PLEASEDONOTCITEWITHOUTPERMISSION

NGOsinKenya:IncreaseStateLegitimacyorUnderminePopularSupport?

PresentedattheWorkshoponPoliticalTheory&PolicyAnalysisColloquiumSeriesIndianaUniversity

Monday,September20,2010

JenniferN.BrassSchoolofPublicandEnvironmentalAffairs

brassj@indiana.edu

“TheproblemactuallyiswhenNGOsarenotthere.Thewananchidon’tcarethatgovernmentisnottherewhenNGOsarethere.Aslongasoneisthere,allisokay.Butifnonearethere,thentheygetangryatgovernment.”(CivilServant,Machakos,2008)“IfIdon’tgiveservices,Iwon’tbeelected.”(CountyCouncilMember,MasakuCounty,2008)

IntroductionInthepasttwodecades,thepositionofnon‐governmentalorganizations(NGOs1)in

developingcountriesaroundtheworldhasshiftedfromthatofminorandlittle‐discussedplayersfocusingonthewelfareofthepoortomajor,centralactorsontheworldstageofdevelopment,receiving,insomecases,moredonorfundsthantheirstatecounterparts(Chege1999).Insub‐SaharanAfrica,thisshiftaroseinpartfromdonorfrustrationwithopaqueandinefficientstate‐basedsystemsfordevelopment,whichspawnedaninterestinaccountabilityandgovernancemechanismsinvolvingnon‐stateactors,includingNGOs.NGOshaveincreasinglybeenseenbydonorsasmoreefficient,effective,flexibleandinnovativethangovernments,tobeother‐orientedandideologicallycommittedtodemocracyandparticipatorypro‐poordevelopment,andtobemoreaccountableandtransparentthanthegovernment(Bratton1989,Fowler1991,Owitietal2004).ThenumbersofNGOsthroughoutthedevelopingworld,andAfricainparticular,haveskyrocketed.Kenya,forexample,haswitnessedavirtualexplosioninthenumberNGOsinthecountry:in1974therewereonly125NGOsinKenya;by2006,over4,200hadbeenregisteredwiththegovernment(RepublicofKenya2006). ThispaperanalyzestheimpactinthegrowthofNGOsonstate‐societyrelations.Itbeginswithrecognitionthat–accordingtomanytheoriesofAfricanpolitics–African

1TheKenyangovernmentdefinesanNGOas:“Aprivatevoluntarygroupingofindividualsorassociationsnotoperatedforprofitorforothercommercialpurposesbutwhichhaveorganizedthemselvesnationallyorinternationallyforthebenefitofthepublicatlargeandforthepromotionofsocialwelfaredevelopment,charityorresearchintheareasinclusivebutnotrestrictedtohealthrelief,agricultural,education,industryandsupplyofamenitiesandservices”(RepublicofKenya1992).TheWorldBankusesasimilarlybroaddefinition.

statespredicatetheirlegitimacyonthepromiseofdistributingservicestothepopulace(Jackson&Rosberg1984,Young1988,Fowler1991,Kanyinga1996,Schatzberg2001).Itthenaskswhathappenswhenanewsetofactors–non‐governmentalorganizations(NGOs)–stepsbetweenstateandsocietytodeliverhealthcare,education,water,etc.Doestheintroductionoflargelyforeign‐fundedorganizationsinterrupttheflowoflegitimacyfromcitizentostate?SeveralprominentscholarswritingonNGOsandserviceprovisioninAfricasuggestthatthisoccurs(Bratton1989,Rothschild&Chazan1998,Martin2002,Sandberg1994).Intheliteratureonthesubject,governmentlegitimacyisunderminedasNGOscallfordemocracy,accountabilityandruleoflaw,therebydevelopingactivecivilsocietyinoppositiontothestate.AndithappensasAfricanscompareresponsive,generous,transparentandparticipatoryNGOstotheirlesseffectivegovernment. ThispaperanalyzestherelationshipbetweenNGOprovisionofserviceandgovernmentlegitimacyinKenya.ItexamineswhethertheKenyangovernmentmustprovideservicesitselfforlegitimacytoaccruetoit.WhenNGOsprovidecheapservices,distributedonorresourcesgenerously,andfosterademocraticspiritofparticipation,istheresultlowerpopularperceptionsofgovernment?

Therearefourprinciplefindingsinthispaper.First,mostindividualshavelimiteddirectcontactwithNGOs.Second,atthesametime,NGOsaregenerallyregardedfavorably–oftenbettersothantheirgovernmentcounterparts.Third,contactwithandappreciationforNGOsdoesnottranslateintodistasteforgovernment.ThosesoundingthealarmthatNGOsareoverwhelminggovernmentlegitimacyinweakstatesarecryingwolf–ifperhapsunintentionally.InKenya,thereislittleevidenceofNGOsreplacingthegovernmentasthenewlegitimateauthority.Indeed,inmyfindings,NGOpresenceeitherbolstersthewaypeopleviewtheirstate,orseeminglyhaslittleeffect.2Finally,whileNGOsappearnottohaveamajorimpactonlegitimacyingeneral,thereissomeevidenceofasignificantdifferenceinresponsesfromurbanversusruraldwellers,andonmoreversusless‐educatedindividuals.3

Thispaperisorganizedasfollows:thefirstsectiondescribesthedatacollectedandused.Next,thepaperexaminesdescriptivestatisticalfindingsfromthedata,followedbymodelsthatcontrolforavarietyoffactorsandprovidechecksofrobustness.Drawingoninterviewsconductedwhilethesurveywastakingplace,thefollowingsectionprovidesseveralexplanationsforthestatisticalresults,highlightingthemostconvincingexplanations.Thepaperconcludeswithabriefoutlineoffactorsthatdoaffectlegitimacy.

TheDataSurveyInstrumentandInterviews

Thispaperdrawsondatafromanoriginalsurveyof501individualstakenatthehouseholdlevelinthreedistricts(thecapital,Nairobi,MbeereandMachakos),inAugust2Ina2009APSRarticle,ChrisBlattmanalsofindsthatNGOsdonotaffectpoliticalparticipationlevelsamongex‐combatantsinnorthernUganda.3 Inacompanionpaperusingaseparatedatasetthatmeasureschangeovertime,IfindthatthelegitimacyoftheKenyanstatehasactuallynotdecreasedinastatisticallymeaningfulwayinthepost‐independencetimeperiod,oncedemographicandpoliticalfactorsaretakenintoconsideration.Moreover,NGOs’relationshipwithlegitimacyisusuallypositive;whereitisnegative,theimpactofNGOsislessstrongthanareotherfactors.

andSeptember2008.4Randomsamplesof150respondentsweretakeninthetwofulldistricts,MbeereandMachakos.Samplingwasdoneinclustersinurban,ruralandperi‐ruralareas(so‐calledmarkettowns),inproportiontothepopulationofeachlocationorsub‐locationaccordingtothemostrecentgovernmentcensus.Notmorethantenrespondentscamefromanyonecluster.5Bothdistrict‐widesampleswerethereforemeanttoberepresentativeofthedistrictasawhole.

TocompareresponsesoftheaverageKenyan(fromMachakosandMbeere)tourbanKenyans,arandomsampleof101respondentsinNairobi“middleclass”neighborhoodswasalsoconducted.6Theseneighborhoodswerecarefullyselectedforbeinglong‐standingcommunitiesinhabitedlargelybyeducatedandformal‐sectoremployed–butnotwealthyor“elite”–Kenyans.Forthemostpart,theywereneighborhoodswhereonemightexpectcivilservantsandtheirfamiliestolive.Arandomsampleof100respondentsinMachakosTown,thedistrictheadquartersofMachakos(population150,000),wasalsotakeninordertocompareurbanvs.ruralviewpointsoutsideofthecapitalcity.7

Tosupplementtheresponsesfrommysurveyinstruments,Iconductedin‐depthone‐on‐oneinterviewswithregularKenyans,NGOworkers,civilservantsandpoliticians.Theseinterviewshelpinterpretthestatisticalfindings.RespondentsoftenprovidestoriesofmechanismsbywhichNGOsandlegitimacyarerelated.

Variables

ContactwithNGOsismeasuredthroughtwosetsofquestionsintheoriginalsurvey.First,respondentsareaskedwheretheygotoreceivebasicservices–education,healthcare,HIV‐specifichealthcare,drinkingwaterandsecurity.Second,otherquestionsdirectlyaddressperceptionsofthenumberofNGOsintheirareaaswellastheirindividualcontactwiththem.AsidefromcontactwithNGOs,somequestionsaimtoassesswhetherregularKenyansmakecomparisonsbetweenNGOsanddifferentaspectsofgovernmenttodeterminewhichtheyviewmorefavorably.

Inthispaper,legitimacyisoperationalizedinseveraldifferentways.ItfollowstheWorldValuesSurveyandAfrobarometermodels,askingquestionsconcerningperceptionsoftherighttogoverninproceduralterms.Theseaskrespondentssuchthingsaswhethertheybelievestatecourtsshouldalwaysbeobeyed;whetherpolicehavetherighttoenforcelaws;whetherthestatehastherighttocollecttaxes;andwhetherthelawsrepresentthewaytheyshouldberuled.Questionsalsoaskaboutvotingandothermeasuresofcivic4WhileIdesignedthesurveyandoversawitsimplementation,IhiredtheNairobi‐basedinternationalfirmSteadmanGrouptotranslatethesurveytothreelocallanguages(Swahili,KambaandKimbeere),administeritandinputthedatatodigitalformatusinghigh‐speedscanners(15%ofwhichwererescannedand10%ofwhichwerecheckedmanuallyforqualitycontrolpurposes).Steadman,whichwasintheprocessofbeingacquiredbytheBritishcompanySynovate,isthelargestsurveyfirminEastAfricaanddoesconsiderableworkforclientslikeUNDP,Afrobarometer,andotherAmericanscholars.Inadditiontoscholarlyresearch,theyhaveanextensivecommercialmarketingand24‐hourmediamonitoringbusiness.5InMbeere,samplingtookplacesintwenty‐threelocationsrepresentingtheentiredistrict.InMachakos,samplingwasdonein25locations.6TheseneighborhoodsincludedUmojaI,Mbotela,BuruBuru,Bahati,Maringo,OfafaJericho,Uhuru,Kayole,Makadara,andDonholm.7InMachakosTown,samplingwasdoneinBondeni,Katoloni,Kariobangi,KenyaIsrael,College,Mjini,Miwani,Muthini,EastleighandMajengoII.

participation.Thesehaveallbecomecommonmeasuresoflegitimacyinsurveyresearch(WorldValuesSurvey,Afrobarometer,Gilley2006and2006b),andremainvalidregardlessofslightvariationsinthedefinitionoflegitimacyemployed.

