, we implicitly assumed that the effects of UN and US economic sanctions aretime‐invariant. However, there is some reason to believe that the detrimental impacteconomicsanctionsexertonGDPgrowthisdecreasingovertime.First, the lengthofasanctionperiodmay indicate thestrengthof the incumbentpoliticalregime.Arguably,the longer the target state’s government can withstand the economic and politicalpressure associatedwith economic sanctions, the lower are the expectations that thesanction measures actually trigger desired changes in the political and socialenvironment.Thismayrestoreinvestors’confidenceinthestabilityofthetargetstate’spoliticalandlegalsystem.Second,aftersometimehaspassed,thetargetgovernmentaswellas theeconomicagentswithin the targetcountrymayadapt to thenewsituationand learn how to successfully evade sanctionmeasures, reducing the economic coststhey incur. Finally, sanctions which are particularly harmfulmay also be particularlyeffectiveandwillthusbeliftedsooner.Toexaminethedevelopmentofthesanctioneffectsovertime,weextendEquation(1)by interacting thesanction indicatorswithavariable thatmeasures theyearselapsedsince the respective sanction has been imposed.10 The results are shown in Table 4.Column (7)provides the estimates for thebinary sanction variables,whereasColumn(8) offers insight into the dynamics when distinguishing between different sanctioncategories.

10Notethatwealsoconsideredinteractionsofthesanctionvariableswiththesquarednumberofyears

elapsed since the respective sanction has been imposed to capture non‐linearities in the impact ofsanctionsovertime.However,theresultingestimatesyieldimplausibledynamicsand,therefore,arenotshownbutavailableonrequest.

18

Table4:TheimpactofsanctionsonGDPgrowthovertime:binarysanctionvariableanddifferentsanctionlevels (7) (8)log(realGDP/capita)t‐1 –0.08 *** –0.08 ***opennesst‐1 0.02 *** 0.02 ***log(population)t‐1 –0.06 *** –0.06 ***politicalterrort –0.55 ** –0.65 ***polityscoret –0.12 *** –0.13 ***interstateconflicttminor –2.24 ** –2.32 **war –6.76 *** –6.57 ***internalconflictw/ointerventiont minor –0.30 –0.09war –3.68 *** –3.47 ***internalconflictw/interventiontminor –1.37 –1.33war –6.67 *** –6.57 ***UNsanctions(yes/no)t –4.88 *** ––– …*years 0.32 *** –––USsanctions(yes/no)t –1.89 *** ––– …*years 0.17 *** ––– UNsanctionst mild ––– –3.50 ***mild*years ––– 0.25 *moderate ––– –4.57 ***moderate*years ––– 0.11 severe ––– –13.10 ***severe*years ––– 1.64 ***USsanctionst mild ––– –2.19 ***mild*years ––– 0.19 *moderate ––– –2.73 **moderate*years ––– 0.34 ***severe ––– 0.34severe*years ––– –0.01 R2 0.18 0.19Observations 2160 2160Countries 68 68 Notes:DependentvariableistheannualgrowthrateofGDPpercapitain2005USdollars.Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.Our findings for thebinary sanction indicators suggest that theeffectsofUNandUSeconomicsanctionsvaryconsiderablyover timeasboth the linearandthe interaction

