Community Development In Mbeere

in Mbeere CommunityDevelopment Martin Walsh Schoolof African andAsian Studies Universiw of Sussex

April1994

draft working paper preparedfor the ODA ESCOR-fundedproject Rural Livelihood Systemsand Fann/1,{on-farmLinkages in Lower Embq Kenya,19721 to 1992-3 (Research SchemeR4816)

Community Developmentin Mbeere

Contents: 1 The Rootsof Community Development 1.1 A Brief History of Community Development 1.2 Community Developmentin Kenya 1,2.1 Community Developmentin the ColonialPeriod 1.2.2 Harambeeand Community Development 2 Community Devclopmentin Mbeere 2.1 The ChangingRole of the State? 2.1.1 TechnicalInterventionsand Aid Projects 2.1.2 The Institutionalisationof Community Development 2.1.2.1 The District Focusfor Rural Development 2.1.2.2 PoliticalAsendas 2.1.2.3 The Emperor'sNew Clothes 2.2 GrassrootsInitiatives 2.2.1 Womcn'sGroups 2.2.2 Rotating Savingsand Credit Associations 2.2.3 Schoolsand Churches 2.3 The Roleof NGOs 3 Conclusions Bibliography

Community Developmentin Mbeere

I The Rootsof Community Developmcnt

In the pastdecadethe twin conceptsof communitydevelopmentandlocal participationin the planningprocesshavebecomeessentialcomponentsof state-(donorandrecipient)Africa. This is evident,for sponsoreddevelopmentplansandproposalsin sub-Saharan example,in recent(1992) proposalsfor the provisionofBritish aid to strengthen developmentplanningandimplementationcapacityin Embu,Meru andTharaka-Nithi Districts. The termsof referencedrawnup by the British OverseasDevelopment Administration(ODA) for the consultancyto preparetheseproposalsspecifiedthat they "...will be madein the contextofa process-project with participatoryplanningasa key objective",whilea wholeannexofthe resultingreportwasdevotedto the subjectof communitydevelopment(ODI 1992).

1.1 A Brief Ilistory of Community Development

This ernphasison communitydevelopmentis, aswe shallsee,equallyprominentin the developmentstrategydrawnup by the Kenyangovemment.Despiteits apparentnovelty, it hasa long pedigreedatingbackto the yearsfollowing the secondworld war, in Kenya aswell as elsewherein the developingworld. The historyof communitydevelopmentasa developmentpolicy providesimportantinsightsinto currentpracticeand suggestsa numberoflessonswhich contemporarydevelopmentpractitionerswould do well to take

note of It alsosetsthe scenefor an examinationof communitydevelopmentin Mbeere, its effectiveness to dateandthe potentialfor the future.

approachto naturalresourcemanagement, In a recentreviewof the community-based Hassett(1994: 7) arguesthat communitydevelopmentin generalhasevolvedfrom the . The first of theseis the tradition mergingof two distincttraditionsof rural development organisations(NGOs) and of "grassroots"developmentpursuedby non-government churchesthroughoutAfric4 Asia andLatin America. The secondis the tradition of state management ofrural development,involving "penetration"by stateagencies(includingthe of govemmentministriesanddepartments) administrativeapparatusandrepresentatives into rural areasin order to imolementnationalDolicies.

wasthe An early attemptto mergethe grassrootsapproachwith statemanagement statein CommunityDevelopmentProgrammein lndia"initiatedby the newly-independent the early 1950s. As Hassettremarks,despiteits populistrhetoricthis endedup mainlyas a "top-down" programmeof agriculturalextensionin which the views of ordinaryfarmers were not well represented(1994: 8). Nonetheless"communitydevelopment"initiatives were startedin manycolonialandindependentcountriesaroundthis time. The emphasis wasvery muchon extensionandtraining,particularlyin agriculture,though often with evenlesseffectthan in the Indiancase. In the 1960sthe communitydevelopment approachgaveway to the developmentprojectasa vehiclefor overseasaid and investment,andthe rise of the developrnentplanningindustry. "Whereas",Hassettwrites, "'communitydevelopment'hadbeenbasedon a moral critiqueofthe socialand institutionalbasesof rural poverty,economicplanningtreatedpovertyasa technical problem,which couldbe solvedby the applicationof scientifictechniques.The developmentplansofthe 1960sand 1970swere almostuniversallystronglycentralised

andrelied on 'experts',often expatriates,to determinepeople'sneedson their behalf' (1994:8).

Meanwhile,an altemativetradition wasbeingdevelopedin the NGO sector,throughthe communitywork of radicalactMstsin Latin America. This tradition wastheorisedand widelypopularised in thewritingsofPauloFreire(1972). Duringthe 1970sthe Freirean and "empowerment",hada significant thought,with its stressupon "conscientisation" impactuponthe practiceof NGOs, especiallythoseworking in Asia, andsawthe of manylocal NGOs with a concemto eliminatepovertyandits corollaries. emergence Many NGOs beganto uselocal staf andfieldworkersto encouragecommunitiesor groupsofthe rural poor to start their own smallenterprisesanddevelopmentactivities, experienceandcapacitiesso that they would not be andto developtheir own management permanentlydependenton extemalaid.