Otherquestionsaskmoredirectlyaboutpopularsupportforandtrustingovernment.Whileperhapsnotidenticaltolegitimacy,popularsupportismorecomprehensibletotheaveragepersonthantheabstractconceptlegitimacy.ThisisparticularlytrueintheruralAfricancontext,where“legitimacy”mustoftenbetranslatedintolocallanguagesanddiscussedwithpeoplewhohavehadlimitedaccesstopost‐primaryeducationopportunities.Popularsupportthereforeservesasagoodproxyforlegitimacyinthiscontext,especiallywhencombinedwiththemorecommonsurveymeasures.

Awidevarietyofcontrolvariableswereincludedinbothsurveys.Issuesofeducationlevel,socio‐economicclass,viewsontheeconomy,accesstomedia,firstlanguage,gender,ageandurbanizationlevelareallconsidered.

FindingsMostKenyanshavelittledirectcontactwithNGOs Simpledescriptivestatisticsprovidethefirstfindinginthispaper:countertosuggestionsthatNGOsareoverwhelmingthestate,mostKenyansactuallyhavelittledirectcontactwiththem.Althoughanot‐insignificantproportionofservicesareprovidedbynon‐governmentalorganizations,manyoftheseserviceswereprovidedinjointNGO‐governmentarrangements.Thegovernmentprovidesthemajorityofservices,asseeninthetablebelow.8

TableOne:PrimarySocialServiceProvisionbyTypeofOrganization

PrimarySchoolType ResponseFrequencyPercentofSub­

TotalGovernment 342 75.0%Private 66 14.5%JointGovernment&Missionary 24 5.3%Missionary/FBO 21 4.6%Don'tKnow 3 0.7%Sub‐Total 456

HealthcareProviderType ResponseFrequencyPercentofSub­

TotalGovernment 538 70.8%Private 127 16.7%Missionary/FBO 89 11.7%NGO 2 0.3%JointGovernment&Missionary 1 0.1%Don'tKnow 3 0.4%

8Totalresponsesreachover501becauserespondentswereaskedtonamemorethanoneschool,medicalfacilityorsecurityprovider,ifapplicable.

Sub‐Total 760

SecurityServiceProvider ResponseFrequencyPercentofSub­

TotalGovernment:Police 425 47.4%Government:AreaChief 19 2.1%CommunityPolicingInitiative 200 22.3%Vigilantes9 129 14.4%PrivateGuards 100 11.2%Other(myself,dogs,family,God) 21 2.3%NoAnswer 2 0.2%Sub‐Total 896

Elsewhereinmywork,IshowthatthetypesofservicesofferedbyNGOsareusuallyindirectassistanceintheformoftraining,bursaryprovision,HIVawarenessandpreventionactivities,etc(Brass2010,ChapterThree).Outsideofremotepartsofthecountry,whereintervieweesavowthatNGOsdoprovidecoreservices,thesesurveyresultsareprobablygeneralizabletotherestofKenya.

Clearly,NGOsarenotoverwhelmingthestateinprovidingbasicservices,thoughtheyarecontributingtoserviceprovision.Indeed,mostKenyansinthesurveyhadhadlittleornocontactwithNGOs,(excludingchurches).10WhileNGOsarequitevisibleontheinternationalsceneandamongtheeliteinmostcities,mostwananchisurveyedarenotawareofbeingindirectcontactwiththem.WhenaskedwhethertheythinktherearemanyorfewNGOsintheirarea,only33.4percentgaveresponsesof“many”“some”or“few”NGOsinthearea,while53.3percentsaid“none.”11Similarly,25.5percenthavebeenapproachedbyanNGOprovidinggoodsorservicesintheirareaatleastonce,with24.2percentofrespondentshavingdeliberatelygonetoanNGOforassistance.12

Thelevelofcontactisdependenttosomeextentonwhereapersonlives.Thereisavastdifferenceinthepercentagegivinganon‐zeroresponseinMachakosTown–ahighof57percent–versusmiddleclassNairobiorMbeeredistrict(seetablebelow).ThiscorrespondstoandconfirmstheconclusionpresentedinBrass(2010,ChapterTwo),thatNGOslocatewheretheyareneeded,butalsowhereitisconvenienttoreachagreatnumberofpeople.MiddleclassNairobiresidentsdonotparticularlyneedNGOs,andwhileruralMachakosandMbeereresidentsdoneedthem,theyarelessconvenientthanMachakosTown.(Inparticular,thereisnolargetowninMbeere;it’sdistrictheadquartersisnotonapavedroad,buttwelvekilometersdownabumpymurramroad.)ThesefindingsregardingNGOpenetrationalsoconfirmdatacollectedbythegovernmentNGOboard

9VigilantesintheKenyancontextdonothavequitethesameconnotationasintheWest.Usuallytheyaregroupsofyoungmenwhohavetakenprovisionofsecurityoftheirneighborhoodorvillageuponthemselvesforasmallfee.10ThequestionsdidnotspecifywhatwasmeantbyNGO,leavinginterpretationopentotherespondent–althoughitimmediatelyfollowedaquestionaskingrespondentsto“defineNGO.”Clearly,giventhehighrateofchurchattendanceinKenya,thesefiguresdonotreflectcontactwithchurches.11Respondentschosefrom“Verymany,Some,VeryFew,NoneorDon’tknow.”12.4percentofrespondentssaidtheydidn’tknow.12Thesegroupsarenotmutuallyexclusive.

(2006),whichsaysthatMachakoshadahigherconcentrationofNGOsthanMbeere.Inthisgroupofpeople,then,aboutone‐quartertoone‐halfofthepeoplefeltexposuretoNGOs.

InyouropinionaretheremanyNGOsworkinginyourareaorfew?

TotalPositiveResponse VeryMany Some VeryFew None

Don'tKnow

UrbanMachakos 57.0% 1.0% 18.0% 38.0% 39.0% 4.0%MachakosDistrict 30.7% 1.3% 10.0% 19.3% 64.7% 4.7%UrbanNairobi 28.7% 1.0% 5.9% 21.8% 43.6% 27.7%MbeereDistrict 26.7% 2.7% 6.0% 18.0% 58.0% 15.3%

Inaddition,aboutone‐third(33.5percent)ofallrespondentssaid,“Idon’tknow”

whenaskedtodefineNGO.Whileit’spossiblethishadsomethingtodowiththewayenumeratorsaskedthequestion,it’sequallylikelythatmostKenyansarenotasfamiliarwiththe“NGORevolution”asweareintheWest,especiallyamongthoseininternationaldevelopmentandpublicadministrationcircles.Thisstatisticalsosuggeststhat,unlikeintheWest,regularKenyansdonotassociatemostmissionorchurch‐basedeffortsasbelongingtotheNGO.Onlyatinyfractionofwell‐educatedKenyans,however,didn’tknowwhatanNGOis.Indeed,ofthosewhocouldnotdefine“NGO,”147ofthe168respondents(87.5percent)wereclassifiedinthelowestofsevensocio‐economicbracketsinKenya,withanother14(8.3percent)locatedinthesecond‐lowestbracket.Whilethemajority(72%)ofthetotalrespondentsfellintothesetwoextreme‐povertybrackets,thisdatacouldbeinterpretedtomeanthatNGOsarenotreachingthepoorestofthepoor.Itcould,however,alsomeanthatNGOs,particularlydevelopmentandrelieforganizationsassociatedwithchurchesorreligiousinstitutions,arenotidentifiedquaNGOsbythepoor.Clearly,thereisasignificantlevelofcontactwithNGOs,butitisnottheoverwhelmingpresencesometimesportrayed.NGOvs.GovernmentLegitimacy KenyanswhodohaveopinionsaboutNGOs,however,tendtothinkveryhighlyofthem–oftenmorepositivelythantheythinkofthegovernment.NGOsinKenyaareconsideredlegitimateorganizations,givenrespectandappreciationbymostrespondents:“peoplearejusthappywithNGOs”(2008‐32).Forexample,whenrespondentswereaskedtworelatedquestionsregardingthemorallegitimacyofNGOs,roughly58percentand56percentrespondedaffirmatively,whileonlyabout13percentand16percentrespondednegatively(therestwereneutralordidn’tknow).Specifically,respondentswereaskedtowhatextenttheyagreedordisagreedwiththestatements,“NGOsingeneralsharemyvaluesanddotherightthing”and“WhatNGOsbelieveisgoodforKenyaisthesameaswhatIthinkisgoodforKenya.”13

ThoughtsonNGOscomparefavorablyorwereaboutequaltothoseonthegovernment.Whenaskedthefirsttwoquestionsaboutgovernment,favorableresponseswereverysimilartothoseforNGOs.65percentofrespondentsagreethatthegovernmentofKenyasharestheirvalues,while56percentfeltthatwhatthegovernmentofKenya13Respondentswereaskedtochoosefromtheoptions“stronglyagree,”“agree,”“neitheragreenordisagree,”“disagree,”and“stronglydisagree.”Theycouldalsoanswer“don’tknow,”and13and12percentdidsorespectively.

believesisgoodforKenyaisthesameaswhattheybelieveisgoodforKenya.MorepeoplehadnegativeviewsofgovernmentthantheydidofNGOs,however;21and27percentofrespondentsdisagreedorstronglydisagreedwiththestatementsrespectively.