19

termsareofnotablesizeandstatisticallysignificant.TheinitialnegativeinfluenceofUNsanctionsonGDPgrowth is–4.88pp,which isnotably larger(inabsolute terms) thanthe correspondingestimate in ourbaseline specification (–2.30pp; seeColumn (3) ofTable2).However, thisdetrimentaleffectbecomessmallerovertime.WitheveryyearthatpassesaftertheimpositionofaUNsanctiontheadversegrowtheffectdecreasesby0.32 pp.We obtain a very similar picture forUS sanctions. In the year inwhich aUSsanctionisadopted,thetargetstate’sGDPgrowthratedecreasesby–1.89pp.AsinthecaseofUNsanctions,thiseffectdiminishesby0.17ppwitheveryyearthatpassesaftertheimpositionofUSsanctionmeasures.Strikingly,both,theinitialandthetime‐varyingeffect ofUS sanctions are significant at the1% level,whereas the estimate for theUSsanctionindicatorinaspecificationwithoutaninteractiontermisonlysignificantatthe10%level(seeColumn(3)ofTable2).To facilitate interpretationof the interaction termsand toglean further insights intothe development of the sanction effects over time we graphically illustrate time‐dependentmarginaleffectsalongwith90%confidencebandsforthebinaryUNandUSsanctionindicators.Figure2:TheimpactofsanctionsonGDPgrowthovertime:binarysanctionindicator

UNsanctions USsanctions

Notes: Figure shows impact of sanctions on GDP growth over time for the binary sanction indicator.EstimatesarebasedontheresultsfromColumn(7)inTable4.Thedottedlinesrepresent90%confidenceintervals.As Figure 2 shows, UN sanctions exert a significant negative influence on the targetcountry’sGDPgrowthfor10years,whereastheadverseeffectofUSsanctionslastsfor7years. In addition, the detrimental impact of UN sanctions is significantly larger (in

‐7‐6‐5‐4‐3‐2‐101

1 2 3 4 5 6 7 8 9 10 11 12‐7‐6‐5‐4‐3‐2‐101

1 2 3 4 5 6 7 8 9 10 11 12

20

absolute terms) than thatofUS sanctions throughout the first eight years.11Thus, theeconomiccostsUNsanctionsinflictonthetargetstatearenotablylargerthanthoseofUSsanctionsasUNsanctionsexertastrongernegativeinfluenceonGDPgrowthandarealsolongerlasting.We obtain similar results when distinguishing between different levels of sanctions(Column(8)ofTable4). Initially,mild (–3.50pp),moderate (–4.57pp),andsevere (–13.10 pp) UN sanctions have a huge negative influence on the target country’s GDPgrowth. These adverse effects are mitigated over time by 0.25 pp (mild), 0.11 pp(moderate),and1.64pp(severe)witheveryyearthatpassesaftertheimpositionoftherespective sanctionmeasures. Turning to US sanctions, the picture from our baselinespecification changes a bit: in addition to mild sanctions (–2.19 pp), also moderatesanctions(–2.73pp) initiallyexertasignificantnegative influenceonGDPgrowth; theeffectofthelattersanctioncategorywasinsignificantwhencomputinganaverageeffectoverthewholesanctionperiod(seeColumn(6)inTable3).Forboth,mildandmoderateUSsanctions,weobserveasignificantdecreaseofthedetrimentaleffectovertime(mild:0.19 pp; moderate 0.34 pp). The impact of severe sanctions, however, remainsinsignificant.Figure A1 in the Appendix graphically illustrates the corresponding time‐dependentmarginaleffectsfordifferent levelsofsanctions.Theresultsarequalitativelythesameasforthebinarysanctionindicators.TheinfluenceofUNsanctionsislongerlastingthanthat of US sanctions. Mild, moderate, and severe UN sanctions exert a statisticallysignificant influenceonGDPgrowth for7,14,and6years, respectively,whereasmild,moderate, and severe US sanctions have a detrimental effect for 6, 4, and 0 years,respectively.Finally, as in Section 5.1, we explore the robustness of our findings with fourmodificationstothesampleperiod.Inadditiontothesanctionperiod,weconsider(i)awindowof five years around, (ii) awindowof three years around, (iii) the five yearsbefore,and(iv)thethreeyearsbefore.TableA7intheAppendixpresentstheresultsforthe binary sanction variables and Table A8 for the specifications in which we takeaccountoftheseverityofthesanctions.The most important finding from these robustness tests is that, in contrast to thespecificationswithoutaninteractionterm(seeTablesA3andA4intheAppendix),the11Thedifferenceis–1.72ppafter8years(t=–1.83*)and–1.56ppafter9years(t=–1.58).