The quiet successofthese "grassroots"initiativescontrastedmarkedlywith the failure of anddonor-fundedprojects. As Hassettnotes(1994. 9manylarge-scalestate-sponsored l0), developmentpractitionersincreasinglyquestionedthe designandphilosophybehind on technologicalsolutionsto suchprojects. Theywere criticisedfor their over-emphasis problems(the quick "technicalfix") andcorrespondingdisregardfor the developmentof humanresources;for the way in whichthey wereplanned,by professionalplannerswith little knowledgeof local conditionsandwith a minimumof consultationwith the people (includingthe "beneficiaries")who werelikely to be mostdirectly affected;andfor the (theiruseofa "blueprint"ratherthan"process" inflexibilityoftheir implernentation approach).In the courseofthe 1980sthis led to a growing intereston the part of state in of development, (bothdonorsandrecipientsof aid)in the socialdimensions agencies institutionalandorganisationaldevelopment,andin participatoryandprocessplanningas development:in other words to a renewedinterestin community a meansto sustainable

"the returnto development,asdevelopedin interveningyearsby the NGOs. This, (Shepherd1992,citedin Hassett1994:10),has,in the termsof communitydevelopment" and development the mergerofthe traditionsof grassroots Hassett's analysis, completed statemanagement.

While this accountofthe historyof communitydevelopmentis reasonablyaccuratein outline,it doesencodea significantbias. It is written from the perspectiveof current orthodoxy,which offersunqualifiedsupportfor the integrationof a participatory, communitydevelopmentperspectiveinto developnrentplansandprogrammes.It is possible,however,to take a more cynicalview, andrememberthe lessonsofthe CommunityDevelopmentProgrammein India. In this andother casesthe mergerof the two traditionsof developmentidentifiedby Hassettmight be betterdescribedasthe by the other (the state),where appropriationofone (grassrootsactionandrepresentation) the stateadoptedthe rhetoric of popularparticipationbut continuedto manage developmentin traditionalauthoritarianfashion. An examinationofthe history of communityparticipationin Kenyaprovidessomesupportfor this scepticalposition,and suggeststhat the cunent enthusiasmfor comrnunitydevelopmentandthe way in which it is construedandput into practiceshouldbe subjectto more critical scrutinythan it usually is.

1.2 Community Developmentin Kenya

1.2.1 Community Developmentin the ColonialPeriod

Communitydevelopmentasa policy was originallyintroducedinto Kenyaby the British in a documentsentto all the colonies colonialadministration.It madeits first appearance

referredto by his in November1948by the British ColonialSecretaryand subsequently name,asthe "CreechJonescircular". The introductionofthis new policy - alsoreferred - represented a reinterpretation of to at timesas'masseducation' and'socialdevelopment' colonialindirect rule asa d1'namicratherthan staticpolicy (shadesofblueprint and process!)responding in the colonies.It was to changeandthe demandfor development locatedwithin the overall statedobjectiveof guidingcoloniesto eventual"selfgovernment".

Accordingto Hill (1991:23), from whomthis accountis drawn,theBritishgovernment had a very definitesetof political objectivesin mind: "The CommunityDevelopment policy wasdesignedin part to meetcriticismof colonialrule by liberal lobbiesin the colonialcentre,but without putting colonialinterestsin jeopardy;in part it was also intendedto defuseoppositionby nationalistgroupsin the colonieswhich were callingfor political independence."This is a far cry from the "moralcritique ofthe socialand institutionalbasesof rural poverty" referredto by Hassett,andit is temptingto seesimilar ofthe policy. Certainlycommunity agendasat work in conternporarymanifestations developmenthasbecomeintimatelylinkedin somecontextswith the post-ColdWar conceptsof"good governance"and "democratisation"beingtouted by the westempowers with their new found political morality andfoistedupontheir former alliesin the developingworld.

The Kenyangovemmentalsohasits political agendas,not leastofwhich arethe needto defusepolitical opposition. It satisrylocal aspirationsand,like its colonialpredecessor, alsohasan interestin formulatingpolicieswhich areconsonantwith the wishesof its donors,andaremore likely to attracttheir aid. Communitydevelopmenthelpsto achi€ve all of theseobjectives.But herewe arejumping aheadof our argument: let us return to discussionofthe colonialoriginsof communitydevelopment.

The implernentationof the new communitydevelopmentpolicy was slow andundramatic, andhinderedin manyareasby the outbreakofthe Mau Mau uprisingandthe declaration ofa StateofEmergencyin 1952. At first communitydevelopmentwas addedto the tasks of District Officers: later on it becamethe responsibilityof newly-appointedDistrict took over CommunityDevelopmentOfficers. In somedistrictsAfricanssubsequently thesepositionsfrom British colonialofEcers. They did so asAssistantAdministrative Officers,the most seniorpost to which Africanswere admitted,their salariespaid by District Councilsratherthan centralsovernment.

