Social!v.!Conservative!Democracies!and!Homicide!Rates! % % % % % ICAT%Working%Paper%Series% January%2012%!

Transcription

1 ICATWorkingPaperSeries January2012!!!!! Social!v.!Conservative!Democracies!and!Homicide!Rates! Luc!Noiset!and!Marcus!Marktanner! KennesawStateUniversity

2 ABSTRACT! Economictheorypredictsthatdemocracieshavelessinequalitythanautocracies.Thereisalso evidence that inequality and homicide rates correlate positively. While homicide rates should thereforebelowerindemocracies,recentstudiesfindoppositeresults.toresolvethispuzzle, we distinguish between high and low incomeninequality democracies (conservative vs. social democracies).weincorporatethisideainarationalnbehaviormodelinwhichamedianvoter potentially joins the criminal class if his calls for equitable social development are not met. Empirical results confirm that homicide rates are higher in conservative democracies where inequalityishighdespitestrongdemocraticelectoralprocedures. Keywords: Homiciderates,comparativeeconomics,conservativevs.socialdemocracy, economicsofcrime,incomeinequality! 2!!Socialv.ConservativeDemocraciesandHomicideRates

3 1.!INTRODUCTION! In his background paper to the 2011 World Bank Development Report Fearon (2011) notes,asothershaveearlier,thatincomeinequalityisamajordeterminantofhomiciderates (see, for example, Neumayer, 2003; LaFree and Tseloni, 2006; Hall and McLean, 2009; and WilkinsonandPickett,2009).Hefurthernotesthatdemocraciestendtohavehigherhomicide ratesthanautocraticcountries.theseobservationsareindeedpuzzling. Economic theory predicts that democracies should have less income inequality than authoritarian regimes. In democratic countries, the poorer majority can presumably vote for policies that tax the richer minority to limit the degree of income inequality; thus one would expectbothlessincomeinequalityandlowerhomicideratesinmoredemocraticcountries,not theotherwayaround. Of course, this simple view of democracy is not always realized in the face of the complexitiesofrealworldpolitics.inreality,democraciesaroundtheworlddifferfromeach other in many significant ways. Since democracies are not homogeneous entities, it may be promisingtoexplorewhetherdifferenttypesofdemocraciesgeneratedifferenthomiciderates. Thisistheobjectiveofthisstudy. Althoughmanystudiespointtoincomeinequalityasarobustexplanatoryvariablefor homicideratesacrosssocieties,thenexusbetweenincomeinequalityandthecharacteristicsof democraciesismoredifficulttoestablish.thisstudyusesthewidelyappliedpolity'scorefrom the Center for Systemic Peace to characterize democracies. The Polity score classifies procedural characteristics of regimes on a spectrum from autocracy to democracy on a scale ranging from minus ten to plus ten. This study gives particular attention to regimes with an InternationalConflictAnalysisandTransformation.January2012! 3

4 average Polity score greater than five between 1960 and All regimes within this range areconsideredestablisheddemocracies. By regressing the Polity score of established democracies on countrynlevel Gini coefficients,weshowthatestablisheddemocracieswithhigherpolityscoresdoinfacthave,on average,lowerincomeinequality.mirroringaclassificationsuggestedbyhsu(2008),wethen usearesidualanalysistoidentifycountrieswithpositiveandnegativeresidualsasconservative andsocialdemocracies,respectively.usingtheseresiduals,anindexiscreatedthatreflectsthe degree to which each democracy can be classified as a more conservative democracy on the one hand, or a more social democracy on the other. Empirical results using this index as an explanatoryvariablesuggestthathigherhomicideratesarecharacteristicofconservative,but not of social, democracies. These findings allow for the conclusion that it is not more fully institutionalizeddemocraticproceduresandcivilfreedomsthatleadtohigherhomiciderates, butratherthefailureofsomedemocraciestotranslatedemocraticprocedures,institutionsand freedomsintoequitablesocialdevelopment. Theremainderofthispaperisorganizedasfollows.Sectiontwoprovidesabriefreview ofthelargeliteratureonthedeterminantsofhomiciderates.insectionthree,asimplerational choiceeconomicmodelispresentedinwhichdemocraciesthatareunabletoprovideequitable social development produce higher homicide rates. The presentation and discussion of the empirical results follow in section four. The paper concludes with a summary of the main findings. 4!!Socialv.ConservativeDemocraciesandHomicideRates

5 2.!LITERATURE!REVIEW! Theliteratureonhomicidesexplainshomiciderateswithdemographic,socioNeconomic, andpoliticalfactors.oneofthefirstresearchedvariableswasage.neison(1857)andgoring s (1913) work on England and Wales were probably the first comprehensive studies in this regard.theirstudiesfoundthatmostmurderersbelongedtotheagegroupbetween16and25. More recently, Pampel and Gardner(1995) provide a comprehensive literature review of the agenhomicidenexus,givingspecialemphasistotheinteractionbetweenageandseveralsocion economicvariables. Urbanizationisanotherfactorthatreceivesparticularattention.LodhiandTilly(1973) study the crime rates of early 19 th century France. The authors find no evidence that urbanization produces more tension and crime. Jalil and Iqbal (2010), however, argue that urbanization has increased crime rates in Pakistan. Theoretical arguments can be made in either direction. Urbanization is associated with more poverty and points to a direct effect betweenurbanizationandcrimerates.thecounterargumentisthaturbanizationbringsabout developmentandjobs.generalizationsinthisareaarethereforedifficulttomake. The comparative criminology literature also discusses inequality as an explanatory factor.blauandblau(1982),forexample,findthatallegedracialandgeographicalpatternsof violenceintheusbecomeinsignificantonceonecontrolsforeconomicinequality.alsousing US data, Brush (2007) finds that levels of inequality and homicide rates are positively associated. Surprisingly, however, the same regressions on differences generate negative results. Referencing Levitt (2004), Brush (2007) suggests that this is most likely due to InternationalConflictAnalysisandTransformation.January2012! 5

6 unmeasured timenspecific effects like more policing, rising incarceration rates, declining crack consumption,andlegalizingabortionbeingabsorbedbythecoefficientonincomeinequality. Political characteristics are often behind demographic and socioeconomic factors. Neumayer (2003) reports evidence for good governance to reduce violent crime. Pridemore and Kim (2006) explain the increase in homicide rates in Russia during the transformation processbyapplyingdurkheim s(1893)anomietheoryandlinkingittoahighlysignificantselfn developedindexofnegativesocioeconomicchange.closelyrelated,lafreeandtseloni(2006, p.30)findthatcountriesintransitiontofreemarketsexperienceanincreaseinhomiciderates, and they note that, a growing number of regional studies are consistent with the idea that democratization has been associated with rapidly increasing violent crime rates. They also findthathomiciderateshaveincreasedformanydemocraciesinthesecondhalfofthe20th century.theimpactofdemocracyonvariouscrimeseveritiesisexaminedbylin(2007),who finds that less severe crimes are more common in democracies and more severe crimes, like homicides,lesscommon;afindingatoddswiththefindingsoffearon(2011)andlafreeand Tseloni(2006).Finally,HallandMclean(2009)lookatstructuralcharacteristicsofdemocracies in a study more closely related to the analysis presented in this paper. They find that US neoliberalism is more prone to the spread of violence than Western Europe s social democracies,andtheystatethatit seemsquitelikelythatover90percentofushomicidesin known circumstances can be associated in one way or another with economic motivations and/orsituationsandlocalesdefinedprincipallybytheirdisadvantagedpositionsinthesocion economicstructure (HallandMcLean,2009,p.315). 6!!Socialv.ConservativeDemocraciesandHomicideRates