Organization“sharesmyvaluesanddoestherightthing”

GovernmentofKenya NGOsStronglyAgree 7.4% 5.2%Agree 57.3% 52.7%Neitheragreenordisagree 13.8% 17.6%Disagree 18.4% 10.4%Stronglydisagree 2.8% 2.2%DK 0.4% 12.0% Whatorganization“believesisgoodforKenyaiswhatIthinkisgoodforKenya”

GovernmentofKenya NGOsStronglyAgree 8.0% 5.6%Agree 47.9% 50.3%Neitheragreenordisagree 16.4% 14.8%Disagree 22.6% 11.8%Stronglydisagree 4.6% 4.2%DK 0.6% 13.4%

Respondentswerealsoasked,“TowhatextentdoyouthinkthatKenya‐basedor

internationally‐basedNGOshavetheinterestsofthepeopleinmind?”Justlessthan63percentoftotalrespondentsthoughtthatKenya‐basedNGOs“sometimesorusually”havetheinterestsofthepeopleinmind,whileafull75percentofpeoplefeltthiswayaboutinternationalNGOs.14Asoneinformedobserversaid,“ThecommonmanlikesNGOstentimesgovernment!Goaskanyoneonthestreet”(2006‐6).WhenaskedwhetherKenyanpoliticiansorcivilservantshavetheinterestsofthepeopleinmindonly33percentand53percentofKenyansrespectivelyresponded“sometimes”or“usually”–considerablyfewerpeoplethinkgovernmentofficialshavetheirinterestsinmindthandoNGOs.

Towhatextentdotheseorganizationshavetheinterestsofthepeopleinmind?

KenyanPoliticians

KenyanCivilServants

KenyanNGOs

InternationalNGOs

Neverhaveinterestsinmind 35.9% 14.6% 6.6% 2.4%Rarelyhavetheinterestsinmind 27.9% 27.4% 19.2% 13.6%Sometimeshavetheinterestsinmind 27.5% 33.1% 42.5% 36.9%Usuallyhavetheinterestsinmind 5.6% 19.8% 20.4% 38.3%NoAnswer 3.0% 5.2% 11.4% 8.8% Putinbargraphform,weseetheproportionofrespondentsthinkingfavorablyaboutpoliticiansistheinverseofthatforinternationalorganizations.Ifweweretodraw

14Respondentsweregiventheoptions,“never,”“rarely,”“sometimes,”and“usually”havetheinterestsofthepeopleinmind.

linesfollowingtheblue(politicians)andpurple(internationalNGOs)barsbelow,theywouldformanX.Civilservants(inred)andKenyan‐basedNGOshaveasimilartrendasinternationalNGOs,iflesspositive,particularlyatthehighest“usually”frequency.

Towhatextentdotheseorganizationshavetheinterestsofthepeopleinmind?

Finally,whenrespondentswereaskedhowmuchconfidencetheyhadinvarious

governmentalandnon‐governmentalorganizationsandinstitutions,rangingfromindividualslikethepresidenttoorganizationslikethepolicetoKenyan‐basedNGOs.“GovernmentofKenya”ratedveryhighlyasanaggregatedcategorywithnearly74percentofrespondentsgivingapositiveassessment.NGOsdidaswell–particularlyinternationally‐basedones,inwhichalmost69percentfeltconfident.Arguably,theaggregategovernmentcategoryratedsohighlybecauseKenya–andAfricagenerally–hashighlypersonalisticpoliticalsystems,leadingtoatendencyamongcitizenstoassociatethestatewiththepersonofthepresident.Nearly81%ofrespondentssaidtheyhadagreatdealorquitealotofconfidenceinpresidentKibaki.15ConfidencelevelsinorganizationslikeParliament,thecivilserviceandthepolicewereagoodbitlower,at61,58and54percentconfidencelevelsrespectively.

Howmuchconfidencedoyouhaveinthefollowingorganizationsorinstitutions?

GenerallyPositive

Agreatdeal

Quitealot

Notverymuch

Noneatall NA

President 80.8% 37.1% 43.7% 16.8% 2.2% 0.2%GovernmentofKenya 73.7% 30.1% 43.5% 22.0% 4.0% 0.4%InternationalNGOs 68.5% 25.0% 43.5% 19.4% 6.2% 6.0%PrimeMinister 62.1% 23.6% 38.5% 27.2% 10.6% 0.2%Parliament 61.3% 19.4% 41.9% 31.4% 5.8% 1.6%KenyanNGOs 58.7% 21.4% 37.3% 26.4% 8.4% 6.6%

15Thisisaremarkablyhighpercentage,giventhatKibakiiswidelythoughttohavestolenthepresidentialelectionthattookplacelessthanayearbeforethesurveywasconducted.Someofthiscanbeexplainedbythefactthattherewaslittleelectoralviolenceintheareaswherethesurveywasconducted,andmostKambasandMbeeres,thelargestethnicgrouplivinginMachakosandMbeererespectively,tendtosupportKikuyupoliticianslikeKibaki.Moreover,Kibaki’svicepresidentcomesfromtheKambagroup.

0.00%5.00%10.00%15.00%20.00%25.00%30.00%35.00%40.00%45.00%

KenyanPoliticians

KenyanCivilServants

KenyanNGOs

InternationalNGOs

CivilService 58.1% 16.4% 41.7% 37.1% 4.2% 0.6%Police 53.7% 13.4% 40.3% 33.5% 12.0% 0.8%

ItmaybeworthnotingthatnearlyalloftheresponsesinthissectionofquestionsaremuchhigherthanonewouldhaveexpectedaprioriforasurveyofKenyans.16Wolf,etal.(2004)foundaverysimilarsituationinthedatatheycollectedforthefirstwaveoftheAfrobarometerresearchdoneinKenya,in2003.Theyrefertotheoverwhelminglypositiveresponsesas“euphoria,”stemmingfromthepresidentialdemocraticturnoverofpowerinlate2002aftertwenty‐fouryearsofsemi‐authoritarianrule.Thistendencytoward“euphoric”answersonthepartofKenyansdoesnotinvalidatetheresponses.Infact,itaccordswiththepositivefindingsongovernmentlegitimacyfromadifferentsetofdatausedelsewhereinmyresearch(Brass2010,ChapterFive).

NGOImpactonLegitimacy? MoralLegitimacyClearly,peopledothinkhighlyofNGOs.ButhowdoestheirexposuretoNGOsimpactthewaythattheyperceivegovernmentandthestatebroadlyspeaking?DoesexposuretoNGOscorrelatewithhigherorlowerlevelsoflegitimacyofgovernment?ItestthisquestionusingseveralmeasuresoflegitimacyandNGOcontactorinfluence,comparingbetweenwholedistrict(rural)andurbansamples.17Fortheseanalyses,contactwithNGOswasquerieddirectly,makingthelinkagesbetweenNGOsandlegitimacymoretransparent.

Holdingallelseconstant,thereisnotasignificantimpactonperceptionsofgovernmentfromcontactwithNGOs.PeoplewhohavesoughtoutNGOsforagoodorservicewithintheyearpriortothesurveydonothavestatisticallysignificantlydifferentviewsofgovernmentthanthosewhohaven’t.Andtheredoesnotappeartobea“transfer”oflegitimacyaffectfromgovernmentagenciestoNGOs–meaningthatpeoplewhoviewNGOsveryhighlydonotviewgovernmentlessfavorably.WhileIcouldshowtheseoutcomesinmyriadpermutations,Idisplayonlyafewbelow.Theseshouldnotbeconsideredexhaustive,butrepresentative. Iuseanumberofvariableslargelyascontrolvariables.Ibeganbyimaginingwhatmightcorrelatewithhigherorlowerlevelsoflegitimacybesidesmyprincipleindependentvariableofinterest,NGOs.Iincludedquestionsmeasuringthesecontrolsinmysurveyinstrument(seetablebelow).Controlsincludecontinuousvariablesmeasuringage,educationlevel,socio‐economicclass,18andviewsonthepresentstateofthenational16WhiletherespondentsweretoldthatthesurveywasconductedforaresearcherattheUniversityofCalifornia,theenumeratorswereadultKenyans.Theyweretrainedenumeratorsinstructedtodressandspeakappropriatelyfortheregiontheywerein,andtospeakinlocallanguageswhenpreferredbytherespondent.Forthesereasons,Idonotbelieverespondentswereansweringthewayan“outsider”wouldwantthemto,buthowtheythoughtthey“should”moregenerally.17Thewhole‐districtsampleincludesarepresentativesampleofallofMachakosandMbeeredistricts.Theurbansampleincludes101respondentsfromworkingclassNairobiand100respondentsfromurbanMachakosTown.18Ireportonlytheregressionsusingacompositemeasureofclassrangingfrom1(extremepoverty)to17(extremelywealthy),basedontheLivingStandardsMeasurementSurvey(LSMS)developedintheearly1980sintheWorldBank,whichhasbeenusedinmorethan70countries(http://go.worldbank.org/WKOXNZV3X0).Ialsotestedtheresultsusingdifferentmeasuresofeconomic

economy(usingaLikertscale),19aswellasdummyvariablesforsexandsetting(inthefulldistrictsampleonly).Ialsoincludeameasuretogaugetherespondents’politicalactivitylevel,usingtheirhistoryofattendanceatprotestsordemonstrationsandwhethertheyaregenerallythetypeofindividualwhoraisesissueswithgovernmentadministratorsorpoliticians(seeAppendicesfortables).20AlthoughNGOlevelsarenotsignificantcorrelateswithhigherorlowerperceptionsofthegovernment,severaloftheotherfactorsare.

Asmykeyindependentvariableofinterest,IrelyonasurveyquestionregardingwhetherapersonhasactivelysoughtoutorvisitedanNGOwithinthepastyearfortraining,information,goodsorservices.21IusethisparticularmeasureofNGOcontactinpartbecausethisquestionasksrespondentsabouthavingtakeninitiativetowardNGOs,andwemightexpectthatproactiveNGOseekersmightbebothmorelikelythanotherstotransferlegitimacy,andtoalreadybeirritatedbythegovernmentthanthosewhohavenotactivelysoughtgoodsorservicesfromanNGO.Ialsousethisvariableforpracticalreasons–thereareasignificantnumberofindividualswhohavesoughtoutanNGOatleastonceinthepastyear(aboutone‐thirdofthetotalsample),andthereareremarkablyfew“don’tknow”or“NA”responses–only3respondentsoutof501.22

TableofVariables

DependentVariables(stataname)

Description

Govtsharesmyvalues

Measuresmorallegitimacy,askingwhethertherespondentagrees(3),disagrees(1)orisneutral(2)withthestatement,“TheGovernmentofKenyaingeneral,"sharesmyvalues"and"doestherightthing”

AgreementwithGovernment

Measuresmorallegitimacy,askingwhethertherespondentagrees(3),disagrees(1)orisneutral(2)withthestatement,“WhattheGovernmentbelievesisgoodforKenyaisthesameaswhatIthinkisgoodforKenya.”