Theseofficersput the new policy into practicein a numberofways. This includedthe associations andsports oflocation Councils,women'sclubs,ex-servicemen's organisation activities. Most importantwasthe new emphasisupon'self-helpprojects' Accordingto Hill, however,theseappearto havebeenlargelycommunallabourundera new name: "They were unpaid,organizedby chiefs,approvedand sometimescontributedto by District or Local Councils. The work centredon buildingprimaryschools,earthdams (water reservoirs)andwater catchments,maintainingor extendinglocal roads,and soil conservationwork. Severalofthe projectswere probablyinefficientin their useof manpowerowing to poor organization,lack of co-ordinationanddeficienciesoftechnical support"(1991:26). As a resulttheycontinuedto facecriticismfrom localactMstsand nationalistpoliticians. As Hill remarks,"The reality of muchcommunity development.. .wasfar from the notion of self-helpasbeingbasedon voluntarism,popular participationandlocal decision-making.CommunityDevelopmentphilosophyin the activity to allowunderthis rubricanydevelopment colonialworld ofthe 1950sappeared invoMng the communityasa whole which wasjudgedby the authoritiesto be to its benefit"(1991:26-27).

It is not difficult to draw parallelsbetweencommunitydevelopmentin the late colonial period andits currentmanifestations.Someof theseparallelsarebasedupon real continuitiesof policy andpractice,othersreflectbroadsimilaritiesin intentionsand outlook. Hill's accouatof colonialcommunitydevelopmentinitiativesin Kitui District and elsewherepointsto a numberoflocal experimentswhich areinterestingfrom this point of view. OneDistrict Commissionerin Kitui formeda District DwelopmentTeamof departmentheadsto establisha district developmentstrategy- an earlyattemptat integated local planning. Theteamadopteda planfor eachlocationto build an elementaryschoolanda dispensary,and seta targetoffifty damsa yearto be built in the district. This plan entaileda considerableinput ofcompulsorylabourandwas met with demandsfrom the KenyaAfrican Union that the work be paid Thesedemandswere ignoredandthe work went on until the targetswere largelyachieved.At the sametime the colonialagthoritieswere ableto congfatulatethemselveson the fact that suchprojects kept their subjectsbusyandleft themwith little time to indulgein politics (Hill I 991: 2728).

Despitethe apparentsuccessof someprojects,it was evidentthat their negativeimpact upon local perceptionsofthe administrationandits developmentefforts demandeda more sensitiveapproach.Oneearlypointer in this directionwasprovidedby research undertakenin 1949in SouthNyanzaDistrict by Philip Mayer, an anthropologistemployed that report(1951),Mayerrecommended In hispublished bythe colonialadministration. insteadof compulsionor evendemonstrationtechniques,muchvillageimprovement,and especiallyagriculturalmodernisation,could be undertakenby utilisingtraditionalwork party institutions. In 1953MachakosDistrict was chosenfor an experimentalproject conductedalongthe lineswhichMayer had suggestedfor SouthNyanza. This project washeadedby an African AdministrativeOfficer, JohnMalinda,andinvolvedusing od myetlrya(singtlu muethya),as lhe traditionalKambawork parties,the neighbourho

groups- the objectbeingto get roundhatredofthe basisofnew communitydevelopment existingforced communallabour system.Local electedcommitteesofthe newmyethya on whatprojectsto undertakeandhow to undertakethemandensurethe took decisions complianceof groupmembers.The Machakos myethyaundeftooka wide rangeoftasks includingbuilding schools,soil conservationmeasures(especiallyterracing),makingfarm boundaries(on newlyregisteredland),constructingdamsandother forms of water supply andstorage(for livestockanddomesticuse),clearinglocalroadsandpaths,andbuilding new houses.In the late 1950sthe Machakosprojectbecamesomethingof a showpiece, andthe work party modelof communitydevelopmentappearsto havebeensubsequently adoptedby districtofficialsin manyotherpartsof Kenya(Hill 1991:31-35).

As Hill argues,the Machakosproject serveda numberof purposes.Largepartsof MachakosDistrict were seriouslyeroded,andattemptsby the administrationto reduce livestockholdingsin the areahadfailedmiserably.Therewas somerisk that the Kamba mightjoin the Mau Mau rebellioq andit wasrecognisedthat their economicaspirations would haveto be met in definiteways. TheMachakosproject providedan ideal opportunityto do so, andit owed muchof its successto the fact that it was designedand headedby a local Kambaofficial, who phrasedthe projectasa reform of communallabour in the directionof local organisationanddecision-making.Needlessto say,the new myethyasystemdid not meetwith universalapproval,and someareasit was still associatedwith compulsorylabourandattackedby nationalistsfor this reasonQlill I 99I : madeby the 35-38). In spitethis criticism,themyethya,alongwith otherinnovations colonialadministrationin the nameof communitydevelopment,were to providepotent era. in the post-colonial modelsfor sovemments