7 3. A SIMPLE RATIONAL CHOICE MODEL OF HOMICIDE AND ECONOMIC GAIN Thewidespreadempiricalresultthathomicideratesandincomeinequalityare positivelycorrelatedsuggeststhattheremightbesomevaliditytothehallandmclean(2009) conjecture.thisideaismodeledbelow.inthemodel,homicideratesarehigherasthe voice ofthemedianvoterislesssuccessfulinamelioratingexcessiveinequality. Consideranindividualwhoneedstomakeachoicebetweenjoiningthe homicide class (i.e.becomingamurderer,forexample,byjoiningagang)andexitingcivilsociety,or remainingapeacefuldemocraticcitizen.amurdereranticipatesaneconomicreward,r,'by joiningthehomicideclass,orequallybycommittingamurder.'uponjoiningthehomicideclass, thisrewardishiddenfromtaxauthoritiesandconstitutestheonlysourceofpotentialincome forthemurderer.thealternativetogainingincomefrombecomingamurdereristogenerate incomeincivilsociety,includingparticipationinthedemocraticallyestablishedincome redistributionprocess;forexample,byparticipatinginaprogressiveincometaxora redistributivesocialpensionsystem.themodelassumesforsimplicitythatnoindividualcan earnincomefromparticipatinginboththehomicideclassandthenonncriminalcivilsociety activity.finally,economicrewardforjoiningthehomicideclassissmallenoughtobeappealing onlytothoseindividualswithinitialincomesmallerthanthemedianincome.thus,itisa relativelypoorindividualwhomayhaveaneconomicincentivetokillanother,buthighincome individualshavenoincentivetocommitmurderforeconomicgain. Lettheprobabilityofescapingpunishmentbep.Iftheassailantgetscaughtand convicted,thesentenceiss.amurderer sutilityfunction,u M,canthenbewrittenas InternationalConflictAnalysisandTransformation.January2012! 7

8 U = pr -(1- p) S (1) M ThemurdereroperatesinademocracythatprovideshimwiththeutilityfunctionU D 'ifhe choosestoremainamemberofcivilsociety,specifically UD 2 τ 2v = ( 1 τ) y+ τ (2) Equation2modelsademocracyinwhicheachcitizenpaystheproportionaltaxrate,τ, andalltaxrevenuesareequallyredistributedacrossmembersofcivilsociety.forsimplicity,the averageincomeissettoone,sothateachindividualgetsanamountτasaredistributiongrant fromthegovernment.theempiricalfactthatincomedistributionsarealwaysskewedtothe right,meaningthaty mode <y median <y mean,isalsoadoptedhere.everyonewithanincomelower (higher)thany mean willhencebeanetreceiver(payer)fromtheredistributionsystem. Thethirdtermontherighthandsideofequation2indicatesthattaxationisassociated withdistortionaryeffects,whosesizedependspositivelyonthetaxrate,τ,andonan exogenousdemocracyindicatorv.thedemocracyindicator,v,isdefinedfor0<v 1andcanbe interpretedasthemedianvoter s voice inthedemocraticprocess.ifv'isone,themedian voteriswellrepresentedbythedemocraticprocess,andthedistortionaryeffectoftaxationis equaltoonehalfofthesquareofthetaxrate.thedistortionaryeffectincreasesasvgoes towardszero.inpractice,factorslikecorruption,politicalmistrust,andcapitalflightcontribute tolowvoice.thepresenceoflowdemocraticvoicecreatesantingovernmentsentiments, spawningmoresociallycostlyarticulationmechanismssuchaspopularprotests,strikesorriots, whichalsomakegovernment(includingthetaxadministrationsystem)lessefficient. 8!!Socialv.ConservativeDemocraciesandHomicideRates

9 Supposethatanindividualmemberofthesocietyisconfrontedwiththechoiceof joiningthehomicideclasswiththeutilityfunction(1)orremainingaregularcitizenwiththe utilityfunction(2).allcitizensknowthatthetaxauthoritywillimplementthemedianvoter s optimumtaxrate,whichshedeterminesbymaximizingequation(2)withrespecttoτ.the optimumproportionaltaxrateτ appliedtoallcitizenswillthenbe ( ) τ * = v 1 y m (3) wherey m isthemedianincome.becausethemedianincomeissmallerthanone(themean income)andvisatmostone,thetaxrateisdefinedbetweenzeroandone.equation(3) capturestheintuitionthattaxrateswillbehighwhendemocraticvoiceishighand/orwhen inequalityishigh.asalludedtoabove,thepoorermajorityisabletotaxtherichinorderto redistributeincome.thetypicalcitizen sutilitywouldbeceteris'paribusmaximizedifv=1,in whichcaseτ*=1<y m. Basedonconditions(1)and(2),acitizenwillchoosetobecomeamurdererif: 2 τ * y+ pr (1- p) S > ( 1 τ *) y+ τ * 2v (4) Equation(4)canbeconvenientlyreNarrangedas: 2 τ * 2v (1 ) > τ * ( 1 ) (5) pr p S y Thelefthandsideofinequality(5)isthenetgainfromjoiningthehomicideclass,therighthand side(rhs)isthenetbenefitfrombeingamemberofcivilsociety.thelefthandsideof(5)is straightforwardtointerpret:anincreaseinthelikelihoodofescapingpunishmentwillincrease InternationalConflictAnalysisandTransformation.January2012! 9

10 homiciderates,ahigherexpectedsentencedecreasesit.thederivativeoftherhsof(5)with respecttotheoptimumtaxrateandseparatelywithrespecttovoice,showsthatanincreaseof eitherfactorincreasesthenetbenefitfromredistributionforcitizenswithincomeslessthan themedianincome.thelikelihoodforbecomingamurdererdecreasesaccordingly. Specifically,takingthederivativeontherhswithrespecttoτ*andusing(3)givesthe followingresult: > 0 if ym > y drhs = -y + ym= 0 if ym = y (6) dτ * < 0 if y < y m Hence,thenetbenefitfromincreasingtheredistributivetaxispositive(negative)forpeople withincomeslower(greater)thanthemedianincome. Similarly,amarginalincreaseinvoicegeneratesthepartialderivative 1+ y > 0 if y < 2 drhs ( 1 y )[ 1 2y+ y ] 1+ y = = 0 if y = dv y < 0 if y > 2 m m m m m (7) Condition(7)impliesthatmorevoice,asexpected,deliversanetgaintothepoor.Butsincey m issmallerthanone,morevoicedoesnotbenefitallcitizenswithincomeslessthantheaverage income.thelowerthemedianincome,andthereforethehighertheinequality,thelowerthe incomethresholdatwhichanincreaseinvoiceincreasesthenetutilityfromredistribution. 10!!Socialv.ConservativeDemocraciesandHomicideRates