21

resultsfortheUSbinarysanctionindicatoraswellasformildUSsanctionsarerobustwith respect to modifications to the control sample. The results for moderate USsanctions, however, become insignificant in both specifications in which the controlwindowendsafterthesanctionperiod(Columns(A21)and(A22)inTableA8).Turningto UN sanctions, the results remain qualitatively unchanged.12 To summarize, therobustnesstestsconfirmthatitiscrucialtoaccountfortime‐variationwhenestimatingtheeffectsofUSsanctionsonthetargetcountry’sGDPgrowth.7.Conclusions

Inthispaper,weempiricallyassesshow(i)multilateraleconomicsanctions imposedbytheUnitedNationsand(ii)unilateralsanctions imposedbytheUnitedStatesaffectthetargetstates’realGDPgrowth.Weaugmentastandardgrowthmodelby indicatorvariables forUNandUSsanctions,also taking intoaccount that thereasonseconomicsanctions are imposed—that is, engagement in interstate conflicts, autocratictendencies, and political repression—might in themselves affect a country’s economicdevelopment. Our sample includes 68 countries and covers the period from 1976 to2012.Ourresultssuggest,first,thatsanctionsimposedbytheUNhaveasignificantinfluenceoneconomicgrowth.Onaverage, the impositionofUN sanctionsdecreases the targetstate’srealpercapitaGDPgrowthrateby2.3–3.5pp.Aninvestigationofthedynamicsofthe sanction effects reveals that the detrimental influence decreases over time andbecomes insignificant after 10 years. We find that comprehensive UN economicsanctions—embargoesonalmostalleconomicactivitybetweenUNmemberstatesandthe sanctioned country—have the strongest influence; they trigger a reduction in realGDPgrowthbymorethan5pp.Ourfindingsarerobusttomodificationsofourempiricalspecificationthatcontrolforpotentialchangesinacountry’sinstitutional,political,andsocialenvironment.Moreover,wecompareannualrealGDPgrowthratesduringactualUN sanction periods and the years after an unsuccessful attempt by the UN SecurityCounciltoimposesanctions(i.e.,whentheimpositionofsanctionswaspreventedbyavetoofapermanentmemberof theUNSC).Our findingssuggest thatrealGDPgrowth12NotethattheinteractiontermforthebinaryindicatorbecomesinsignificantinColumns(A15)–(A18)

ofTableA7whichimpliesthatsanctionsexertanegativeinfluenceonthetargetcountry’sGDPgrowthfor11–15 years, depending on the specification. The same holds for the interaction terms of mild andmoderatesanctionsininColumns(A19)–(A22)ofTableA8.

22

declinesonlywheneconomicsanctionsareactuallyimposed,indicatingthatourresultsarenotdrivenbyomittedfactorsthatcoincidewithUNsanctionepisodes.Second,theeffectofUSsanctionsismuchsmallerandrobustonlywhenallowingfortime‐variationin the effect of sanctions on growth. The imposition of US sanctions decreases GDPgrowthinthetargetstateoveraperiodof7yearsand,onaverage,by0.5–0.9pp.Our results suggest that multilateral UN sanctions are indeed harmful to the targetcountry’s economy and have a (much) stronger adverse effect than unilateral USsanctions. Whether these sanctions are an appropriate (irrespective of theireffectiveness) tool for compelling governments to comply with the UN’s interestsremainsunclear,especiallyinlightofthefrequentcriticismthattheyoftencausemoredamagetothepoorthantothepoliticalelite.Aninterestingandusefulextensionofthisworkwouldbe todiscover theconsequencesofeconomicsanctions forpoverty in thetargetcountry.13

13Atthetimeofthiswriting,suchananalysisisvirtuallyimpossible,chieflyduetothelargenumberof

missing country/year observations.World Bank poverty data are based on primary household surveydata obtained from government statistical agencies and World Bank country departments and rarelyavailableduringperiodsofsanctions.