1.2.2 Harambeeand Community Development

Althoughthis fact is not widely advertised,the colonialpoliry andpracticeof community in 1963. afterKenya'sindependence formedthebasisfor manydevelopments development The ideaof communitydevelopment,andthe participationof all Katyans in that "Harambee"(derivedfrom a development,was enshrinedin the new nationalsloganof "Let's pull together!"). Since Swahiliwork-gangcry andusuallytranslatedinto Englishas hasbeenthe first formaluseofthe termin 1963,Harambee Kenyatta's @resident-to-be) motto on the Kenyannationalcrestandthe customaryrallying-cryat political andother ralliesandmeetings.The centralmessageof Harambeewas self-reliance,andthis was expressedmost concretelyin the rural self-helpmovement.As a political sloganandthe catchwordfor an ideology,Harambeewasexplicitly meantto contrastwith colonialstate control andits manifestationin forced communallabour. In practice,however,it meant the continuationof manyofthe samekinds of interventio4 the adaptedtnwethyasystem andits analogueswrit largeandlegitimisedby the new nationalist(andtherefore unopposable)ideology. The rhetoricwas - and still is - different,but the methodsremain remarkablysimilar,if muchmorerefined.

The self-helpmovementwasco-ordinatedandmonitoredby a new Ministry of CommunityDevelopmentand SocialServices.In the first two decadesof independence this achievedmuchmorethan its colonialprogenitorscould everhaveimagined. Thousandsof primaryschoolsandhundredsof secondaryschools(known asTlarambee Schools')werebuilt with substantialcontributionsof moneyandlabourfrom local communities.Likewisemanyother kinds of local amenitywere constructedwith Harambeecontributionsandlabour: healthfacilities,unmetalledroads,improvedwater suppliesandcattledipsto namebut a few. For a time in the 1970srural developmentin 'Ujamaa" and Kenyaseemedto be synonymouswith the Harambeemovement,muchas l0

rural socialismdefinedthe policy andpracticesof developmentin Nyerere'sTanzania.It hasbeenestimatedthat self-helpprojectsaccountedfor roughly30% of all rural capital formation,and,between1967and 1973,11.4yoof nationaldevelopmentexpenditures (Gachuki1982;Holmquist1982).Theeconomiccontributionofthe state,however,was minimal:in 1979,for example,only onepercentofKenya'scapitaldevelopmentbudget was devotedto self-helpprojects@arkanet dl. 1979,cited, togetherwith the above in McCormackel al. 1986:47). references,

by the mid-1980sthereweresignsthattheHarambee Despiteits earlysuccesses, movementwasrunningout of steam. "Harambee!"asa nationalsloganhadbeendiluted to his Moi's additionof "Nyayol"(meaning"footstepsl"),a reference by President in 1978. The intentionto follow in the footstepsofKenyatta"whomhe succeeded Ministry of CommunityDevelopmentand SocialServiceshadlong sincebecomethe Ministry of Cultureand SocialServices,with responsibilityflorcommunitydevelopment falling to one departmentwithin it, the Departmentof SocialServices.The term communitydevelopmentitself was dropped,andthe Department'sCommunity DevelopmentAssistants,working at locationallevel andstill paidby their local district councils,becameSocialDevelopmentAssistants.Harambeehadbecomethe common namefor any collectivefund-raisingevent,for whateverpurpose,andthe financial exactionsof chiefsandtheir assistantsin the nameof Harambeewerewidely resented, muchascommunallabourhadbeenresentedin the colonialperiod. Now, however, resentmentwas not directedat foreign ovemrle,but at the widespreadsysternof comrption which hadbecomeendemicwithin local administration.To manypeople Hararnbeehadbecomea tax imposedby the rich uponthe poor, a far cry from its original purpose.

u

The last major resort of the self-helpprogr,unmewasin the women'sgtoup movement, co-ordinatedby the Departmentof SocialServicesandmonitoredby the Women'sBureau within the ministry. This will be discussedat lengthbelow. Otherwise,asBarkanet al. (1979:23')noted,self-helpwas skewedtowardsthe provisionof socialservicesfor the membersof rural communities.Exceptamongwomen'sgroups,little emphasiswas placed upon increasingrural productioq andthe earlier(colonial)link betweenself-helpand agriculturaldevelopmentwas largelylost. The apparentdemiseof self-helpand communitydevelopmentasan aspectof developmentpolicy was matched,however,by projectswhich requiredminimumlocal participationand, the rise of larger,donor-assisted in manycases,alsoinvolveda minimumof consultationwith the local beneficiaries- if indeedthe beneficiarieswere meantto be local people. From the point ofview of some politiciansandofEcialsthis shift certainlypaidbetter: therewas now a lot more moneyto be madeout oflarge aid projects- the largerthe better- than local communitieswhich had alreadybeenmilkeddry by local officials. Investmentandexperimentationin community developmentinitiativeswere left largelyto the NGOs andchurches,whoseactivitieswill alsobe examinedbelow.

in part,to the transitionidentifiedby Hassett(1994, corresponds, Thischangeof emphasis and discussedin section1.I above)from a communitydevelopmentapproachto the more of developmentby the statethroughlarge-scaleaid projects. The explicit management difference,in Kenyaat least,is that communitydevelopmentitselfhas alwaysbeen managedby the stateasa policy andpracticewhich stemsfrom a variety of motives,both transparentandopaque,economicandpolitical.