11 4. EMPIRICAL ANALYSIS Data$and$Methodology$ Theabovemodelcapturestheideathatdemocracieswithlowerresponsivenesstothe voiceofthemedianvoterarelikelytohaveaweaker socialcontract, whichcontributesto higherlevelsofeconomicinequalityandhigherhomiciderates.todeterminewhetherornot theempiricalevidencesupportsthisnotion,datawascollectedacrosscountriesforpercapita GDP,incomeinequality,polity,taxrevenue,medianage,conflicthistory,andtypeof democracy.countrieswereidentifiedasdemocracies,accordingtohsu s(2008)classification. Shedefinesdemocraciesassuchwhentheyhold fair,multipartyelections. Buildingupona classificationproposedbyespingnandersen(1990),shefurtherlabelsdemocraciesas conservativeif statusdifferentialsarepreserved or modestsocialninsuranceplansdominate andassocial'democracieswhen universalsocialrightsareexpanded (Hsu,2008,pp.8N10). Onlythosedemocracieswithavailabledataforallvariableswereincluded,whichleft thefinaldatasetwith89countries.unlessotherwisenoted,allvariablesare2000n2008 averages.thecountryobservationscanbeplacedintothefollowinggeographicareas:subn SaharanAfrica(SSA),MiddleEastandNorthAfrica(MENA),SouthAsia(SA),EastAsiaandthe Pacific(EAP),CentralandEasternEurope(CEE),LatinAmericaandtheCaribbean(LAC),and WesternEuropeandNorthAmerica(WENA).Theappendixcontainsamoredetaileddata descriptionandthecompletedataset.table1showsthenumberofobservationsbyregionas wellasregionalaveragesforthevariables.itisapparentthattherearesignificantregional differences.amongthevariables,thepercentageofsocialdemocracyperregionisparticularly interesting.thetablerevealsthatsocialdemocraciesfallexclusivelyintothewenacluster.(in InternationalConflictAnalysisandTransformation.January2012! 11

12 fact,whatisnotrevealedisthatallnorthamericancountriesareclassifiedbyhsuas conservativedemocracies,thusmakingsocialdemocraciesinprincipleonlyawestern Europeanphenomenon.) Table1alsoshowsthatincomeinequality,asrepresentedbytheGiniindex,willlikely beastrongpredictorofhomiciderates.thewenaregionalclusterhasboththelowest homiciderateandthehighestshareofsocialdemocracies,whilethelacgrouphasboththe highesthomiciderateandhighestshareofconservativedemocracies.thisobservationforetells thecriticalresultintheempiricalanalysistofollow. Table$1$ Regional$Characteristics$of$Dataset$$ (2000;2008$averages,$unless$indicated$otherwise)$ Region SSA MENA SA EAP CEE LAC WENA Numberofobservations Homiciderates GDPpercapita 1,855 16,826 2,001 4,570 9,714 7,255 33,316 Giniindex Taxrevenues(ofGDP) MedianAge Polity2score 3.2 N Socialdemocracypercentage* Conservativedemocracypercentage* MEPVtotalcivilviolenceandwarconflict score(sumof1980n2008values)percountry *ThesumofthepercentagesofsocialandconservativedemocracydoesnotadduptooneasHsuconsidersother regimeclassificationsaswell,namelydictatorship,militarydictatorship,civilwar,onepartydemocracy, communist,islamicrepublic,andeuropeancolony.subnsaharanafrica(ssa),middleeastandnorthafrica(mena), SouthAsia(SA),EastAsiaandthePacific(EAP),CentralandEasternEurope(CEE),LatinAmericaandtheCaribbean (LAC),andWesternEuropeandNorthAmerica(WENA). Fearon(2011,p.1)says: Homicideratestendtobehigherindemocraciesversusautocracies.Thisistrue bothacrosscountriesandwhenwelookattheeffectoftransitionstodemocracy withincountries.thismaybeacausaleffect itmaybethatauthoritarianstates havemoreaggressive,oppressive,and/orcompetentpoliceforcesthando 12!!Socialv.ConservativeDemocraciesandHomicideRates

13 typicalnewdemocracies,makingforlesscrimeingeneral.itcouldalsobea measurementissue perhapsautocraciesarelessinclinedtoreporthomicides. Anotherconclusionthatmaybewarrantedisthatitisnotdemocracythatfertilizesthe soilforhomicides,butthefailureoftheestablishedsocialcontracttoadequatelysafeguard againstthevagariesofsocioeconomicinequities.hallandmclean(2009)makethisargument inacomparisonbetweentheusandeurope,butarguemostlyheuristically.theempirical evidenceprovidedbelowlendssomesupporttotheirclaims. ThepolityscorevariableisusedbyFearon(2011)andotherstomeasurethedegreeof democracy,butthepolityscoreonlycapturesdemocraticprocedures.democracieswithsimilar polityscorescanhavedifferentestablishedbeliefsandtraditionswithrespecttothenatureof thesocialcontract.ifitisnotdemocraticprocedures,butthenatureofthesocialcontract reflectedineconomicinequalitymeasures,whichexplaindifferencesinhomiciderates,taking intoaccountincomeinequalitymeasuresacrossdemocraciesshouldeliminatethesignificance ofthecoefficientonthepolityscorevariable. Inafirststep,thedatasetislimitedtoobservationswithanaveragepolityscoreof greaterthanfivebetween1960and2008.thiswillpresumablyassurethatallobservationsare consolidated democraciesthathold fair,multipartyelections. Thisapproachisusedto avoidequatingyoungdemocracieswithestablisheddemocracieswheresocioneconomic dynamicsarelessturbulent. Onthisrestrictedsample,thefollowingregressionisrun: ( ) Gini Polity average u = i β β (8) Polity > i i InternationalConflictAnalysisandTransformation.January2012! 13