23

References

Aizenman, J. and Marion, N.P. (1993), Policy uncertainty, persistence and growth,ReviewofInternationalEconomics1(2),145–163.Alesina, A., Özler, S., Roubini, N., and Swagel, P. (1996), Political instability andeconomicgrowth,JournalofEconomicGrowth1(2),189–211.Alesina, A. and Perotti, R. (1996), Income distribution, political instability, andinvestment,EuropeanEconomicReview40(6),1203–1228.AliMohamed,M.andShah,I.(2000),SanctionsandchildhoodmortalityinIraq,Lancet355,1851–1856.Allen, S. H. (2008), The domestic political costs of economic sanctions, Journal ofConflictResolution52(6),916–944.Andreas,P.(2005),Criminalizingconsequencesofsanctions:Embargobustinganditslegacy,InternationalStudiesQuarterly49,335–360.Cortright,D.andLopez,G.,eds.(2000),Thesanctionsdecade:AssessingUNstrategiesinthe1990s,Boulder,CO:LynneRienner.Crawford,N. C. andKlotz, A. (1999),How sanctionswork:Lessons from SouthAfrica,NewYork:St.Martin’sPress.Daponte,B.andGarfield,R.(2000),Theeffectofeconomicsanctionsonthemortalityof Iraqichildrenpriortothe1991PersianGulfWar,American JournalofPublicHealth90(4),546–552.Dashti‐Gibson,J.,Davis,P.,andRadcliff,B.(1997),Onthedeterminantsofthesuccessofeconomicsanctions:Anempiricalanalysis,AmericanJournalofPoliticalScience41(2),608–618.Drury, A. C. (1998), Revisiting economic sanctions reconsidered, Journal of PeaceResearch35(4),497–509.Eaton, J. andEngers,M. (1992), Sanctions, JournalofPoliticalEconomy 100(5), 899–928.Evenett, S. J. (2002), The impact of economic sanctions on South African exports,ScottishJournalofPoliticalEconomy49(5),557–573.Garfield, R. (2002), Economic sanctions, humanitarianism and conflict after the ColdWar,SocialJustice29(3),94–107.Gibbons, E. andGarfield,R. (1999),The impact of economic sanctionsonhealth andhumanrightsinHaiti1991–1994,AmericanJournalofPublicHealth89(10),1499–1504.

24

Heine‐Ellison, S. (2001), The impact and effectiveness of multilateral economicsanctions:Acomparativestudy,InternationalJournalofHumanRights5(1),81–112.Hufbauer,G.(1998),Sanctions‐happyUSA,InstituteforInternationalEconomicsPolicyBrief 98‐4, available at http://www.iie.com/publications/pb/pb98‐4.htm (accessed inMarch2014).Hufbauer, G., Schott, J., Elliott, K. A., and Oegg, B. (2009), Economic sanctionsreconsidered: History and current policy, 3rd edition, Washington, DC: Institute forInternationalEconomics.Kaempfer,W.H. and Lowenberg,A.D. (1988), The theory of international economicsanctions:Apublicchoiceapproach,AmericanEconomicReview78(4),786–793.Kaempfer, W. H. and Lowenberg, A. D. (1999), Unilateral versus multilateralinternational sanctions: A public choice perspective, International Studies Quarterly43(1),37–58.Kaempfer,W. H., Lowenberg, A. D., andMertens,W. (2004), International economicsanctionsagainstadictator,EconomicsandPolitics16(1),29–51.Marinov,N.(2005),Doespressurefromtheoutsidedestabilizeleadersontheinside?AmericanJournalofPoliticalScience49(3),564–576.Peksen,D.(2009),Betterorworse?Theeffectofeconomicsanctionsonhumanrights,JournalofPeaceResearch46(1),59–77.Peksen, D. and Drury, A. C. (2010), Coercive or corrosive: The negative impact ofeconomicsanctionsondemocracy,InternationalInteractions36(3),240–264.Weiss,T.,Cortright,D.,Lopez,G.,andMinear,L.(1997),Politicalgainandcivilianpain,Boulder,CO:RowmanandLittlefield.Whang,T.(2011),Playingtothehomecrowd?SymbolicuseofeconomicsanctionsintheUnitedStates,InternationalStudiesQuarterly55(3),787–801.Wood,R.M. (2008), “Ahandupon the throatof thenation”:Economicsanctionsandstaterepression,1976–2001,InternationalStudiesQuarterly52(3),489–513.