12

2 Community Developmentin Mbeere

The rest ofthis paperis devotedto an examinationof communitydevelopmentin Mbeere, initiatives,andthe role ofNGOs. andlooksin detailat the role ofthe state,grassroots

2.1 The ChangingRole of the State?

Despitethe rhetoric of Harambee,developmentin Mbeere,aselsewherein independent Kenya,haslargelybeenconstruedasthe responsibilityofthe state,assisted,where the natureofthe state'srole, and necessary, by externaldonors.This sectionexamines askswhetheror not this role is really changing- in the directionof increasing andcommunityparticipationin developmentplanning- asset out in decentralisation policy documentsandprescribedby Kenya'sinternationaldonors,the British ODA included. Accordingto Hassett'sgeneralhistoryof communitydevelopment,outlined arenow mergingwith the hitherto approaches above(section1.1),community-based andintervention,offeringnew promisefor dominanttradition of statemanagement equitableand sustainabledwelopment. Is this the casein Kenya"andin Mbeerein particular?

2.1.1 TechnicalIntelTentionsand Aid Projects

The primary role of the public sectorin the developmentof Mbeereover the pasttwo decadeshasbeendirective. The government,acting mainly through its different ministrieS.hasfosterednumefoustechnicalinterventions,someof them in the form of projects sponsoredby internationaldonorsand implementedwith the help of external 13

organisationsand expatriatestaff. Someof theseprojects, especiallythosedesignedto improve the basic infrastructureand relatedservices,havehad a tremendousimpact upon Mbeere. Almost every aspectof life hasbeenaffectedby the constructionof roads,bridges,dams,schools,clinics,andthe provisionof electricity,telephone services,and improvedwater supplies- to mentionjust someof the more obvious 'development', in Mbeereas elsewhere,hasbecomelargely developments.Indeed synonymouswith suchinterventions.

Riley and Brokensha(1988'.270-278,297-299)provide an overview of state-directed interventionsin Mbeerewhich is consistentwith our own observationsand more generalcritiquesof this "top-down" approachto development. Despitethe obvious chengesin qudrty of life and somepositive impactsupon the local economy,the benefrtshavebeenunevenlyspreadacrossMbeere(with greateradvantagesoften accruing,as might be expected,to the higher potentialzones),nxrny interventionshave failed or had negativeimpacts,and the overall return to developmentinvestmentsmade by the govemmentand its donorshasbeenpoor (imagining that sucha balancesheet could be drawn up, which of courseit cannot,exceptin notional terms)'

Many interventionshavebe€nmadein piecemealfashion, following the policies and progilmmes of the different ministriesand agenciesworking in Mbeere, including the ministries and departmentsresponsiblefor agriculture, livestock, public works, water development,health, and education(the namesand departmentalcompositionof these ministrieshavechangedon numerousoccasionsover the years, usually in responseto political manoeuvresand reshuffles). Other interventionshavebeenco-ordinated,to a great€ror lesserdegree,as part of wider programmes. One of the most important of thesewas the SpecialRural Developmenthogramme (SRDP). In 1970Mbeerewas chosenas one of six administrativedivisions in Kenyato take paft in the SRDP' a t4

nation-wideintegrateddevelopmentprogrirmmewhich was to plry an important role in promoting the institutionalisationof the state-managed, donor-funded,model of developmentin Kenya (compareRiley and Brokensha'sobservationthat the SRDP in particular had "unintentionallyinduceda strongsenseof dependency,of relianceon governmentto provide everything" (1988: 146). The SRDP in Mbeerewas financed by NORAD, the Norwegiangovemmentdevelopmentagency,which spentsomeKshs 17 million in the period from 1970,when the programmewas inaugurated,through to 1977. NORAD alsoprovided technicalassistanceof different kinds as part of the programme.

The main aims of the SRDP were to increaseagriculturalouput, reducerural unemployment,improve agriculnral extensionand social services,encourage decentralisationat divisional level, and to test the replicability of this programme. As RJleyand Brokenshaobserve,theseaims were not met, and five main constraintswere subsequentlyidentified: lack of credit, shortageof farm inputs, difficulties of communication,shortageof waGr, and limited extensionfacilities. To theseRiley and Brokenshaadd another,more general,constraint: failure adequatelyto understandthe existing systemof agriculture(1988:-nG271).