Laws

Measuresprocedurallegitimacy,askingwhethertherespondentagrees(3),disagrees(1)orisneutral(2)withthestatement,“ThelawsofKenyaexpressthevaluesandmoralsofpeopleinthiscountry.”

Police

Measuresprocedurallegitimacy,askingwhethertherespondentagrees(3),disagrees(1)orisneutral(2)withthestatement,“Thepolicealwayshavetherighttomakepeopleobeythelaw.”

status,suchasthetypeofhousingandtherespondents’accesstotelevision.Thesegeneratedsimilarresults,soarenotincludedinthetablesshown.19WhilenotallsocialscientistsagreethatLikertscalesshouldbeusedwithintervalprocedures,manyagreethatanitemwithatleastfivecategoriescanbeusedinthismanner.Forexample,JaccardandWan(1996:4)summarize,"formanystatisticaltests,ratherseveredepartures(fromintervalness)donotseemtoaffectTypeIandTypeIIerrorsdramatically.”20Whilenotreportedinthetablesbelow,Iusedseveraldifferentmeasuresofpoliticalparticipationinclinations,includingbeingregisteredtovote,theirlevelofdisappointmentwiththe2007elections,andhavingvotedinthe2002and2007elections.Allturnedbacksimilarresults.21Thequestionwas:“Inthepastyear,howmanytimeshaveyougonetoanNGOseekingtraining,information,aserviceorforaphysicalgood?”Inthesurvey,respondentschosebetween“Once,twotofourtimes,morethanfourtimesorNever.”Becauseoftherelativelysmallsamplesizesinmydata,Ihaverecodedthevariableasabinaryvariable,where1=hasgonetoanNGOforgoodsorservicesinthepastyear,and0=hasnotgonetoanNGO.22OtherpossiblesurveyquestionsaskedrespondentswhethertheythoughtthereweremanyorfewNGOsworkingintheirarea,andwhetheranindividualhasbeenapproachedbyanNGOforservicesorgoodsinthepastyear.Sincethenon‐responseratesonthesequestionsweremuchhigher(12percentand6percentrespectively),makingforasignificantlossinobservationsorpower,IusethequestionaboutvisitinganNGO.

IndependentVariable

GonetoanNGO(Dum_GoneNGO)

SignifieswhetherarespondenthasgonetoanNGOseekingtraining,information,aserviceorforaphysicalgood,inthepastoneyear.1=Yes;0=No.

ControlVariables Age(Age) Ageinyears,rangesfrom18to85Gender(Gender) Male(0)orfemale(1)Education(highest_ed)

Highestlevelofeducationobtained.Scalefrom1=Noformaleducationto7=Finisheduniversity.

Class(sclass1_17)

Variablemeasuringwealthbasedonaquestionnaireaskingwhichof22itemstherespondent’shouseholdcontains.Rangesfrom1(extremepoverty)to17(verywealthy).ThistoolisbasedontheLivingStandardsMeasurementSurvey(LSMS)developedintheearly1980sintheWorldBank,whichhasbeenusedinmorethan70countries(http://go.worldbank.org/WKOXNZV3X0).

EconomicViews(econ_condtns)

Responsetothequestion,“Ingeneral,howwouldyoudescribethepresenteconomicconditionsofthiscountry?”Scalefrom1=VeryBadto5=VeryGood,with3beingneutral.

Protest(dum1protest)

Respondentswereaskedwhether,ascitizens,theyhadeverattendedaprotestmarchordemonstration.Dummyvariableforhavingattended(1)orhavingneverdoneso(0).

RaisedIssuewithgovernment(raised_issue)

Respondentswereaskedwhether,ascitizens,theyhadevergottentogetherwithotherstoraiseanissuewithalocalpoliticianoradministrator.1=Yes;0=No.

Inthetwotablespresentedbelow,Ipresentthelog‐oddsfromorderedlogistic

regressionsinwhichthedependentvariablemeasuresrespondents’agreementwiththestatements:“TheGovernmentofKenyaingeneral,‘sharesmyvalues’and‘doestherightthing’”and,“WhattheGovernmentbelievesisgoodforKenyaisthesameaswhatIthinkisgoodforKenya”respectively.23Thesequestionsmeasurethegovernment’smoralrighttogovern,akeycomponentinstatelegitimacy.

RuralDistricts UrbanAreas

ModelOne ModelTwoModelThree ModelFour ModelFive ModelSix

VARIABLES

Govtsharesmyvaluesanddoestherightthing

Govtsharesmyvaluesanddoestherightthing

Govtsharesmyvaluesanddoestherightthing

Govtsharesmyvaluesanddoestherightthing

Govtsharesmyvaluesanddoestherightthing

Govtsharesmyvaluesanddoestherightthing

dum_goneNGO 0.08 0.08 0.03 ‐0.05 0.09 0.13 (0.30) (0.30) (0.30) (0.36) (0.37) (0.37)highest_ed ‐0.04 ‐0.04 ‐0.05 0.25* 0.38*** 0.38*** (0.09) (0.09) (0.10) (0.13) (0.14) (0.14)sclass1_17 ‐0.18** ‐0.18** ‐0.17** 0.01 ‐0.04 ‐0.03 (0.07) (0.07) (0.07) (0.04) (0.04) (0.04)gender 0.26 0.26 0.38 0.38 0.58* 0.54*

23Thedependentvariableisanordinalvariable,inwhich1=doesnotagree,2=neutral,and3=agree.

(0.27) (0.27) (0.28) (0.31) (0.32) (0.33)age 0.00 0.00 0.01 0.00 0.00 0.00 (0.01) (0.01) (0.01) (0.01) (0.01) (0.01)setting3 2.23** 2.23** 2.08* (1.10) (1.10) (1.11) dum1protest 0.77* 0.76 0.74 ‐0.33 ‐0.95 ‐0.97* (0.42) (0.47) (0.47) (0.52) (0.58) (0.59)raised_issue 0.02 0.02 1.96*** 1.84*** (0.31) (0.32) (0.61) (0.61)econ_condtns 0.24** 0.19 (0.12) (0.15) CutPoint1 ‐1.41** ‐1.41** ‐0.74 0.16 0.85 1.27 (0.59) (0.59) (0.66) (0.79) (0.82) (0.92)CutPoint2 ‐0.66 ‐0.66 0.01 0.86 1.59* 2.00** (0.59) (0.59) (0.66) (0.79) (0.83) (0.93) Observations 29724 297 294 196 196 188Coefficientsarelog‐odds;Standarderrorsinparentheses***p<0.01,**p<0.05,*p<0.1

Examiningthetableabove,wefindthatevenforthesetofindividualswhohavesoughtoutanNGOatleastonceinthepastyear,thereappearstobenosignificantcorrelationbetweenNGOuseandviewsongovernmentlegitimacy.Ifanything,havingvisitedanNGOinthepastyearraisesarespondent’sviewofthegovernmentveryslightly.Usingpredictedprobabilities,wecanmaketheregressioncoefficientsmoresubstantiallymeaningful.Forexample,lookingatModelThreeofthefullruraldistrictssubsampleabove,ifwesetthevaluesoftheindependentvariablesforage,educationlevel,socialclassandviewsontheeconomyattheirmean,andlookatamale,rural,politicallyinvolvedandactiverespondent,wefindthattheprobabilityoffeelingthatthegovernmentsharestherespondent’svaluesremainsflat,increasingbyanearzero0.005,ifthatpersonhasvisitedanNGO,from0.729to0.734.Thisresultisnotstatisticallysignificant,norisitsubstantivelymeaningful.ThischangelevelforhavingvisitedanNGOissimilarinruralandurbanareas.

Othervariables,however,havebothstatisticallyandsubstantivelysignificantimpactsongovernmentlegitimacy,ascanbeseeninthetablebelow.Again,theseresultsareforaruralmanofaverageage,educationlevel,class*andviewsontheeconomy*,whoispoliticallyactiverural*(*exceptwherespecifiedinthetable).TheseresultssuggestthattheaverageKenyanfindsthegovernmenttobelegitimate.Respondentshaveabouta3in4likelihoodthattheywillbelievethatthegovernment“sharestheirvalues”and“doestherightthing.”However,accordingtotheseresults,anurbaniteisconsiderablymorelikelytofindthegovernmentlegitimatethanisaruraldweller,allelseheldasstatedabove.25The

24Onthedependentvariable,sixresponsesweredroppedfor“don’tknow”answers.Observationnumbersarelowerinthistableandthosethatfollowbecauseof“don’tknow”andnon‐responsesinboththisdependentvariableandanumberofcategoricalindependentvariables.25Note,however,thatinthissampleof300,only10respondents,or3.3%arecodedaslivinginurbanareas.

sameholdstrueofapersonwhohasverypositiveviewsonthecurrentstateoftheKenyaneconomy,allelseunchanged.

VariableRangeofChange

PredictedProbabilityofFindingtheGovernmentLegitimateat

BottomofRange

PredictedProbabilityofFindingtheGovernment

LegitimateatTopofRange

ChangeinthePredicted

ProbabilityofFindingtheGovernmentLegitimate

SocialClass

FullRange(from0to17) 0.76 0.22 ‐0.54**

Setting

FromRuraltoUrban(0to1) 0.73 0.93 0.20*

EconConditions

FromNegativetoPositiveViewoftheEconomy

(1to5) 0.69 0.84 0.15**Thewealthieraruralindividualis,however,themorelikelyheistoviewthe

governmentnegatively.MovingfromapersoninextremepovertytooneoftherichestpeopleinKenya,theprobabilityoftheirfindingthegovernmentlegitimatedecreasesdramatically–froma3in4likelihoodtoa1in5likelihood.Thevariationonthisvariablecanbeseenmoreclearlyinthegraphbelow,inwhichthedarkgrayarearepresentsthepredictedprobabilityoffindingthegovernmentlegitimateovertherangeofeconomicclasses.

Theresultsofthesameregressionslookingattheurbansamplearesomewhatdifferent,butourvariableofinterest,NGOs,isstillnotsignificant,eithersubstantivelyorstatistically.Lookingatfactorsthatdohavearelationshipwithlegitimacy,weseethatwhereaseconomicclasshasasignificantrelationshipwithlegitimacyinruralareas,inurban

NairobiandMachakos,itdoesnotplayasignificantrole(althoughinmodelsfiveandsix,thesignisinthesamedirection).26

Ifwelookatthesametypeofpersonintheurbansampleaswedidintheruralones–thatis,ifwelookatamalerespondentofaverageage,education,classandviewsontheeconomy,whohasprotested,raisedissueswithhisadministrators,andvisitedanNGO–wefindthatthepredictedprobabilityofhisrespondingaffirmativelyaboutgovernmentlegitimacyis0.75(modelsixabove).IfhehasnotvisitedanNGO,itis0.74,althoughthisisnotstatisticallyorsubstantivelysignificantthisisaboutthesamelikelihoodasitwasinruralareas.