25

Appendix

TableA1:ListofsamplecountriesAfrica(22).Angola,Cameroon,CentralAfricanRepublic,DemocraticRepublicCongo,Eritrea, Ethiopia,Gambia,Guinea‐Bissau,Kenya, Liberia, Libya,Malawi,Niger,Nigeria,Rwanda,SierraLeone,Somalia,SouthAfrica,Sudan,Uganda,Zambia,Zimbabwe.America(16).Argentina,Bolivia,Brazil,Chile,Colombia,Cuba,Ecuador,ElSalvador,Guatemala,Haiti,Honduras,Nicaragua,Panama,Paraguay,Peru,Uruguay.Asia (19). Afghanistan, Cambodia, China, India, Indonesia, Iran, Iraq, Israel, Jordan,Lebanon, Myanmar, North Korea, Pakistan, South Korea, Syria, Thailand, Uzbekistan,Vietnam,Yemen.Europe (10). Azerbaijan, Belarus, Bosnia and Herzegovina, Croatia, Macedonia,Montenegro,Poland,Romania,Serbia,Turkey.Oceania(1).Fiji.

26

TableA2:Variabledescriptionanddatasourcesg(realGDP/capita).100( / 1 ,in2005USdollars.log(realGDP/capita).100 ,in2005USdollars.openness.100 / .log(population).100log .Source:UN.political terror.Terrorscalemeasuringphysical integrityrightsviolationsbasedonUSStateDepartmentratings;rangesfrom1(lowestvalue)to5(highestvalue).Source:PoliticalTerrorScale.polityscore.Polityscalevariable;rangesfromstronglydemocratic(+10)tostronglyautocratic(–10).Source:PolityIVDatabase.interstate conflict. Interstate armed conflict between two ormore states; indicatorvariablesforminorconflicts(between25and999battle‐relateddeathsinagivenyear)andwars(atleast1,000battle‐relateddeathsinagivenyear).internalconflictw/ointervention.Internalarmedconflictbetweenthegovernmentofastateandoneormoreinternaloppositiongroup(s)withoutinterventionfromotherstates;indicatorvariablesforminorconflictsandwars.internalconflictw/intervention.Internationalizedinternalarmedconflictbetweenthe government of a state and one or more internal opposition group(s) withintervention from other states on one or both sides; indicator variables for minorconflictsandwars.Source:UCDP/PRIOArmedConflictDataset.UNsanctions.AsdefinedinTable1.Source:OwncollectionandWood(2008).USsanctions.AsdefinedinTable1.Source:Hufbaueretal.(2009),Wood(2008),andowncollection.