14 andtheresidualsu i arestored.theresidualscanbeinterpretedasinequalitythatisnot explainedbythelevelofproceduraldemocracy. Wethentestwhethercountrieswithpositive(negative)residualsare,onaverage,more likelyconservative(social)democracies.thisisdonebyregressingtheresidualsfrom(8)on Hsu s(2008)classificationofcountriesassocialandconservativedemocracies(abbreviatedby SDandCD,respectively): u = i γ γ Polity 1SD + i γ2cd + i ε (9) > i TheregressionresultsinTable2illustratesthat,despitethesmallsamplesize,Hsu s(2008) SocialDemocracydummyprovidessomeexplanationforthenegativeresiduals. Table$2$ Explaining$the$Residuals$with$Hsu s$(2008)$social$democracy$dummies$(eq.$9)$ $ DV=u i Model1 Intercept N6.10 (3.75) SocialDemocracyDummy + N7.54* (4.17) N 21 RNsquared 0.15 Standarderrorsinparentheses,***=significantatp<1,**=significantatp<5,*=significantatp<10 + OnlyHsu ssocialdemocracydummywasregressedbecauseofperfectmulticollinearity;allobservationsinthe sampleclassifyaseithersocialorconservativedemocracy(seeappendix). AlthoughHsu s(2008)datasetisavaluablemotivationforthispaper,herclassifications involveadegreeofsubjectivity.analternativeperspectivethatcorrespondscloselytohsu s (2008)classification,butinteractsdirectlywithavailableinequalitydata,istoclassifycountries withpositiveresiduals,u i,inequation(8)asconservative,andthosewithnegativeresidualsas 14!!Socialv.ConservativeDemocraciesandHomicideRates

15 socialdemocracies.twonormalizeddemocracytypeindicesarethuscomputed(onefortheu i >0andonefortheu i <0)asfollows: i min Democracy Type j = 10 umax umin u u withj=[csdx u i <0,CCDX u i >0] (10) Theconsolidatedconservativedemocracyandtheconsolidatedsocialdemocracyindex areabbreviatedccdxandcsdx,respectively.thus,eachindexrangesfromzeroto10.ahigher numbersuggestsagreaterinclinationtowardsafeguardingagainstsocioeconomicinequities (socialdemocraticpolicies)ontheonehand,ortowardfewersafeguards(conservative democraticpolicies)ontheother.inafinalstep,thecsdxandtheccdxwereaddedtomodel specificationssimilartothoseperformedbyfearon(2011).inotherwords,ourregression includesinadditiontoallobservationswithapolityscorebetweenn10and+10,whichcaptures proceduraldemocracycharacteristics,avariableforconsolidatedconservativeandsocial democracyasaproxyforsubstantivedemocracy.tables3to5summarizetheempiricalresults. Table3showsthatmoredemocraticcountrieshave,onaverage,lowerlevelsofincome inequalityaspredictedbyeconomictheory.thescatterplotalsoshowsthatmostlylatin Americancountriescanbefoundabovethetrendline,specificallyColombia(COL),Venezuela (VEN),Jamaica(JAM),andCostaRica(CRI),whileEuropeancountriesaregenerallybelowit. $ InternationalConflictAnalysisandTransformation.January2012! 15

16 Table$3:$ $ Inequality$vs.$Democracy$(Eq.$8)$for$Countries$with$ Average$Polity$Score$>5$between$1960$and$2008$ DV=Gini Model ScatterPlot Constant 67.90*** (10.4) 60.0 Gini vs. Avg. Polity Score ZAF COL Polity N3.36*** (1.15) 48.3 VEN CRI JAM N 21 Gini 36.7 LKA GRC TUR IND ISR USA ITA IRL BEL CHE CAN RNsquared 0.31 AUT DEU 25.0 FIN NOR SWE Polity Standarderrorsinparentheses,***=significantatp<1,**=significantatp<5,*=significantatp<10 BeforeturningtothediscussionofusingdemocracyNclassificationvariables(CCDXandCSDX)in theregressions,alookatthebivariatecorrelationcoefficientsofallvariablesusedinthe subsequentanalysismaybeuseful.table4showsthestrongdirectcorrelationbetween homicideratesandincomeinequality(r=0.66).italsorevealsthecorrelationsbetween homicideratesandcsdxandccdx(r=n0.37andr=0.31).amongtheregionalfixedeffects, LatinAmericaandtheCaribbean(LAC,r=0.50)andWesternEuropeandNorthAmerica(WENA, r=n0.68)areprominent.lookingatthecontrolvariables,thehighnegativecorrelationbetween 16!!Socialv.ConservativeDemocraciesandHomicideRates

17 medianageandhomiciderates(r=n0.52)isnoteworthy.table4alsoalertsustopossible multicollinearityproblems.forexample,gdppercapitaisparticularlystronglycorrelatedwith medianage(r=0.82)andurbanization(r=0.75).lastly,asexpected,thecsdxindexisfairly highlycorrelatedwiththewesterneuropeandnorthamericadummy(r=0.44)whiletheccdx indexiscorrelatedwiththelatinamericaandcaribbeandummy(r=0.33). Table$4$ Correlation$Matrix$ lnhc lnyppp Polity2 Medage urban Gini Taxrev civviol lnhc 1.00 N N0.52 N N N N0.34 N N N0.68 lnyppp N N N N N0.29 N Polity N N0.17 N0.40 N0.21 N medage N N N N0.02 N0.59 N0.07 N0.19 N N urban N N N N N0.45 N Gini 0.66 N N0.53 N N N N0.07 N0.11 N0.05 N N0.41 taxrev N N N N N0.29 N N civviol 0.14 N0.27 N0.04 N0.23 N N N0.04 N N0.11 N0.07 N0.21 CSDX N N N0.08 N0.16 N N0.09 N0.12 N CCDX N N N0.04 N0.04 N0.08 N0.08 N N0.04 SSA 0.32 N0.60 N0.17 N0.59 N N0.02 N0.04 N0.16 N N0.15 N0.15 N0.15 N0.26 N0.25 N0.27 MENA N N0.40 N N N0.04 N0.09 N0.04 N N0.09 N0.09 N0.15 N SA N0.07 N0.29 N0.21 N0.19 N0.45 N0.11 N N0.08 N0.15 N N0.09 N0.15 N0.15 N0.16 EAP 0.00 N0.10 N0.11 N0.05 N0.13 N0.05 N N0.09 N0.08 N0.15 N0.09 N N0.15 N0.15 N0.16 CEE N N N0.11 N0.12 N0.14 N0.26 N0.15 N0.15 N N0.26 N0.28 LAC N N0.11 N0.07 N N0.25 N0.15 N0.15 N0.15 N N0.27 WENA N N N N0.04 N N0.16 N0.16 N0.28 N Table5presentsinitialregressionresults.ModelIisFearon s(2011)simplest specification,usingpercapitaincomeandpolityasindependentvariables.modelirevealsthe unexpectedpositivesignforpolity.inmodelii,thecsdxandccdxindiceswereadded,bothof CSDX CCDX SSA MENA SA EAP CEE LAC WENA InternationalConflictAnalysisandTransformation.January2012! 17