Inurbanareas,education,gender,andpriorpoliticalinvolvementhaveasignificantrelationshipwithgovernmentlegitimacy.Holdingallthesameexcepteducationlevel,andsettingittothehighestpossibleeducationlevel,thepredictedprobabilityoffindingthegovernmentlegitimateincreasesto0.83,asizable,butnotamajorincreaseoverthepersonofaverageeducation.Inthetownorcity,however,apersonwiththelowestpossiblelevelofeducationisfarlesslikelytofindthegovernmentlegitimate,withapredictedprobabilityofdoingsoofonly0.38.Thismakessenseifweconsiderthaturbandwellerswiththeleasteducationarelikelytohavelowemploymentopportunitiesandbewithoutthelandandcommunityavailabletotheirruralcounterparts.Theymayblamethegovernmentfortheirsituation,whilethemosteducatedinurbanareasaremuchmorelikelytohaveemploymentandsteadyincome–particularlyinthemiddleclassneighborhoodsofNairobithatwerevisited.

Variable RangeofChange

PredictedProbabilityofFindingtheGovernmentLegitimateat

BottomofRange

PredictedProbabilityofFindingtheGovernment

LegitimateatTopofRange

ChangeinthePredicted

ProbabilityofFindingtheGovernmentLegitimate

EducationFullRange(from1to7) 0.38 0.83 0.45**

Gender

FromMaletoFemale(0to1) 0.81 0.89 0.08*

RaisinganIssuewithPoliticianorAdministrator

FromNotoYes(0to1) 0.48 0.88 0.40***

Protesting

FromHaveNeverProtestedtoHave

Protested(0to1) 0.93 0.79 ‐0.14*

Inurbanareas,havingapropensitytowardpoliticalactivityisalsosignificantly

correlatedwithviewsongovernmentlegitimacy.Ifapersonhasevertakeninitiativeto

2688%ofthepopulationintheruraldistrictssampleareclassifiedinthepoorestthreeoftheseventeencategories,whereasinoururbansample,only35%ofrespondentsfallintotheseextremepovertycategories.Thismightexplainthediscrepancybetweentheurbanandruralsamplesonthewealthandeducationvariables.

raiseanissuewithalocaladministratororpolitician,theyaremuchmorelikelytofeelpositiveaboutthegovernmentthaniftheyhavenot–goingfromabouta1in2chanceofapositiveviewofgovernmenttoa9in10chance.Ifthey’veprotestedordemonstrated,however,theyarelesslikelytofindthegovernmentlegitimate.Thiscanbeunderstoodtomeanthatpeoplewhoraiseanissuewithanindividualgenerallyhavehadgoodexperiencesindoingso,whilethosewhohaveprotestedremainslightlymoredisenchantedwithgovernmentthantheymightotherwise,allelseheldasstatedabove.

TestsofRobustnessAsmeasuresoftherobustness,themodelsusedabovewereusedwithother

measuresofgovernmentlegitimacy.Thesecondmeasuredrawsonaquestionthataskswhetherrespondentsfeeltheirgovernmentholdsthesamevaluesandbeliefsastheydo.Theresultsofanorderedlogisticregressionarepresentedbelow.Thesamevariablesaresignificantusingthisalternatemeasureofgovernmentlegitimacy,withthesignspointingthesamedirection,inbothurbanandruralareas.NotethathavinggonetoanNGOinthepastyearisagaininsignificant(andpositive)inallmodels.

RuralDistricts UrbanAreas

ModelOne ModelTwoModelThree ModelFour ModelFive ModelSix

VARIABLES

WhatgovtbelievesisgoodforKenyaiswhatIthinkisgood

WhatgovtbelievesisgoodforKenyaiswhatIthinkisgood

WhatgovtbelievesisgoodforKenyaiswhatIthinkisgood

WhatgovtbelievesisgoodforKenyaiswhatIthinkisgood

WhatgovtbelievesisgoodforKenyaiswhatIthinkisgood

WhatgovtbelievesisgoodforKenyaiswhatIthinkisgood

dum_goneNGO 0.46 0.46 0.45 0.29 0.40 0.44 (0.29) (0.29) (0.29) (0.34) (0.35) (0.35)highest_ed ‐0.01 ‐0.01 ‐0.01 0.21* 0.30** 0.29** (0.09) (0.09) (0.09) (0.12) (0.13) (0.13)sclass1_17 ‐0.18** ‐0.18** ‐0.18** ‐0.04 ‐0.07* ‐0.07 (0.07) (0.07) (0.07) (0.04) (0.04) (0.04)gender ‐0.14 ‐0.14 ‐0.04 ‐0.10 ‐0.02 0.02 (0.26) (0.26) (0.26) (0.29) (0.30) (0.31)age 0.00 0.00 0.01 ‐0.00 ‐0.00 ‐0.00 (0.01) (0.01) (0.01) (0.01) (0.01) (0.01)setting3 2.51** 2.51** 2.38** (1.09) (1.10) (1.10) dum1protest 0.52 0.53 0.50 ‐0.94* ‐1.27** ‐1.25** (0.37) (0.42) (0.42) (0.51) (0.54) (0.55)raised_issue ‐0.01 ‐0.01 1.14** 1.02** (0.29) (0.29) (0.50) (0.50)econ_condtns 0.20* 0.17 (0.11) (0.14) CutPoint1 ‐1.16** ‐1.16** ‐0.61 ‐0.23 0.20 0.58

(0.56) (0.56) (0.62) (0.75) (0.78) (0.87)CutPoint2 ‐0.24 ‐0.25 0.33 0.32 0.76 1.11 (0.55) (0.56) (0.62) (0.75) (0.78) (0.88) Observations 296 296 293 196 196 188Coefficientsarelog‐odds;Standarderrorsinparentheses***p<0.01,**p<0.05,*p<0.1

Next,Iturntoadifferentsortofmeasureoflegitimacy,procedurallegitimacy,whichisthedegreetowhichcitizensbelievethegovernmenthastherighttogovernasmeasuredthroughperceptionsofitsprocedures–itsrighttocollecttaxes,toenforcelaws,andsoon.Respondentswereaskedquestionspertainingtoprocedurallegitimacy,astheyareontheWorldValuesSurveys(www.worldvaluessurvey.org),Afrobarometer(www.afrobarometer.org)andotherregionalperceptionsurveys.Respondentswereaskedhowmuchtheyagreeordisagreewiththefollowingstatements:a)ThelawsofKenyaexpressthevaluesandmoralsofpeopleinthiscountry;andb)Thepolicealwayshavetherighttomakepeopleobeythelaw.Thesequestionsprobeverydifferentelementsofgovernmentprocedure,theabstractlawsofthecountryandtheday‐to‐dayimplementationandenforcementofthoselaws.Orderedlogisticregressionoutputisshownbelowforbothquestions,andisquiteconsistentwithintheruralandurbansamples.

RuralDistricts UrbanAreasVARIABLES laws police_right laws police_right dum_goneNGO 0.357 ‐0.328 0.353 0.431 (0.342) (0.398) (0.390) (0.459)highest_ed 0.0460 ‐0.0995 0.156 0.201 (0.106) (0.133) (0.145) (0.163)sclass1_17 ‐0.306*** ‐0.211** ‐0.0212 ‐0.0496 (0.0779) (0.0897) (0.0463) (0.0498)gender 0.240 0.161 0.325 ‐0.0613 (0.313) (0.399) (0.340) (0.370)age ‐0.00881 ‐0.0197 0.0173 0.0100 (0.0103) (0.0128) (0.0154) (0.0161)dum1protest 0.695 0.0913 ‐1.461** ‐1.541*** (0.501) (0.629) (0.580) (0.541)raised_issue ‐0.113 0.573 0.651 0.115 (0.338) (0.468) (0.517) (0.524)econ_condtns 0.143 0.0380 0.506*** ‐0.0245 (0.135) (0.164) (0.173) (0.162) CutPoint1 ‐1.889** ‐3.712*** 0.873 ‐0.974 (0.740) (0.954) (0.969) (1.069)CutPoint2 ‐1.137 ‐3.040*** 1.549 ‐0.341 (0.733) (0.941) (0.973) (1.062) Observations 283 282 185 185

Standarderrorsinparentheses***p<0.01,**p<0.05,*p<0.1

RegressionresultsagainsupportthefindingthathavinggonetoanNGOforinformation,servicesorgoodsdoesnotsignificantlycorrelatewithhigherorlowerlevelsoflegitimacy,allelseheldconstant.Asformorallegitimacy,economicvariablesamongruralrespondentsandpoliticalparticipationinurbanareasappeartobemosthighlycorrelatedwithviewsonthestate’srighttogovernment.

MakingSenseoftheData ThestatisticsabovesuggestthatNGOsdonotunderminestatelegitimacy.Ifanything,theoppositeappearstobetrue:NGOcontactseemstoimproveoverallsupportforthestate.WhilescholarshiphassuggestedthatAfricanstatesmakeasocialcontractwiththeircitizensinwhichthestategivesservicesandpeoplegivethestateloyalty,itseemsthatinfactthepeoplegiveloyaltyregardlessofwhogivesservices.Why?

Interviewdatafrommyfieldresearchhelpstomakesenseoftheseregressions.Inthissection,IproposeseveralpossiblemechanismsbywhichNGOshavetheirimpact.Thesearedividedintothreetypesofmechanisms:onefocusingondeliberatepoliticalobfuscation,anotheroncognitivelimitationsamongAfricansorresearchers,andthethirdonthetypesofexpectationscommonamongthepoor,particularlytheruralpoor.SomeofthesemechanismswerebroughttomyattentionduringqualitativefieldworkinterviewswithNGOleadersandworkers,politiciansandcivilservants–meaningthattheyreflectthebiasesofmyrespondents,mostofwhomwererelativelywell‐educatedmiddleandupperclassKenyans.