27

Table A3: The impact of sanctions on GDP growth: robustness test for different timewindowsandthebinarysanctionvariable (A1) (A2) (A3) (A4)log(realGDP/capita)t‐1 –0.11 ** –0.13 ** –0.12 *** –0.13 ***opennesst‐1 0.02 ** 0.02 ** 0.02 * 0.02 *log(population)t‐1 –0.06 * –0.03 –0.02 –0.01 politicalterrort –1.04 ** –1.30 ** –1.38 *** –1.81 ***polityscoret –0.12 * –0.13 * –0.11 –0.15 *interstateconflictt minor –1.81 –1.65 –1.93 –1.42war –7.19 ** –6.88 ** –7.22 *** –7.24 ***internalconflictw/ointerventiont minor –0.83 –1.33 –1.31 –1.25war –5.46 ** –6.20 ** –7.02 *** –6.78 ***internalconflictw/interventiont minor –0.61 –2.73 –2.41 –2.93war –5.77 ** –5.91 ** –6.26 *** –5.87 ***UNsanctions(yes/no)t –3.01 ** –3.54 ** –2.65 ** –3.20 ***USsanctions(yes/no)t –0.52 –0.55 –0.83 –0.81timewindow [–5;+5] [–3;+3] [–5;0] [–3;0] R2 0.20 0.23 0.22 0.25observations 1337 1106 1045 915countries 68 68 68 68 Notes:DependentvariableistheannualgrowthrateofGDPpercapitain2005USdollars.Inadditiontotheactualsanctionperiod,Columns(A1)and(A2)includeawindowofonlyfiveandthreeyearsaroundthis period, respectively. Columns (A3) and (A4) restrict the sample to five and three years before thesanctionperiod(whichisalsoincluded),respectively.Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.

28

Table A4: The impact of sanctions on GDP growth: robustness test for different timewindowsanddifferentsanctionlevels (A5) (A6) (A7) (A8)log(realGDP/capita)t‐1 –0.11 ** –0.13 ** –0.12 *** –0.14 ***opennesst‐1 0.03 ** 0.02 ** 0.02 0.02 *log(population)t‐1 –0.05 * –0.03 –0.02 0.00 politicalterrort –1.07 ** –1.33 ** –1.42 *** –1.89 ***polityscoret –0.13 ** –0.14 ** –0.12 –0.15 *interstateconflictt minor –1.82 –1.64 –1.84 –1.28war –7.47 ** –7.32 ** –7.72 *** –7.97 ***internalconflictw/ointerventiont minor –0.79 –1.35 –1.30 –1.29war –5.33 ** –6.18 ** –6.90 *** –6.78 ***internalconflictw/interventiont minor –0.59 –2.81 –2.55 –3.26war –5.68 ** –5.89 ** –6.27 *** –5.93 ***UNsanctionst mild –2.41 ** –3.26 ** –2.97 ** –4.06 ***moderate –4.06 ** –4.08 ** –2.84 ** –2.95 **severe –5.15 ** –6.32 ** –6.65 *** –7.80 ***USsanctionst mild –0.96 –1.02 –1.63 ** –1.80 *moderate 0.21 0.50 0.94 1.45severe 0.27 0.06 0.86 0.99timewindow [–5;+5] [–3;+3] [–5;0] [–3;0] R2 0.20 0.23 0.23 0.26observations 1337 1106 1045 915countries 68 68 68 68 Notes:DependentvariableistheannualgrowthrateofGDPpercapitain2005USdollars.Inadditiontotheactualsanctionperiod,Columns(A5)and(A6)includeawindowofonlyfiveandthreeyearsaroundthis period, respectively. Columns (A7) and (A8) restrict the sample to five and three years before thesanctionperiod(whichisalsoincluded),respectively.Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.

29

TableA5:TheimpactofUNsanctionsonGDPgrowth:robustnesstestincludingvetoedUNresolutionsandthebinarysanctionvariable (A9) (A10) (A11)log(realGDP/capita)t‐1 –0.11 *** –0.11 *** –0.11 ***opennesst‐1 0.03 ** 0.03 ** 0.03 **log(population)t‐1 –0.04 –0.04 –0.04 politicalterrort –1.29 *** –1.33 *** –1.34 ***polityscoret –0.28 *** –0.28 *** –0.29 ***interstateconflictt minor –8.69 *** –9.09 *** –9.25 ***war –8.63 *** –8.78 *** –8.77 ***internalconflictw/ointerventiont minor 0.65 0.63 0.62war –5.47 *** –5.52 *** –5.46 ***internalconflictw/interventiontminor –3.70 –3.69 –3.67war –5.22 *** –5.18 *** –5.11 ***UNresolutionvetoed(yes/no)t –0.37 1.46 1.72 UNsanctions(yes/no)t –2.93 *** –2.89 *** –2.86 ***timewindow [0;+1] [0;+3] [0;+5]R2 0.22 0.22 0.22observations 982 982 982countries 33 33 33 Notes:Dependentvariable is theannualgrowthrateofGDPpercapita in2005USdollars.Thedummyvariable ‘UNresolutionvetoed (yes/no)t’ takes thevalue1during theyearof theveto inColumn (A9),during a three‐year window (including the year in which the veto took place) after a veto in Column(A10),andduringafive‐yearwindowafteravetoinColumn(A11).Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.