18 whicharesignificantandcarrytheexpectedsign.thepolityvariableisstillsignificantwitha positivesigninthisspecification.modeliiiaddssocioeconomiccontrolvariables.thecsdx variableisnownolongersignificant,butstillcarriestheexpectedsign.thelossofsignificanceis likelyattributedtomulticollinearity.medianageandcsdxhaveacorrelationcoefficientof r=0.38.bothpolityandccdxremainsignificantwiththeexpectedsign.modelivadditionally includesregionaldummies.inmodeliv,thepolityvariableisnolongersignificantwhileccdx stillis.despiteexpectedmulticollinearitywithccdx,thelatinamericaandcaribbeandummy isalsosignificantwiththeexpectedpositivesign. SimilartoFearon(2011),modelsVtoVIIIinTable6reproducespecificationsItoIVwith theginicoefficientincludedineachmodel.theginivariableissignificantwithapositivesignin allmodels.theadditionoftheginicoefficientinmodelvleavespolitysignificantandpositive. AsopposedtomodelII,however,theadditionofGinitoModelVIrendersCSDXinsignificant, althoughtheexpectedsignprevails.thisislikelyattributabletomulticollinearitybetweengini andcsdx.thesignificanceandsignofpolityandccdxisnotaffectedbygini.polityandccdx arealsosignificantinmodelviiwhilethecontrolvariablemedianageisnolongersignificant. Lastly,theonlysignificantvariablesinmodelVIIIareGini,CCDX,andagainmedianage(which maybeaspuriousresultgiventhefactthatitisnotsignificantinmodelvii). 18!!Socialv.ConservativeDemocraciesandHomicideRates

19 ModelI DV=lnHC (Fearonspecif. Tbl.3, ModelI) Constant N9.190 (4.757)* lnyppp (1.118)*** lnyppp2 N0.209 (0.065)*** Polity (0.020)*** Table$5:$ Homicides$and$Kind$of$Democracy$(N=89)$ N5.414 (4.484) (1.059)** N0.153 (0.062)** (0.019)*** ModelII ModelIII ModelIV N6.316 (4.350) (1.037)** N0.159 (0.060)*** (0.019)*** Gini CSDX N0.116 (0.063)* CCDX (0.055)*** N0.070 (0.064) (0.055)*** MedAge N0.061 (0.021)*** Urban (0.007) CivViol N0.001 (0.004) TaxRev N0.003 (0.016) (4.844) (1.155) N0.002 (0.068) (0.020) N0.013 (0.058) (0.047)** N0.049 (0.026)* (0.007) (0.004) (0.014) WENA N0.622 (0.596) LAC (0.494)** CEE (0.428) EAP (0.442) MENA N0.529 (0.533) SSA (0.413)* RNsquared Standarderrorsinparentheses,***=significantatp<1,**=significantatp<5,*=significantatp<10 InternationalConflictAnalysisandTransformation.January2012! 19

20 DV=lnHC Constant lnyppp (0.985) lnyppp 2 Table$6:$ Homicides$and$Kind$of$Democracy$(N=89)$ ModelV (Fearonspecif., Tbl.3, ModelII) N3.901 (4.032) N0.077 (0.058) Polity (0.017)** Gini (0.011)*** N2.654 (3.975) (0.967) N0.068 (0.057) (0.017)** (0.012)*** CSDX N0.055 (0.056) CCDX (0.051)** 20!!Socialv.ConservativeDemocraciesandHomicideRates ModelVI ModelVII ModelVIII N2.834 (4.189) (1.064) N0.077 (0.061) (0.019)* (0.016)*** N0.043 (0.060) (0.053)** MedAge N0.008 (0.025) Urban (0.007) CivViol (0.004) TaxRev N0.003 (0.015) (4.826) N0.658 (1.183) (0.069) (0.020) (0.018)* (0.059) (0.046)** N0.044 (0.026)** (0.007) (0.004) (0.014) WENA N0.477 (0.589) LAC (0.532) CEE (0.456)* EAP (0.419) MENA N0.628 (0.525) SSA (0.432) RNsquared Standarderrorsinparentheses,***=significantatp<1,**=significantatp<5,*=significantatp<10

21 Fearon s(2011)findingofapositiverelationshipbetweenthepolityscoreandhomicide ratesis,atfirstsight,remarkable.however,theresultshereshowthatitispossibleto eliminatepolity ssignificanceafterincorporatingthenatureofdemocracyandcontrollingfor socioeconomicvariablesandregionalfixedeffects.themostrobustvariablesforexplaining homicideratesareinequalityandtheccdxindex.forademocracywithrelativelyhighincome inequality,ccdxisameasureofthedistancebetweenitsactualanditsexpectedlevelof incomeinequality.totheextentthatthepolityvariable,whichisameasureofthelevelof democracy,hashadexplanatorypowerinempiricalstudiesofhomiciderates,itislikelythe resultofafailuretosufficientlytakeintoaccountthedifferencesinincomeinequalityacross otherwiseseeminglysimilardemocracies. TheCSDXvariableperformedweakly,butthisismostlikelytheresultof multicollinearity.theproblemofmulticollinearityisinfactomnipresentinthisempirical analysis.futureresearchmaythereforefocusonabetterunderstandingofthestructural interactionamongtherighthandsidevariables.despitethesetechnicalconcerns,theresults lendsupporttotheconclusionthatitisnotdemocracythatleadstomorehomicides,butrather thefailureofsomedemocraciestoprovideforequitablesocialdevelopment.ultimately,the consolidatedconservativedemocracyindexishighlyrobustandsignificantinexplaining homiciderates. 5. CONCLUSIONS TherelationshipbetweendemocracyandhigherhomicideratesreportedbyFearon (2011)ispuzzling.Toresolvethispuzzle,themainargumentofthispaperisthatamore InternationalConflictAnalysisandTransformation.January2012! 21

22 differentiatedlookatdemocracyisnecessary.weproposethatitisnotthelevelofprocedural democracyitselfthatisimportant,whichiswhatthepolityivscoreemphasizes,butratherthe moresubstantivevariablethatreflectstheextenttowhichdemocraticpoliciesserveasa meansforequitablesocialdevelopment. WediscussthisideawithasimplerationalNbehaviortheoreticalmodelandfroman empiricalperspective.themodelsuggestsatransmissionmechanismfromlessvoicetoless redistribution,whichinturnincreasesapoorcitizen spropensitytoseekeconomicgain throughviolence.empirically,wecapturethisideabydistinguishingbetweenestablished democracieswithlowandhighinequality.welabelthesetwodemocracyformsassocialand conservativedemocracies,respectively.theempiricalresultssuggestthathigherhomicide ratesarestatisticallysignificantlylimitedtoconservativedemocracies,whicharedemocracies inwhichincomeinequalityisrelativelyhigh.theseconservativedemocraciesthatfailtocreate anequitablesocioeconomicsystemhaveagreaterlikelihoodforviolencethansocial democraciesandauthoritarianregimes.!! 22!!Socialv.ConservativeDemocraciesandHomicideRates