UnderstandingNGOs’ImpactonGovernmentLegitimacyCause Explanation AssessmentPolitical GovernmentsuccessfullytakescreditforNGOs’work.

Peopleexpectservicesofgovernment,andtheyviewthegovernmenttohavedeliveredifitcreatesanenvironmentinwhichanyserviceproviderdelivers,includingNGOs.

Sometimes.Occasionally.

Cognitive AfricansdonotunderstandthedifferencebetweenNGOsandthegovernment–theyareallgovernmentintheeyesofwananchi.

ScholarlyerrorinassumptionsonthesourcesofAfricanstatelegitimacy.

Rarely.Sometimes.

Poverty Africancitizenexpectationsoftheirgovernmentareextremelylow,particularlyamongtheruralpoor.

Frequently.

ThefirstexplanationregardingtheneutralorpositivecorrelationbetweenNGOs

andgovernmentlegitimacylargelyinvolvespoliticalsleight‐of‐hand:theargumentisthatpoliticians,fromlocalcountycouncilorstoMembersofParliament,claimcreditfortheservicesthatNGOsprovide,therebyclaimingthepositivepopularsentiment.Governmentsdonotwantcreditforsocio‐economicprogresstogotonon‐governmentalorganizations(Bratton1989:572)–sogovernmentofficialsandevenadministratorsoftenclaimresponsibilityfortheworkthatothershavedone.Asaresult,thelegitimacyderivedfromprovidingservicesaccruestothegovernment,aswellasNGOs.

Credittakingassumesmanyforms,rangingfromtherelativelysubtledroppingofNGOs’namesintheauthorshipofjointlycreateddocuments(2008‐24)topoliticiansandadministratorspubliclyandunambiguouslydeclaringtheypersonallyarrangedfor–orevenfunded–anNGOtoworkinthearea.ThesehappeninConstituencyDevelopmentreports(2008‐10),projectopeningsorlaunches(2008‐17),dedicationplaquesoninfrastructure(2008‐18,2008‐14),andotherevents.Onoccasion,politiciansalsoattempttogarnerpoliticallegitimacybyactingthehero,publicallydenouncingNGOsforunrestorproblemsinthearea(2008‐54).AlocalpoliticianinMachakosDistrictexplainedtherationale:“Thecommunitymightnotliketoknowwheretheirservicescomefrom,butforpoliticians,whenitcomestoelectiontime,itreallymatters!Sotheyreallymakesuretospecifytothepeople”(2008‐35).

SomeNGOsencouragethegovernmenttogainlegitimacyfromtheorganization’swork–theyseetheirroleasimprovingthestateasawhole,notunderminingit.Assuch,theyarehappytoshareaccruedserviceprovision‐basedlegitimacywithlocalcivilservantsandpoliticians.Onthewhole,however,NGOrepresentativesintervieweddidnotwantcredittogotogovernmentunlessgovernmentactorsactuallycontributedinsomeway.WhileNGOsdon’tworktounderminethestate,27manywouldliketoseegovernmentbeingmoreresponsivetoitscitizenry.

Relatedtothis,somesuggestthattheGovernmentgainslegitimacybycreatinganenvironmentinwhichNGOschoosetobepresentandconducttheirwork–workwhichoftencouldjustaseasilybeundertakenelsewhere,suchasinothercountrieswithneedypopulations.IndeedmanyoftheNGOsinsomepartsofKenyaserveneighboringcountries,especiallySudanandSomalia,fromwithinKenya,sincetheatmosphereforworkinginthecountryissomuchlesshostileinKenya.OnesmallNGOleader,whenaskedhowNGOsaffectlegitimacysaid,“thecommonpeopleendupthankfultogovernmentforallowingtheNGOstooperatethere”(2008‐13).

Asecondexplanationreliesonaratherpessimisticviewofthecognitiveabilitiesofwananchi.ThislineofreasoningholdsthatregularKenyansarenotsmartenough(ornotwelleducatedenough)tounderstandthedifferencebetweenNGOsandgovernment.Thus,whenwananchiseeservicescomingfromanorganizationexternaltothevillageortown,theyassumethattheservicescomefromsomebranchofgovernment.Interviewdata,however,suggeststhatthisisnotthecase–thoughconfusiondoeshappensometimes.

Foronething,particularlyinpoororremoteareas,thestartofanewserviceprovisionprogramisoftenbignews–everyoneinthevillageknowsaboutit,howitcameabout,andbywhom.Projectsarediscussedseriouslythroughoutthevillage.Communitieshavemanagementcommitteesaswell,whoworkwithNGOstocompleteprojects–so“theyknowwhobroughtwhat!”(2008‐33).Peoplealsoknowthatit’suncommonforgovernmenttofundverylocalizedprograms:“Soifsomethingcomes,theyknowit’snotfromgovernment”(2008‐18).EvenifindividualsdonotknowparticularsabouttheNGObringinggoods,informedobserversinsistthattheyknowthatit’snotthegovernment(2008‐34,2008‐50,2008‐52).

Inaddition,NGOsmakeapointofinformingthecommunityinwhichtheydoaprojectofthesourceoftheprogram–byincludingthemintheprojectasithappens,27Cannon(1996:263)similarlyfoundthatNGOsinUgandaaretheretomeetneeds,notworryaboutwhetherornottheyareundermininggovernmentbyprovidingservices.

holdingmeetingsinthevillage,and/orleavingtheirlogooninfrastructureorsignboardsbuilt(2008‐18,2008‐25,2008‐30,2008‐32).Assuch,it’snoteasyforapoliticiantosuccessfullyclaimcreditforsomethings/hedidnotdo:forexample,whenaDistrictOfficerinoneareatriedtotakeoveracommunityboreholeprovidedbyalargeNGO,hecouldnotconvincethecommunitythathedrilledit–theyhadbeenactivelyinvolvedwiththeNGOintheprojectandknewthetruth(2008‐14).Inmanycases,“wananchiaremobilizedbeforetheprojectandthroughout,sotheyknowthatwhenapoliticiantakescredit,it’susuallynottrue”(2008‐54).

Still,confusionofthistypedoeshappensometimes,allowingNGOactivitiestoreflectpositivelyonstatelegitimacy.Mostinformedobserversthoughtthatsuchmisunderstandingswereverycontextspecific–theyaremorelikelyinextremelypoororremoteareas(2008‐31),orinsituationswheremuchisuncertain,asinrefugeeorinternallydisplacedpersonscamps(2008‐14).Confusionhasstartedtoarisearoundrelativelynewgovernmentdecentralizationoffundinginitiativesaswell–thingsliketheLocalAreaTransferFund,governmentbursaryfunds,andlocalHIV/AIDSfundsincreasinglylooklikeNGOprograms,butarerunbythegovernment.Alocalresearchcompany,ResearchInternational,hasfoundagooddealofconfusionbetweentheseprogramsandthoseofNGOswhenithasconducteditsownresearchonnon‐stateprovisionofservices(2008‐59).AsNGOsandgovernmentworktogethermoreandmore,theeasilydrawnlinebetweenthemhasblurredforwananchiaswell:“Peopledon’tknowwhat’sgovernmentfunded,what’sNGO,especiallynowthattheyareworkingsomuchtogether,”accordingtoamid‐levelcivilservant(2008‐37).

Mostcommonly,however,theattitudeofpeopledependsontheirlocal‐levelpublicadministrator–theChieforSub‐ChiefoftheLocationorSub‐Location.“Forlocalpeople,thegovernmentistheAssistantChief,”afterall(2008‐30).And“ThecommonpeopleendupthankfultogovernmentforallowingtheNGOstooperatethere,”since,“MostpeopleknowthatgovernmenthastoallowtheNGOtoparticipatebeforetheystateinanarea…”(2008‐13).

Insteadofcitingmisunderstandingonthepartofwananchi,wecanalsomakesenseofthisdatainathirdway,byreassessingourownassumptionsasscholars.AreourpriorsabouttherelationshipbetweenAfricanstates,serviceprovisionandlegitimacywrong?Perhapswhereprevioustheoriesweremistakenisatthelevelofunderstandingexpectations.PoliticalscientistanddevelopmentexpertNormanUphoffoncewrotethatlegitimacyderivesfrom“havingsatisfiedpeople’sneedsandforhavingmettheirnormativeexpectationsovertime”(1993:614,emphasisadded).DatafrominterviewsheldconcomitanttothesurveysuggestthatNGOsdonotlowerlegitimacy–particularlyinruralareas–becauseKenyans’expectationsofthegovernmentintheseareasarealreadyextremelylow(2008‐32).Mostinformedobserversfeltthatwananchidon’tmakethelinkbetweenNGOprovisionandgovernmentnotfulfillingitsendofthesocialcontract(2008‐30,2008‐34,2008‐51)because,“Peopletendtounderstandthatgovernmentcan’tdoitall.Theyseethegovernmentrepresentativelivingintheirsameconditions,etc.”(Ibid.)Theyalsomightnotunderstandthattheypayvalue‐addedtax(VAT)ongoodstheypurchaseinformalstores:“Alotofpeopledon’tthinkthattheypaytax,sotheydon’tthinkgovernmentneedstogiveanything”(2008‐27).Theydon’tbecomefrustratedbecausetoalargeextent,theyarenotholdingexpectationsthataren’tbeingmet.Inhisseminalworkonrebellion,

Gurr(1970)identifiesthisrelativedeprivationasthekeysourceofhumanviolenceandrebellion.

RegularKenyansseemtounderstandandacceptthelimitationsoftheirownandgovernment’ssituation(2008‐18).Thereisasenseofunderstanding–evenempathy–onthepartofwananchi:“Peopleunderstandgovernmenthasanobligation,butnocapacitytoreachdownsofar.Theyknow.”(2008‐25)Evenwhentheythinkit’sthegovernment’sresponsibilitytotakecareofthem,theyunderstandthatNGOssimplytrytofillinwheregovernmentcan’tprovide(2008‐23,2008‐24).Thisunderstandingmayberealistic–Kenya’spost‐independencedevelopmenthistorysuggestsitsreasonableforpeopletomakedowithwhatisavailable.

Yetonsomelevel,wananchibothunderstandthatthegovernmentisunlikelytoprovideforthem,andexpectonanabstractlevelthatthegovernmentwillstilldoso.AnNGOtalkingaboutbringingfinancialservicestoamarkettowninMbeere,forexample,said,“Peopleexpectgovernmenttodothings,likebringabankhere.Butpeoplealsounderstandthegovernmentisslow,sotheyarehappythatthe[organization]isherefornow,evenifitsnotafullbank…Inthelong‐term,though,theyexpectalotmorefromgovernment,becauseNGOswillleaveoneday.”(2008‐45).