30

TableA6:TheimpactofUNsanctionsonGDPgrowth:robustnesstestincludingvetoedUNresolutionsanddifferentsanctionlevels (A12) (A13) (A14)log(realGDP/capita)t‐1 –0.11 *** –0.11 *** –0.11 ***opennesst‐1 0.03 ** 0.03 ** 0.03 **log(population)t‐1 –0.03 –0.03 –0.04 politicalterrort –1.29 *** –1.33 *** –1.35 ***polityscoret –0.32 *** –0.33 *** –0.33 ***interstateconflictt minor –8.69 *** –9.11 *** –9.27 ***war –8.95 *** –9.10 *** –9.10 ***internalconflictw/ointerventiont minor 0.46 0.43 0.43war –5.60 *** –5.65 *** –5.59 ***internalconflictw/interventiontminor –3.93 * –3.91 * –3.89 *war –5.42 *** –5.37 *** –5.30 ***UNresolutionvetoed(yes/no)t –0.28 1.54 1.81 UNsanctionst mild –1.83 –1.77 –1.73moderate –3.98 *** –3.97 *** –3.95 ***severe –6.20 ** –6.18 ** –6.19 **timewindow [0;+1] [0;+3] [0;+5]R2 0.22 0.22 0.22observations 982 982 982countries 33 33 33 Notes:Dependentvariable is theannualgrowthrateofGDPpercapita in2005USdollars.Thedummyvariable ‘UNresolutionvetoed(yes/no)t’ takesthevalue1duringtheyearofthevetoinColumn(A12),during a three‐year window (including the year in which the veto took place) after a veto in Column(A13),andduringafive‐yearwindowafteravetoinColumn(A14).Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.

31

FigureA1:TheimpactofsanctionsonGDPgrowthovertime:differentsanctionlevelsUNmildsanctions USmildsanctions

UNmoderatesanctions

USmoderatesanctions

UNseveresanctions

USseveresanctions

Notes:FigureshowsimpactofsanctionsonGDPgrowthovertimefordifferentsanctionlevels.EstimatesaretakesfromColumn(8)inTable4.90%confidenceintervalsarerepresentedbydottedlines.