23 REFERENCES Boix,C.(2003),DemocracyandRedistribution,Cambridge(CambridgeUniversityPress). Blau,J.R.andBlauP.M.(1982),TheCostofInequality:MetropolitanStructureandViolent Crime,AmericanSociologicalReview,Vol.47(February),pp.114N129. Brush,J.(2007),DoesIncomeInequalityLeadtoMoreCrime?AComparisonofCrossNSectional andtimenseriesanalysesofunitedstatescounties,economicsletters,vol.96,pp.264n 268. EspingNAndersen,G.(1990),TheThreeWorldsofWelfareCapitalism,Princeton(Princeton UniversityPress). Fearon,J.D.(2011),HomicideData,ThirdRevision,BackgroundpaperpreparedfortheWDR 2011Team, Goring,C.(1913),TheEnglishConvict,AStatisticalStudy,London(DarlingandSon,Ltd.) Hall,S.andMcLean,C.(2009),ATaleofTwoCapitalisms:PreliminarySpatialandHistorical ComparisonsofHomicideRatesinWesternEuropeandtheUSA,Theoretical Criminology,Vol.13(3),pp.313N339. Hsu,S.(2008),TheEffectofPoliticalregimesonInequality,1963N2002,UniversityofTexas InequalityProject,UTIPWorkingPaperNo.53, Jalil,H.H.andIqbal,M.M.(2010),UrbanizationandCrime:ACaseStudyofPakistan,27 th AnnualGeneralMeetingandConference,PakistanSocietyofDevelopmentEconomists, LaFree,G.andTseloni,A.(2006),DemocracyandCrime:AMultilevelAnalysisofHomicide TrendsinFortyNFourCountries,1950N2000,AnnalsoftheAmericanAcademyofPolitical andsocialscience,vol.605(may),democracy,crime,andjustice,pp.26n49. Levitt,S.(2004),Understandingwhycrimefellinthe1990s:fourfactorsthatexplainthedecline andsixthatdonot,journalofeconomicperspectives18,pp.163n190. Lin,M.(2007),Doesdemocracyincreasecrime?TheEvidencefromInternationalData,Journal ofcomparativeeconomics,vol.35,pp Lodhi,A.Q.andTilly,C.(1973),Urbanization,Crime,andCollectiveViolencein19thNCentury France,AmericanJournalofSociology,Vol.79,No.2(September),pp.296N318. InternationalConflictAnalysisandTransformation.January2012! 23

24 Neison,F.G.P.(1846),TheRateofMortalityandtheLawofSickness,FromOriginaland ExtensiveDataProcuredfromFriendlySocieties,2 nd Edition,London(Simpkin,Marshall &Co.). Neumayer,E.(2003),GoodPolicyCanLowerViolentCrime:EvidencefromaCrossNNational PanelofHomicideRates,1980N97,JournalofPeaceResearch,Vol.40,No.6 (November),pp.619N640. Pampel,F.C.andGartner,R.(1995),AgeStructure,SocioNPoliticalInstitutions,andNational HomicideRates,EuropeanSociologicalReview,Vol.11,No.3(December),pp.243N260. Pridemore,W.A.andKim,S.(2006),DemocratizationandPoliticalChangeasThreatsto CollectiveSentiments:TestingDurkheiminRussia,AnnalsoftheAmericanAcademyof PoliticalandSocialScience,Vol.605,Democracy,Crime,andJustice(May),pp.82N103. Data$Sources:$ CenterforSystemicPeace(online),MajorEpisodesofPoliticalViolence,1946N2008, Hsu,S.(online),PoliticalRegimeDataset,UniversityofTexasInequalityProject(UTIP), Marshall,M.G.,Jaggers,K.,andGurr,T.R.(online),PolityIVProject:PoliticalRegime CharacteristicsandTransitions,1800N2010, UnitedNations(online),PopulationDivisionoftheDepartmentofEconomicandSocialAffairs oftheunitednationssecretariat,worldpopulationprospects:the2008revision, WilkinsonR.andPickett,K.,(2009)TheLevel:WhyGreaterEqualityMakesSocietiesStronger. BloomsburyPress,NewYork. WorldBank(online),HomicideRateDataset1995N2008, WorldBankDevelopmentIndicatorDatabase(online), WorldHealthOrganization(online),HistoricalTrendSeriesDataonDeathsfromRoadTraffic Crashes,SuicideandHomicide, 24!!Socialv.ConservativeDemocraciesandHomicideRates

25 InternationalConflictAnalysisandTransformation.January2012! 25 APPENDIX Variables$and$Sources$ Variable Transformation Abbreviation Source LogHomiciderate per100,000 population Averageof availabledatafor the2000n2008 period,naturallog lnhc HomicideRateDataset1995N2008(WorldBank, countriesfromsubnsaharaafrica,latinamerica andthecaribbean,eastasiaandthepacific, SouthAsia,CentralandEasternEurope (transitionaleconomies),andthemiddleeastand NorthAfrica.DataforWesternEuropeanand NorthAmericancountrieswastakenfromthe WorldHealthOrganization shistoricaltrendseries dataondeathsfromroadtrafficcrashes,suicide andhomicide( LogGDPpercapita, PPP (2005dollars) Averageof availabledatafor the2000n2008 period,naturallog lnyppp 2011WorldBankDevelopmentIndicator Database. SquareofLogGDP GDPPPPSquared lnyppp2 TaxRevenueas aofgdp Averageof availabledatafor the2000n2008 period Taxrev Urbanpopulation (oftotal) Urban Giniindex Gini Polity2score Averageof availabledatafor the2000n2008 period Polity Marshall,Jaggers,andGurr,PolityIVProject: PoliticalRegimeCharacteristicsandTransitions, 1800N2010,

26 Dummyforsocial andconservative democracy Socialdemocracy (SD)whenall observations between2000and 2005werelabeled assuch.similarfor conservative democracy(cd). SDandCD Hsu,S.(online),PoliticalRegimeDataset, UniversityofTexasInequalityProject(UTIP), Medianage In2009 Medage PopulationDivisionoftheDepartmentof EconomicandSocialAffairsoftheUnitedNations Secretariat(2009).WorldPopulationProspects: The2008Revision, Civiltotalscore Sumofallciviland ethnicpolitical violenceandwar scoresbetween 1981and2008. Civtot CenterforSystemicPeace,MajorEpisodesof PoliticalViolence,1946N2008, 26!!Socialv.ConservativeDemocraciesandHomicideRates