IntervieweesunwittinglyrecalledSchatzberg’s(2001)extendedanalogyinPoliticalLegitimacyinMiddleAfrica:Father,Family,FoodoftheAfricanstateasfamily.ComparingthestatetoamotherandNGOstoacaringneighborhoodlady,oneseniorKenyanNGOworkersaid,

“Imaginethatyouhaveamother,whoissupposedtoprovideyouwithcare.Butwhatifsomeladyfromtheneighborhoodistheonewhoalwaystakescareofyou.Howdoyoufeel?Youdon’tknowherinterests!You’dratherstillhaveyourmother’slove.So…Youunderstandthatyouarebeingtakencareof,butyoureallywantyourmothertocomeanddoit.Thereisjustsomethingspecialaboutamother’scare,andthatotherwoman,youdon’tknowwhyshe’sdoingit,whatothermotivationsshehas…”(2008‐10)Eventhoughpeoplemightwanttheirowngovernmenttolookafterthem,several

NGOmanagersindependentlyworriedthatNGOsinadvertentlylowerpeople’sexpectationsofpublicofficialsandagencies(2008‐18).OnenotedthatwananchipesterthegovernmentforserviceslessfrequentlywhenNGOsareinthearea,becausepeopleknowtheycangettheservicesfromtheNGO(2008‐44).Whilethismeanstohimthatgovernmentcanbelessaccountable,italsosuggeststhatNGOslullKenyansawayfrommakingdemandsontheirpublicadministratorsandofficials.EvenagovernmentofficialsaidthatpeopleexpectsomethingfromNGOsonly,takingpressureoffgovernment,“PeoplehavecometoassociateNGOswithgettingstuff.Theydon’texpectasmuchfromgovernment”(2008‐39).Ageneralizedsentimentof,“Peopleknowgovernmentshouldprovide,butisnotableto,sotheycountonNGOs,”prevailedinmanyinterviews(2008‐54).

Anothergovernmentofficialclarifiedthat,“TheproblemactuallyiswhenNGOsarenotthere.Thewananchidon’tcarethatgovernmentisnottherewhenNGOsarethere.Aslongasoneisthere,allisokay.Butifnone,thentheygetangryatgovernment.”Hepointedout,“YourarelyfindaplacewhereneitherNGOsorgovernmentisthere.”(2008‐39)ThissuggeststhatNGOsmaydecreasetheextenttowhichpeoplemakedemandsongovernment.

Finally,manyinformedobserversthoughtthatNGOsprovidingservicesbenefitsstatelegitimacybecauseinsituationsofextremepoverty,peopleliveintoodiresituations

tocarewhereservicesandgoodscomefrom–theyarejusthappytohavesomeoftheirneedsmet(2008‐12,2008‐26,2008‐33,2008‐34,2008‐45,2008‐47,2008‐49,2008‐50,2008‐51,2008‐53).Thishappinessresultsindiffusesupport–wananchitendtobegenerallymoreoptimisticwhenservicescome.AsawomanworkinginruralMachakosexplainedit,“Desperatepeopledon’tthinkaboutwherethingscomefrom–becauseoftheirpoverty.Theyjustwantmore,soastogetoutoftheplacethey’rein.Sotheygenerallyfeelbetteraboutthegovernment,becausesomeoneishelpingthematleast!”(2008‐12).Mixingsarcasmandcoldreality,anothersaid,fortheuneducated“…solongasthereisfoodonthetable…Actually,ifthereevenisatable…well,theydon’tcarewhere[help]comesfrom”(2008‐26).Anothersaid,“Peoplearejustveryhappy.It’swonderful,theWorldConcern[NGO],forbringingthesethings”(2008‐45).Assuch,“Youdon’tseeNGOsreflectingpoorlyongovernment.YouseegovernmentbeinghelpedbyNGOs”increasingoverallsatisfactionlevels(2008‐50).

Inconditionsofextremepoverty,thefactofanoutsidercomingforpovertyreliefsuggeststopeoplethattheirlivescouldbecomelessprecariousovertime.Theygainthesensethatpeople“outthere”or–“upthere”astheysayinKenya–caresabouttheirsmallvillageanditsproblems.Whetherintentionallyornot(andIbelieveitsnot),wananchitransferthispositivefellow‐feelingtothestate,excusingitslimitedperformance.

TheRealDriversofLegitimacy Contrarytowhatsomescholarshaveasserted,NGOsdonotthreatengovernmentlegitimacyintheplaceswherethissurveywasconducted.Whatdoes?Accordingtothesesurveydata,correlateswithlegitimacydifferbetweenurbanandruralareas.Inruralareas,householdwealthstronglycorrelatewithchangesinviewofstatelegitimacy.Asruralpeoplebecomewealthier,theybecomemorelikelytobeskepticalofgovernment.Andiftheybelievetheeconomyisdoingverypoorly,theytendtofeellesspositiveaboutgovernment.Fortownandcityrespondents,politicalandcivicparticipationappearstomattermost.Urbaniteswhoengagewithgovernmentinapositive,individualmannerbyraisinganissuewiththeirlocaladministratororpoliticiantendtofeelthegovernmentismorelegitimate,whilethosewhoprotestinurbanareasfinditlessso.28

Again,interviewdatahelpsmakesenseofthis.Inparticular,manyinformedobserversfeltthataspeoplebecomebettereducated,wealthierandmoreurban,theybegantoexpectmorefromgovernment,citingtheir“rights”ascitizens.Assuch,theirexpectationsofgovernmentaswellastheirskepticismofittendtobehigher.Aschangescometocommunities,theirdemandsgoup.Theyrealizetheyshouldhavepossibilitiesthatthegovernmentisnotproviding(2008‐25).

Interestingly,whenaskedhowtheythoughtwananchiwouldrespondtothequestion:“Whichstatementdoyouagreewithmore:A)ItistheresponsibilityoftheGovernmentofKenyatoprovidemyfamilywithservices,andIexpectthattheGovernmentitselfwillprovidethem,orB)Itdoesn'tmatterwhoprovidesmyfamilyservicesaslongastheyareprovided,”

manyNGOandgovernmentofficialsinterviewedfeltthateducatedpeoplewhoknowtheirrightswouldanswer“A”(2008‐10,2008‐14,2008‐26,2008‐27,2008‐33,2008‐36,2008‐

28Ofcoursethedirectionofcausationhereisunclear;indeeditispossiblethatlowlegitimacyleadstoprotest.

37,2008‐47).Often,intervieweesmadethedistinctionbetweenthosewhounderstandtheypaytaxesandthosewhodon’t,believingpeoplewhoknowtheypaytaxeswillexpectmorefromthegovernment.OneseniorNGOleadersaid,“Butwepaytaxes!Sothegovernmenthastoprovidefromthosetaxes.It’sunfairotherwise,andit’sunsustainable.ShouldwepaytaxestoNGOs?”(2008‐10).Governmentofficialsfeltthesameway:

“(Laughs)Let’ssaybotharetrue.Theycarrythesameweight,butstatementBisslightlyheavierthanA.Peoplehavebeenenlightened,likeabouttaxes.Before,theyjustthoughtgovernmentistheretomisusefunds.Now,theyevenquerytheirMPinParliament:Whyaretherenolights?Whytheroadisn’ttarmacked?Butitdependsonliteracylevels.Inplaceswithoutliteracy,theywillsayA.Theydon’tcarewhogiveswhat.”(2008‐37)

Inthesurvey,however,urbanitesactuallytendedtorespond“B”considerablymorethantheirruralcounterparts.Movingfromthemostruralsample(Mbeere)tothemosturban(Nairobi),wefindanear‐doublingof“B”answers:36%inMbeere,49%inMachakosDistrict,60%inMachakosTown,and70%ofrespondentsinNairobisaidthatitdoesn’tmatterwhoprovidesservicesaslongastheyareprovided.AsoneNairobiresidentsays:“NairobiansarenowrealizingthattheGovernmentisincapableofsolvingallofthecity'sproblems.TheirsolutionliesinthewillofNairobiansthemselves”(Onyango,1998).

Whichstatementdoyouagreewithmore:ItistheresponsibilityofthegovtofKenyatoprovideservicesandIexpectthegovernmenttoprovidethem;orItdoesn'tmatterwhoprovidesservicesaslongastheyareprovided.

Rural

DistrictsUrbanAreas Total

Doesn'tmatter 128 131 259ResponsibilityofGovernment 172 70 242 Total 300 201 501

Summary&ConclusionsAlthoughscholarsandpoliticianshaveclaimedthatNGOsthreatentoundermine

statelegitimacy,inKenyathereislittleevidenceofthispattern.Indeed,whileNGOsareregardedquitefavorably–andoftenevenmorefavorablythantheirgovernmentcounterparts–theydonottakeawaylegitimacyfromthegovernment.Ratherthanazero‐sumgamewhereapositiveviewofNGOsbringsaboutanegativeviewofthecivilserviceorpoliticians,NGOsappeartoeitherhavenoimpactonpopularperceptionsofthestate,ortoimprovethem.

InplaceslikeKenya,thisfindingmattersforquitepracticalreasons.Inthepast,NGOshavebeendecriedbypoliticiansasunderminingtheirlegitimateauthority.Yetwhenputtothetest,suchallegationsappearnottohavemerit–theymaybeausefulscapegoatforpoliticiansfacingunhappypopulations,butdonotreflectempiricalevidence.Thispaper,whencombinedwiththeresultsofthepreviousone,demonstratesusingdatanotonlyfromtwodifferenttypesofKenyans(secondaryschoolstudentsandadults),butalsoacrossruralandurbanareasthatNGOsarenotdecreasinggovernmentlegitimacyinKenya.