The SRDP's main emphasisin agriculturewas on increasingthe productionof selected cashcrops: cotton, Mexican 142beans,tobacco,castor,and, as a famine reserve, Katumanimaize. However, insufFrcientconsiderationwas given to the various constraintsreferred to above,and aboveall to existing productionpracticesand patterns of resourceallocationat the householdlevel. Where someof thesecrops have since beenmore widely adopted,it haslargely beenbecauseof the removal of someof these constraintsand the developmentof reliable markets(for exampleby B.A.T., British American Tobacco).ratherthan as a direct result of the SRDPand its extensionefforts. l5

Otherwiseagricultural productionin Mbeere,and especiallyin the lower and more arid zones,is gearedmore to coping with this environmentand the ever-presentthreat of droughtthan to producingfor the market. The outcomeof farmers' risk-aversion strategiesis considerablediversity of crop choice, and not the tendencytowardsthe monocultureof cashcropsfound in more high potentialareas.

O,theraspectsof the SRDPalso met with limited success.One of thesewas the atempt to improve water suppliesby meansof a complexnetwork of plpes. In Riley and Brokensha'sjudgement 'this project hasneverworked properly. Water is often not availableat the stand-pipesbecausethe PVC...pipes havebeenbroken, or damagedby road machinery,or there is no dieselor no sparesfor the pump, or the intake on the can be madeof other Ena river is silted or broken" (1988: 299). Similar assessments SRDPinterventions,as well astlose undertakenin the framework of other and more recentdevelopmentprojects. This doesnot meanthat development,understoodin terms of improvedtechnologyand infrastrucfure,and greateraccessto services,hasnot takenplace. As indicatedat the start of this sectionit very clearly has. One of the most successfrrlcomponentsof the SRDPwas the constructionof new roadsand improvementof existing ones,leaving Mbeerewith an unusuallygood network of mainly all-weatherroads. The nation-wideRural AccessRoadsProgrammeof the 1980ssawthe constructionof 210 km of roadsand significantly increasedaccessto marketsin remoteareasof Mbeere, while the constructionof a tarmacroad running from Kiamberethrough Kamburuand Kiritiri to Embu in the mid-1980shashad a considerableimpact on the economyof the whole of southernMbeereand (what is now) GachokaDivision (this road being a spin-off from the constructionof Kiambere Dam, which also displacedmany peopleand provided employmentfor othersin Mbeere).

lo

Other wide-ranginginterventionshavehad similarly mixed rezultsin Mbeere. These include the government'sprogrammeof land reform and registrationof individual freehold title, whosefar reachingeffectswe havedescribedelsewhere,and the various interventionsmadeby the British/ODA-fundedEMI ASAL Programme,which will be discussedbelow. It would not be difficult to generatea catalogueof project successes and failures, though we do not intend to do so here. It is perhapsmore important to considerthe wider implicationsof the top-down, project-oriented,approachto development. This approachhasbecomeso well integratedinto developmentpractice that it is often to difficult to assessit from any other framework than the one which its own practicesand practitionersprovide. While it is tempting to view development solely in terms of projectsand interventions,evaluatedby their technical, economicand social impacts,the current emphasisupon communitydevelopmentinvites us to take a wider perspective;wider, even, than the normal definition of communitydevelopment and participationwould allow. We attemptto do this in the sectionswhich follow.

2.1,2 T\e Institutionalisation of Communit5rDevelopment

In the first part of this paper(section1.2 above)we saw how communitydevelopment in Kenya has, throughoutits history, beensubjectto definition and manipulationby the state. This history is rarely mentionedin the contemporaryenthusiasmfor community development. According to current orthodoxy, this new emphasisupon local participation in the developmentprocesshasevolved in direct responseto its former absenceand the widespreadfailure of projectswhich havenot sufficiently involved the communitiesthey affect. However, knowledgeof the pasthistory of community development,and the continuity of this with the present,might lead us to questionthe orthodox account. 1'f

2.1.2.1 The District Focusfor Rural Development

As notedabove(in section2.1.1),oneofthe aimsofthe SRDPin the 1970swasto promoteadministrative decentralisation andintegratedplanning.Thisaspectof the SRDP failed, primarilybecauseofthe lack of commitmentof senioradministratorsin the sectoral (ODI 1992:E.l). In the early1980s, ministriesto decentralised decision-making however,the governmentembarkedupon a new andapparentlyradicalstrategydesigned to shift the responsibilityfor planningandimplementingrural developmentfrom the headquarters of ministriesto the districts.Thisis the policyknownasthe District Focus for RuralDevelopment', thetitle of the document(the so-called"BlueBook") in whichit is codified(first issuedin 1982andrevisedin 1987).Thenewpolicybecameofficially in 1983. ooerational

TheDistrictFocusfocusesuponthe creationandoperationofa hierarchyofDevelopment Committeeswhich is supposedto transmitdevelopmentproposalsupwardsfrom sublocationalto district level. In theory,then,developmentplanningbeginsat the local level (LDCsandSLDCs). As in theLocationalandSubJocational Development Committees stipulatedin the DistrictFocustslueBook' (Republicof Kenya1987)eachSLDCis to be chairedby the AssistantChief andits membersareto includethe local KANU chairman, councillors, departmentalofficers and headmasters of primary schoolsin the sublocation. The core compositionof the LDC is basicallythe same,exceptat a higher level. It is chairedby the Chief of the location and its membersare to include the relevantAssistantChiefs, the KANU locationalchairman,councillors, departmental officers, local representatives of parastatalsand headmasters of secondaryschoolsin the area. Both LDCs and SLDCs are also to include cooptedlocal leadersand representatives of cooperatives, NGOsandself-helpgroups. The'Blue Book'further

18

statesthat women's organisationsmustbe adequatelyrepresentedin the LDCs and SLDCS.