27 Dataset& Country! Code! Region! Homicide! yppp! Polity! medage! urban! SD!(Hsu)! CD!(Hsu)! Gini! taxrev! civviol! CSDX! CCDX! Albania* ALB* CEE* 5.7* * 7.4* 29.7* 44.2* 0* 1* 31.7* 16.82* 2* 0.00* 0.00* Armenia* ARM* CEE* 2.6* * 5.0* 31.8* 64.4* 0* 1* 33.36* 14.94* 0* 0.00* 0.00* Azerbaijan* AZE* CEE* 2.5* * J7.0* 28.2* 51.5* 0* 0* 35.11* 16.74* 21* 0.00* 0.00* Bosnia*and*Herz.* BIH* CEE* 2.1* * 0.0* 38.9* 45.3* 0* 1* 33.34* 20.9* 24* 0.00* 0.00* Bulgaria* BGR* CEE* 3.1* * 8.9* 41.5* 70.0* 0* 1* 36.3* 20.16* 0* 0.00* 0.00* Croatia* HRV* CEE* 3.2* * 8.4* 41.3* 56.4* 0* 1* 31.28* 20.76* 12* 0.00* 0.00* Estonia* EST* CEE* 9.5* * 9.0* 39.5* 69.4* 0* 0* 36.5* 15.98* 0* 0.00* 0.00* Hungary* HUN* CEE* 2.4* * 10.0* 39.7* 66.0* 0* 1* 28.46* 21.28* 0* 0.00* 0.00* Kazakhstan* KAZ* CEE* 13.4* * J5.6* 29.3* 57.0* 0* 0* 32.75* 12.88* 0* 0.00* 0.00* Kyrgyz*Republic* KGZ* CEE* 8.1* * J0.2* 24.8* 35.8* 0* 0* 32.68* 14.26* 0* 0.00* 0.00* Latvia* LVA* CEE* 8.4* * 8.0* 39.9* 68.1* 0* 1* 35.92* 14.54* 0* 0.00* 0.00* Lithuania* LTU* CEE* 9.5* * 10.0* 39.4* 66.8* 0* 1* 33.99* 16.73* 0* 0.00* 0.00* Macedonia* MKD* CEE* 3.5* * 8.3* 35.7* 64.9* 0* 1* 39.83* 19.78* 0* 0.00* 0.00* Moldova* MDA* CEE* 7.9* * 7.9* 35.0* 43.1* 0* 1* 36.81* 16.65* 7* 0.00* 0.00* Poland* POL* CEE* 1.9* * 9.8* 37.9* 61.5* 0* 1* 33.79* 17.04* 0* 0.00* 0.00* Russia* RUS* CEE* 21.6* * 5.6* 37.9* 73.1* 0* 1* 39.75* 14.85* 44* 0.00* 0.00* Slovenia* SVN* CEE* 1.6* * 10.0* 41.4* 49.7* 0* 1* 30.15* 20.28* 0* 0.00* 0.00* Turkey* TUR* CEE* 3.8* * 7.0* 28.0* 66.7* 0* 1* 41.71* 19* 35* 2.41* 0.00* Ukraine* UKR* CEE* 8.4* * 6.3* 39.4* 67.6* 0* 1* 28* 15.04* 0* 0.00* 0.00* China* CHN* EAP* 1.9* * J7.0* 33.8* 39.5* 0* 0* 41.53* 8.69* 19* 0.00* 0.00* Indonesia* IDN* EAP* 4.0* * 7.1* 27.9* 46.8* 0* 1* 38.5* 12.29* 45* 0.00* 0.00* Lao*PDR* LAO* EAP* 5.6* * J7.0* 20.4* 26.4* 0* 0* 34.69* 11.24* 2* 0.00* 0.00* Malaysia* MYS* EAP* 3.9* * 3.3* 26.0* 66.4* 0* 1* 37.91* 15.57* 0* 0.00* 0.00* Mongolia* MNG* EAP* 12.6* * 10.0* 25.9* 56.8* 0* 1* 34.13* 20.65* 0* 0.00* 0.00* Philippines* PHL* EAP* 8.9* * 8.0* 23.0* 61.8* 0* 1* 44.87* 13.37* 90* 0.00* 0.00* Thailand* THA* EAP* 7.3* * 5.8* 32.8* 32.1* 0* 1* 42.52* 16.32* 6* 0.00* 0.00* Argentina* ARG* LAC* 7.1* * 8.0* 30.2* 91.1* 0* 1* 50.66* 12.18* 0* 0.00* 0.00* Bolivia* BOL* LAC* 16.2* * 8.3* 21.7* 63.7* 0* 1* 58.56* 15.19* 0* 0.00* 0.00* Brazil* BRA* LAC* 25.8* * 8.0* 28.6* 83.5* 0* 1* 56.96* 15.84* 0* 0.00* 0.00* Chile* CHL* LAC* 5.0* * 9.3* 31.8* 87.2* 0* 1* 54.09* 18.2* 0* 0.00* 0.00* International*Conflict*Analysis*and*Transformation.*January*2012!!27*

28 Appendix*1:*Dataset*(Contd.)* Country! Code! Region! Homicide! yppp! Polity! medage! urban! SD!(Hsu)! CD!(Hsu)! Gini! taxrev! civviol! CSDX! CCDX! Colombia* COL* LAC* 52.5* * 7.0* 26.5* 73.3* 0* 1* 58.27* 11.66* 76* 0.00* 10.00* Costa*Rica* CRI* LAC* 7.3* * 10.0* 27.8* 61.2* 0* 1* 48.49* 15.81* 0* 0.00* 9.10* Dominican*Rep.* DOM* LAC* 15.4* * 8.0* 24.8* 65.8* 0* 1* 50.86* 14.62* 0* 0.00* 0.00* El*Salvador* SLV* LAC* 44.3* * 7.0* 23.7* 59.5* 0* 1* 50.05* 12.38* 18* 0.00* 0.00* Guatemala* GTM* LAC* 36.0* * 8.0* 18.7* 46.8* 0* 1* 54.67* 11.38* 35* 0.00* 0.00* Honduras* HND* LAC* 42.6* * 7.0* 20.7* 46.1* 0* 1* 55.89* 15.06* 1* 0.00* 0.00* Jamaica* JAM* LAC* 46.8* * 9.0* 26.1* 52.5* 0* 1* 46.93* 25.74* 0* 0.00* 7.39* Mexico* MEX* LAC* 21.0* * 8.0* 27.2* 76.0* 0* 1* 49.49* 11.66* 10* 0.00* 0.00* Nicaragua* NIC* LAC* 12.0* * 8.2* 21.7* 55.7* 0* 1* 51.32* 15.86* 3* 0.00* 0.00* Panama* PAN* LAC* 11.0* * 9.0* 27.1* 69.7* 0* 1* 55.87* 9.73* 0* 0.00* 0.00* Paraguay* PRY* LAC* 16.2* * 7.7* 22.8* 57.8* 0* 1* 54.27* 11.22* 0* 0.00* 0.00* Peru* PER* LAC* 7.4* * 8.6* 25.3* 71.0* 0* 0* 51.94* 13.58* 24* 0.00* 0.00* Uruguay* URY* LAC* 5.1* * 10.0* 33.5* 91.8* 0* 1* 45.43* 17.03* 0* 0.00* 0.00* Venezuela* VEN* LAC* 39.0* * 5.8* 25.8* 91.7* 0* 1* 46.44* 12.67* 0* 0.00* 2.81* Egypt* EGY* MENA* 0.6* * J4.7* 23.6* 42.6* 0* 0* 32.45* 14.45* 8* 0.00* 0.00* Iran* IRN* MENA* 3.1* * J2.0* 26.3* 66.4* 0* 0* 38.28* 6.69* 16* 0.00* 0.00* Israel* ISR* MENA* 3.9* * 10.0* 29.5* 91.6* 0* 1* 39.2* 27.31* 38* 0.00* 1.75* Jordan* JOR* MENA* 2.4* * J2.2* 22.4* 78.3* 0* 0* 38.3* 20.37* 0* 0.00* 0.00* Morocco* MAR* MENA* 0.6* * J6.0* 25.8* 54.7* 0* 0* 40.76* 22.36* 0* 0.00* 0.00* Qatar* QAT* MENA* 0.9* * J10.0* 30.1* 95.3* 0* 0* 41.1* 20.74* 0* 0.00* 0.00* Tunisia* TUN* MENA* 1.4* * J3.8* 28.6* 64.9* 0* 0* 40.81* 21.18* 0* 0.00* 0.00* Afghanistan* AFG* SA* 3.4* 795.5* J0.8* 16.8* 22.6* 0* 0* 29.4* 5.6* 84* 0.00* 0.00* Bangladesh* BGD* SA* 3.0* * 3.3* 24.1* 25.3* 0* 1* 30.87* 8.09* 6* 0.00* 0.00* Bhutan* BTN* SA* 2.4* * J7.2* 23.8* 29.9* 0* 0* 46.74* 8.91* 6* 0.00* 0.00* India* IND* SA* 3.6* * 9.0* 24.7* 28.6* 0* 1* 36.8* 9.85* 139* 1.39* 0.00* Nepal* NPL* SA* 4.9* 949.0* 0.7* 21.3* 15.3* 0* 0* 47.3* 9.17* 22* 0.00* 0.00* Pakistan* PAK* SA* 5.8* * J3.3* 21.0* 34.6* 0* 0* 31.44* 10.03* 35* 0.00* 0.00* Sri*Lanka* LKA* SA* 7.7* * 5.6* 30.3* 15.3* 0* 1* 40.66* 13.85* 99* 4.92* 0.00* Benin* BEN* SSA* 11.4* * 6.3* 18.3* 39.7* 0* 1* 38.62* 15.98* 0* 0.00* 0.00* Burkina*Faso* BFA* SSA* 3.4* * J0.3* 16.7* 18.0* 0* 0* 39.6* 11.69* 0* 0.00* 0.00* 28!!Social*v.*Conservative*Democracies*and*Homicide*Rates.*!