AppendixA:CorrelatesofIndependentVariablesRuralDistricts goneNGO setting3 age highes~d gender sclas~17 dum1pr~t raised~e econ_c~s goneNGO 1.00 setting3 ‐0.06 1.00 age 0.05 ‐0.03 1.00 highest_ed 0.25 0.11 ‐0.19 1.00 gender ‐0.11 ‐0.16 ‐0.13 ‐0.27 1.00 sclass1_17 0.24 0.20 ‐0.06 0.52 ‐0.12 1.00 dum1protest 0.08 ‐0.07 0.10 ‐0.02 ‐0.26 0.05 1.00 raised_issue 0.12 ‐0.12 0.14 ‐0.05 ‐0.15 0.04 0.51 1.00 econ_condtns 0.09 0.17 ‐0.10 0.11 ‐0.19 0.07 0.02 ‐0.02 1.00UrbanAreas goneNGO age highes~d gender sclas~17 dum1pr~t raised~e econ_condtns goneNGO 1.00 age ‐0.01 1.00 highest_ed 0.24 ‐0.10 1.00 gender ‐0.08 ‐0.01 ‐0.18 1.00 sclass1_17 0.10 ‐0.06 0.47 ‐0.14 1.00 dum1protest ‐0.06 0.10 ‐0.11 ‐0.02 0.08 1.00 raised_issue ‐0.09 0.01 ‐0.13 ‐0.13 0.23 0.32 1.00 econ_condtns ‐0.03 ‐0.12 0.01 0.01 0.11 0.11 0.10 1.00AppendixB:DescriptiveStatistics

RuralDistricts Variable Obs Mean Std.Dev. Min Max GoK_share_~s 299 2.44 0.82 1 3q302 298 2.28 0.85 1 3laws 298 2.56 0.75 1 3police_right 297 2.77 0.58 1 3dum_goneNG~y 298 0.24 0.43 0 1 age 300 35.28 13.79 18 85highest_ed 300 3.81 1.60 1 8gender 300 0.62 0.49 0 1sclass1_17 300 1.88 1.97 1 12dum1protest 300 0.14 0.34 0 1 raised_issue 300 0.30 0.46 0 1econ_condtns 297 1.88 1.09 1 4

UrbanAreasVariable Obs Mean Std.Dev. Min Max GoK_share_~s 200 2.44 0.82 1 3q302 200 2.30 0.89 1 3laws 196 2.51 0.80 1 3police_right 197 2.64 0.70 1 3dum_goneNG~y 200 0.25 0.43 0 1 age 200 31.46 11.67 18 78highest_ed 198 5.27 1.41 2 8gender 201 0.45 0.50 0 1sclass1_17 201 6.18 4.34 1 16dum1protest 199 0.09 0.29 0 1 raised_issue 199 0.16 0.36 0 1econ_condtns 193 1.98 1.13 1 5

AppendixC:FrequencyTables

InthepastyearhowmanytimeshaveyougonetoanNGOseekingtraining,information,aserviceorforphysicalgood?

RuralDistricts UrbanAreas Total

Never 226 151 377 Once 47 28 75 2‐4Times 13 18 31 Morethan4Times 12 3 15 DK 2 1 3 Total 300 201 501 Towhatextentdoyouagreewiththisstatement:ThegovtofKenyasharesmyvaluesanddoestherightthing.

RuralDistricts UrbanAreas Total

StronglyAgree 28 9 37 Agree 166 121 287 Neitheragreenordis 42 27 69 Disagree 54 38 92

Stronglydisagree 9 5 14 DK 1 1 2 Total 300 201 501

Towhatextentdoyouagreewiththisstatement:WhattheGoKbelievesisgoodforKenyaisthesameaswhatIthinkisgoodforKenya.

RuralDistricts UrbanAreas Total

StronglyAgree 30 10 40 Agree 132 108 240 Neitheragreenordis 58 24 82 Disagree 66 47 113 Stronglydisagree 12 11 23 DK 2 1 3 Total 300 201 501 Laws

RuralDistricts UrbanAreas Total

0 2 5 7

StronglyAgree 58 26 84 Agree 158 111 269 Neitheragreenordis 34 21 55 Disagree 38 33 71 Stronglydisagree 10 5 15 Total 300 201 501 Police

RuralDistricts UrbanAreas Total

0 3 4 7

StronglyAgree 83 32 115 Agree 171 121 292 Neitheragreenordis 19 18 37 Disagree 18 23 41 Stronglydisagree 6 3 9 Total 300 201 501

Ingeneralhowwouldyoudescribethepresenteconomicconditionsofthiscountry?

RuralDistricts UrbanAreas Total

VeryGood 0 3 3 Good 40 25 65 NeitherGoodnorBad 41 29 70 Bad 58 45 103 VeryBad 158 91 249 Total 297 193 490 Whatisthehighestlevelofformaleducationthatyouhavefinished?

RuralDistricts UrbanAreas Total

Noformaleducation 11 0 11 Someprimary 56 6 62 Finishedprimary 87 16 103 Somesecondary 30 18 48 Finishedsecondary 85 97 182 Somepost‐secondary 8 15 23 Finishedpost‐seconda 17 32 49 FinishedUniversity 6 14 20 Total 300 198 498 Gender

RuralDistricts UrbanAreas Total

Male 113 110 223 Female 187 91 278 Total 300 201 501 Class‐basedonLSMS(LevelofSocialandMaterialStatus)

RuralDistricts UrbanAreas Total

1 221 35 256 2 26 18 44 3 18 18 36

4 5 19 24 5 5 8 13 6 5 9 14 7 8 20 28 8 6 21 27 9 4 8 12 10 1 12 13 11 0 5 5 12 1 6 7 13 0 5 5 14 0 5 5 15 0 6 6 16 0 6 6

300 201 501 Haveyou,asacitizen,everattendedaprotestordemonstration?

RuralDistricts UrbanAreas Total

Wouldneverprotest 219 140 359 Haven'tprotested,butwouldunderrightcircumstances 40 41 81 Haveprotested 41 18 59 Total 300 199 499 Haveyoueverraisedanissuewithpoliticiansoradministrators?

RuralDistricts UrbanAreas Total

No 209 168 377 Yes 91 31 122 Total 300 199 499

ReferencesBlattman,Chris.2009.“FromViolencetoVoting:WarandPoliticalParticipationin

Uganda,”AmericanPoliticalScienceReview.Brass,JenniferN.2010(forthcoming).SurrogatesforGovernment?NGOsandtheStatein

Kenya.PhDDissertation.UCBerkeley.Bratton,Michael.1989.“ThePoliticsofGovernment‐NGOrelationsinAfrica”World

Development.17:4,569‐587.Bratton,Michael.1989b.“BeyondtheState:CivilSocietyandAssociationalLifeinAfrica.”

WorldPolitics,41:3,407‐30.Etzioni,1965."OrganizationalControlStructure",inJ.G.March,ed.,Handbookof

Organizations,p.650‐677.Fowler,Alan.1991.“TheroleofNGOsinchangingstate‐societyrelations:perspectives

fromEasternandSouthernAfrica”DevelopmentPolicyReview9,53‐84.Gilley,Bruce.2006.“Themeaningandmeasureofstatelegitimacy:Resultsfor72

countries.”EuropeanJournalofPoliticalResearch45:499–525.Gilley,Bruce.2006b.“TheDeterminantsofStateLegitimacy:Resultsfor72Countries”

InternationalPoliticalScienceReview.27,1:47–71.Gurr,Ted.1970.WhyMenRebel.Princeton:PrincetonUniversityPress.Jaccard,JamesandChoiK.Wan(1996).LISRELapproachestointeractioneffectsin

multipleregression.ThousandOaks,CA:SagePublications.Jackson,RobertandCarlRosberg.1984.PopularLegitimacyinAfricanMulti‐EthnicState.

JournalofModernAfricanStudies.22(2):177‐198Kanyinga,Karuti.1996.“ThePoliticsofDevelopmentSpaceinKenya”inSemboja,Joseph

andOleTherkildsen(eds.)ServiceProvisionUnderStressinEastAfrica:State,NGOs&People'sOrganizationsinKenya,TanzaniaandUganda.Copenhagen:CentreforDevelopmentResearch.

Lister,Sarah.2003.“NGOLegitimacy:TechnicalIssueorSocialConstruct?”CritiqueofAnthropology23:2,175‐192.

Osodo,PatrickandSimonMatsvai.1997.“PartnersorContractors?TheRelationshipBetweenOfficialAgenciesandNGOs–KenyaandZimbabwe.”INTRAC:OxfordUK.

Owiti,Jeremiah,OtienoAluokaandAdamsG.R.Oloo.“CivilSocietyintheNewDispensation:ProspectsandChallenges”inOkello,Duncan(ed).2004.CivilSocietyintheThirdRepublic.NationalCouncilofNGOs:Nairobi.

Prewitt,Kenneth.EDUCATIONANDCITIZENSHIPINEASTAFRICA,1966‐1967:KENYASAMPLE[Computerfile].Chicago,IL:KennethPrewitt,UniversityofChicago,Dept.ofPoliticalScience[producer],196?.AnnArbor,MI:Inter‐universityConsortiumforPoliticalandSocialResearch[distributor],1976.doi:10.3886/ICPSR07301

Republic of Kenya: NGO Coordination Board. NGO Database. December 2006. Sandberg,Eve.1994.“Introduction:TheChangingPoliticsofNon‐Governmental

OrganizationsandtheAfricanState”inEveSandberg(ed.)TheChangingPoliticsofNon­GovernmentalOrganizationsandAfricanStates.WestportCT:Praeger,1‐31.

Schatzberg,Michael.PoliticalLegitimacyinMiddleAfrica:Father,Family,Food.2001.Bloomington:IndianaUniversityPress.

Selznick,Philip.1949.TVAandtheGrassRoots.UniversityofCaliforniaPress:Berkeley.Suchman,MarkC.1995.“ManagingLegitimacy.”AcademyofManagementReview20(3).

Sunshine,J.&T.Taylor2003.“MoralSolidarity,IdentificationwiththeCommunity,andtheImportanceofProceduralJustice:ThePoliceasPrototypicalRepresentativesofaGroup’sMoralValues,”SocialPsychologyQuarterly.66:2,153‐165.

Uphoff,Norman.1993.GrassrootsorganizationsandNGOsinruraldevelopment:opportunitieswithdiminishingstatesandexpandingmarkets.WorldDevelopment.21:4.

Wolf,ThomasP.,CarolynLogan,andJeremiahOwiti.2004.“ANewDawn?PopularOptimisminKenyaaftertheTransition.”AfrobarometerWorkingPaper33.

Young,Crawford.1988.“TheColonialStateanditsPoliticalLegacy”inRothschildandChazan(eds)ThePrecariousBalance:StateandSocietyinAfrica,59‐62,