Developmentproposalsare, in theory, passedup from the SLDCs and LDCs and vetted at eachstep, through the similarly constitutedDivisional DevelopmentCommittees (DvDCs) and on to the District DevelopmentCommittee(DDC). The DDC is the most importantofthe institutionsthroughwhichthe DistrictFocusstrategyis applied.It is a largebody chairedby the District Commissioner,with the District DevelopmentOfficer as its secretary,andotherwisecomprisingall district headsof department,Membersof Parliament,district KANU chairmen,local authoritychairmen,DvDC chairmen(who are parastatals,andthe invited of development-related the District Officers),representatives representatives ofNGOs andself-helpgroups. The DDC meetsquarterlyandis assisted by theDistrictExecutiveCommittee@EC), alsochairedby the DistrictCommissioner, but limited in membershipto governmentofficials. The DEC in turn is servedby the District PlanningUnit (DP[D, led by the Distria DevelopmentOfficer andincludingthe District StatisticalOfficer andAssistantDistrict DevelopmentOfficers. Oneof the principaltasksofthesedistrict-level bodiesis to producefive-yearlyDistrictDevelopment Plans@DPs), linkedto annualbudgetsby an annuallyupdatedDistrict Annex: this is supposedto provideboth a work planfor the implementationofthe DDP during the year anddetailsofthe budgetaryprovisionrequired(for a more detailedaccountseeODI 1992:8.5-6).

This looks all very well on paper,but doesit work in practice?Althoughthe District Focusstrategyhadundoubtedlyfocusedattentionon the districtsasplanningand administrativeunits, it hasachievedrelativelylittle in the way of decentralising developmentplanning. Gven the prevailingscarcityof public sectorresources,the DDCs havealmostno firndsto disburseandtherefore,asthe ODI consultancyteamnotes, 19

districts "continuewith routinesof planningthat areintendedto influenceline ministry allocationoffunds at districtlevel(butbarelydo so) andto preparedistrictprojectsthat of havevery little hopeof beingfunded"(1992:E 8). At the sametimethe relationship is both localauthorities andNGOsto theDDCsremainsuneasy:theirrepresentation frequentlylimited andthey often seethe DDCs asattemptingto control themandthereby mandates.This impressionis confirmedby the fact that subverttheir own, independent, the DDCs arealmostwholly managedby the administrationandother seniorgovernment officials,boththroughthe influenceofthe DECsandasa resultoftheir own composition.

As a mechanismfor participatoryplanning,the DDC systemis seriouslyflawed: "Despite ofthe theBlueBook'sespousal ofparticipatoryandbottom-upplanning,the essence systemit describesis for requeststo travel up the system,with decisionsbeingtransmitted Theimplicitplanningmodelis oneof centralisation andcontrol"(ODI 1992: downwards... E8)

Even if the systemdid work in the oppositedirection, tlere are no guaranteesthat it would meetthe participatoryideal. While the District Focusprescribesa mechanism for communityparticipation, it doesnot ensurethat suchparticipationwill be a regular and integral part of the planning process. LDC and SLDC membersare left free to decidehow and when and who is cooptedonto their committees. This leadsto considerablevariation in practice, ranging from adequateto minimal community and women'sparticipationin the LDCs and SLDCs. To give but one examplefrom Mbeere: in 1992Mbita SLDC had nine membersincluding the Assisant Chief, but noneof thesewere women, althoughsomeof the men did representa variety of community interests. This SLDC only met irregularly at the requestof its chairman, and then it was, asone memberdescribed,to draw up requestsfor moneywhich were rarely granted.

As a result there is considerabledissatisfactionin somelocal communitieswith the existing system,which relies heavily upon the characterand actionsof individual chiefs in the absenceof any formal mechanismfor communityrepresentationor involvement. The revisededition of the District Focus 'Blue Book' (Republicof Kenya 1987) recognisesthat the LDCs and SLDCs are not sufficiently active in all districts. It ascribesthis to the fact that their personnelare not equippedwith basic skills in project planning, monitoring, and the preparationof detailedrninutesso that they can communicateeffectively with the DvDCs and DDCs. The suggestedremedyis appropriatetraining. However, in the absenceof funds and with their requests generallymeetingno or only negativeresponsesfrom above,it is unlikely that this would remedymuch(compareWalsh 1992:3).

2.1.2.2 Political Agendas

The District Focusstrategyand its implementation,indeedthe developmentprocessas a whole, is bestunderstoodin its wider political context (seeBarkanand Chege1989). Different actorsacting at different levels havedifferent, sometimesconflicting, political agendas,and we will not attemptto describeall of thesehere. A few main points can, however,be made.