‐6

‐4

‐2

0

2

1 2 3 4 5 6 7 8 9 10 11 12‐6

‐4

‐2

0

2

1 2 3 4 5 6 7 8 9 10 11 12

‐8

‐4

0

4

1 2 3 4 5 6 7 8 9 10 11 12‐8

‐4

0

4

1 2 3 4 5 6 7 8 9 10 11 12

‐12‐8‐4048

1 2 3 4 5 6 7 8 9 101112‐12‐8‐4048

1 2 3 4 5 6 7 8 9 101112

32

Table A7: The impact of sanctions on GDP growth over time: robustness test fordifferenttimewindowsandthebinarysanctionvariable (A15) (A16) (A17) (A18)log(realGDP/capita)t‐1 –0.11 *** –0.13 *** –0.12 *** –0.14 ***opennesst‐1 0.03 ** 0.03 ** 0.02 * 0.03 **log(population)t‐1 –0.06 ** –0.04 –0.03 –0.01 politicalterrort –0.91 *** –1.19 *** –1.24 *** –1.70 ***polityscoret –0.12 ** –0.14 ** –0.10 –0.13 interstateconflictt minor –1.80 –1.53 –1.88 –1.23 war –6.48 *** –6.22 *** –6.46 *** –6.53 ***internalconflictw/ointerventiont minor –0.72 –1.29 –1.02 –0.94 war –5.17 *** –5.95 *** –6.54 *** –6.20 ***internalconflictw/interventiont minor –0.80 –2.87 –2.58 –2.89 war –6.45 *** –6.57 *** –7.14 *** –6.89 ***UNsanctions(yes/no)t –4.65 *** –4.86 *** –4.58 *** –4.95 ***…*years 0.18 0.11 0.20 0.18 USsanctions(yes/no)t –1.60 ** –1.82 ** –2.04 ** –2.27 ** …*years 0.18 *** 0.22 *** 0.27 *** 0.34 ***timewindow [–5;+5] [–3;+3] [–5;0] [–3;0] R2 0.21 0.23 0.23 0.26 observations 1337 1106 1045 915 countries 68 68 68 68 Notes:DependentvariableistheannualgrowthrateofGDPpercapitain2005USdollars.Inadditiontotheactualsanctionperiod,Columns(A15)and(A16)includeawindowofonlyfiveandthreeyearsaroundthisperiod,respectively.Columns(A17)and(A18)restrictthesampletofiveandthreeyearsbeforethesanctionperiod(whichisalsoincluded),respectively.Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.

33

Table A8: The impact of sanctions on GDP growth over time: robustness test fordifferenttimewindowsanddifferentsanctionlevels (A19) (A20) (A21) (A22)log(realGDP/capita)t‐1 –0.11 *** –0.14 *** –0.12 *** –0.15 ***opennesst‐1 0.03 ** 0.03 ** 0.02 0.02 log(population)t‐1 –0.05 * –0.02 –0.01 0.01 politicalterrort –1.05 *** –1.30 *** –1.47 *** –1.96 ***polityscoret –0.14 ** –0.15 ** –0.12 –0.16 *interstateconflictt minor –1.83 –1.44 –1.86 –1.17 war –6.15 *** –6.00 *** –6.46 *** –6.63 ***internalconflictw/ointerventiont minor –0.39 –0.99 –0.57 –0.46 war –4.85 *** –5.76 *** –6.09 *** –5.86 ***internalconflictw/interventiont minor –0.67 –2.91 –2.43 –2.86 war –6.22 *** –6.48 *** –6.79 *** –6.49 ***UNsanctionst mild –3.51 ** –3.83 ** –3.67 ** –4.34 **mild*years 0.13 0.04 0.10 0.07 moderate –3.98 * –3.85 * –3.72 * –3.76 *moderate*years –0.08 –0.15 0.04 0.01 severe –12.52 *** –13.52 *** –14.53 *** –15.77 ***severe*years 1.51 *** 1.43 ** 1.55 *** 1.57 ***USsanctionst mild –1.71 * –1.85 * –2.65 *** –3.09 ***mild*years 0.17 0.19 0.32 ** 0.42 **moderate –2.64 * –2.88 * –1.62 –1.18 moderate*years 0.37 *** 0.43 *** 0.41 *** 0.45 ***severe 0.17 –1.06 0.88 0.06 severe*years 0.03 0.11 0.09 0.19 timewindow [–5;+5] [–3;+3] [–5;0] [–3;0] R2 0.22 0.24 0.25 0.28 observations 1337 1106 1045 915 countries 68 68 68 68 Notes:DependentvariableistheannualgrowthrateofGDPpercapitain2005USdollars.Inadditiontotheactualsanctionperiod,Columns(A19)and(A20)includeawindowofonlyfiveandthreeyearsaroundthisperiod,respectively.Columns(A21)and(A22)restrictthesampletofiveandthreeyearsbeforethesanctionperiod(whichisalsoincluded),respectively.Modelincludescountry‐fixedeffectsandtime‐fixedeffects.***/**/*indicatessignificanceatthe1%/5%/10%level.

Deckblatt 24-2014Submission