29 Appendix*1:*Dataset*(Contd.)* Country! Code! Region! Homicide! yppp! Polity! medage! urban! SD!(Hsu)! CD!(Hsu)! Gini! taxrev! civviol! CSDX! CCDX! Cape*Verde* CPV* SSA* 6.5* * 9.8* 20.9* 56.6* 0* 1* 50.4* 22.97* 0* 0.00* 0.00* Cen.*Afr.*Rep.* CAF* SSA* 26.3* 693.0* 1.0* 19.4* 38.0* 0* 0* 43.57* 6.21* 9* 0.00* 0.00* Congo,*Rep.* COG* SSA* 17.5* * J4.3* 19.4* 59.8* 0* 0* 47.32* 7.72* 12* 0.00* 0.00* Ghana* GHA* SSA* 4.1* * 6.7* 20.4* 47.0* 0* 1* 42.76* 17.1* 1* 0.00* 0.00* Kenya* KEN* SSA* 6.1* * 5.6* 18.3* 20.6* 0* 1* 47.68* 17.52* 9* 0.00* 0.00* Lesotho* LSO* SSA* 25.2* * 7.3* 19.6* 22.7* 0* 1* 52.5* 47.97* 0* 0.00* 0.00* Liberia* LBR* SSA* 24.8* 381.0* 3.0* 18.4* 57.3* 0* 0* 38.16* 0.27* 36* 0.00* 0.00* Madagascar* MDG* SSA* 10.8* 893.3* 7.0* 18.3* 28.3* 0* 1* 47.36* 10.53* 0* 0.00* 0.00* Nigeria* NGA* SSA* 3.6* * 4.0* 18.5* 45.5* 0* 1* 42.93* 0.2* 32* 0.00* 0.00* Senegal* SEN* SSA* 5.0* * 7.8* 17.9* 41.4* 0* 1* 40.22* 16.12* 8* 0.00* 0.00* Sierra*Leone* SLE* SSA* 12.8* 608.2* 4.6* 18.3* 36.6* 0* 0* 42.52* 10.95* 33* 0.00* 0.00* South*Africa* ZAF* SSA* 45.2* * 9.0* 24.7* 58.8* 0* 1* 57.77* 26.02* 21* 0.00* 5.76* Swaziland* SWZ* SSA* 31.6* * J9.0* 19.1* 24.0* 0* 0* 50.68* 25.5* 0* 0.00* 0.00* Togo* TGO* SSA* 12.1* 778.1* J2.9* 19.6* 39.2* 0* 0* 34.41* 15.13* 0* 0.00* 0.00* Uganda* UGA* SSA* 10.6* 893.1* J2.7* 15.5* 12.5* 0* 0* 44.2* 11.49* 34* 0.00* 0.00* Zambia* ZMB* SSA* 11.6* * 4.8* 16.8* 35.0* 0* 1* 46.41* 17.28* 0* 0.00* 0.00* Austria* AUT* WENA* 0.8* * 10.0* 41.4* 66.4* 1* 0* 29.15* 20.57* 0* 3.60* 0.00* Belgium* BEL* WENA* 1.6* * 9.6* 41.1* 97.3* 1* 0* 32.97* 26* 0* 0.78* 0.00* Canada* CAN* WENA* 1.5* * 10.0* 39.6* 80.0* 0* 1* 32.56* 13.94* 0* 0.89* 0.00* Finland* FIN* WENA* 2.2* * 10.0* 41.8* 62.2* 0* 1* 26.88* 22.57* 0* 5.40* 0.00* Germany* DEU* WENA* 0.7* * 10.0* 43.9* 73.4* 0* 1* 28.31* 11.4* 0* 4.27* 0.00* Greece* GRC* WENA* 0.9* * 10.0* 41.3* 60.3* 0* 1* 34.27* 20.71* 0* 10.00* 0.00* Ireland* IRL* WENA* 0.9* * 10.0* 34.3* 60.2* 0* 1* 34.28* 24.67* 0* 0.00* 0.00* Italy* ITA* WENA* 0.9* * 10.0* 43.0* 67.6* 0* 1* 36.03* 22.36* 0* 0.00* 1.12* Norway* NOR* WENA* 0.9* * 10.0* 38.7* 76.9* 1* 0* 25.79* 28.11* 0* 6.27* 0.00* Spain* ESP* WENA* 1.0* * 10.0* 39.9* 76.7* 0* 1* 34.66* 13.23* 0* 0.00* 0.00* Sweden* SWE* WENA* 1.0* * 10.0* 40.7* 84.3* 1* 0* 25* 21.93* 0* 6.90* 0.00* Switzerland* CHE* WENA* 0.8* * 10.0* 41.6* 73.3* 0* 1* 33.68* 10.29* 0* 0.00* 0.00* United*States* USA* WENA* 6.5* * 10.0* 36.5* 80.4* 0* 1* 40.81* 10.99* 0* 0.00* 4.18* * * * International*Conflict*Analysis*and*Transformation.*January*2012.*!29*