So far we havereferred to the Kenyanstateas a monolithic entity, and implied that developmentplanning is virtually the solepreserveof the governmentand ministries in Nairobi. The reality is somewhatmore complicated. Although political power is wielded from the centre- by the President,his chief ministers,and other close associates- the ethnic fragmentationof Kenyaand the absenceof a strong and allpervasiveapparatusof coercion(Kenyais not a military state,nor doesit rely heavily 21

upon a secretpolice force) hasproduceda much finer political balance. The key actors in this balanceare regional power-brokers,operatingat district or wider level (as determinedby the ethnic compositionand alliancesin a particular region) to deliver political supportand legitimacy (most obviously in the form of votes)to the cenfe. In return they are awardedpostsof varying importancein the government- hencethe burgeoningnumberof ministriesand parastatalsin recentyears- while they in turn are expectedby their local supporters(including other, intermediary,power-brokers)to translatetheir influenceinto developmentprojectsand interventionsin their home regions.

This processis active tlroughout Kenya. It is evident, for example,in Meru District, where the seniorpolitician for many yearshasbeenknown nation-wideas "the King of Meru". It is also evidentin Embu District, wherepower haslikewise beenbrokered for many yearsby a singlepolitician, servingas a senior Minister in a successionof different ministriesuntil his retirementat the 1992election. He hasremained, however, district KANU chairman,while his son, althougha memberof anotherparty (the DemocraticParty, DP), has succeededhim as a Memberof Parliament. His role in bringing developmentto different pars of Embu District, including Mbeere(where his family originally camefrom), is legendary: particular interventionsbeing often linked to whicheverministry he happenedto be in chargeof at the time. Storiesof his attempts,usually successfrrl,to frustratethe developmentinitiatives of his rivals are also legion. Developmentprojectsthereforebecomegifts in the handsof politicians, gifts to their supportersin reward or return for their support.

This systemprovidesthe District Focusstrategywith its political logic. From this point of view the decentralisationof developmentplanning, or at leastsomeaspectsof it, to the districts was not primarily intendedto foster communityparticipation and

makethe processmore efFrcient,but was an attemptto rationalisethe systemin which political power and developmentprojectsare regularly tradedfor one another. It is not surprising, therefore,that the DDCs are the District Focus,that the lower levels in the chain of committees(the so-calledsub-DDCs)are largely ineffective, and community participationin the whole processan empty promise.

The District Focusstrategyalso hasother benefitsfor the government. The rhetoric of decenhalisationand participationhasattractedthe attentionand approvalof Kenya's donors, and to this extent madethem more willing to provide financial assistanceand supportthan they might havedoneotherwise(especiallyin an era when the catchwords are good-governance,democratisationand accountability). The ODA's funding of consultanciesto draw up proposalsfor District Supportprogrammesin Isiolo (separately)and Embu, Meru and Tharaka-NithiDistricts can be interpretedas one of the preliminary fruits of this. The District Focusalso providesthe governmentwith a meansof monitoring and to someextentcontrolling the developmentactivities of NGos. All throughthe 1980sthe governmentexpressedincreasingconcernabout the proliferation of NGOs and its lack of control over them, accusingsomeNGOs of possessingpolitical objectivesin conflict with its own. This concernintensifiedduring the campaigl for a multi-party system- a campaignwhich directly involved a number of churchorganisations- and generateda numberof attemptsto regulateand monitor the NGO community, for examplethroughthe introductionof strict registration requirements. The District Focussystemgives the govemmentan opporhrnityto monitor and channelthe activities of NGOs at district and local levels, though any attemptat exercisinggreatercontrol is ineviAbly underminedby the willingnessof local politicians and power-brokersto contracttheir own allianceswith NGOs.

ZJ

2.1.2.3 The Emperor'sNew Clothes

Kenya'sbilateraldonorshavea very differentinterestin the District Focusstrategyandthe waysin which it is implemented.While awareof the fact that the systemis subjectto political manipulation,the donor agenciestend to view this asa form ofunwarrantedand unwantedinterferencewhich detractsfrom the properobjectivesofthe strategyand rendersthe systemlessefficientthanit couldbe. Theseobjectivesare,of course,those whicharestatedin theDistrictFocustslueBook' andotherpolicydocuments (seesection 2.1-2.1above),objectives whichechothe overtpurposes of colonialcommunity policy- Justastherewasmoreto colonialpolicythanwasgivenpublic development expression,so the modernagenciescarrytheir own political agendas,couchedin the rhetoricofthe "New World Order". Theirprincipalunderlyingobjectiveis, arguably,to recreateKenya(and other countrieslike it) in our own image,so that the development processcanbe controlledandmarketsopenedup without the needfor continuedlargeand (to us) economicallyinefficientpay-offsto nationalpoliticiansandlocal power-brokers_

On this interpretation,then,the grandsubtextof developmentpolicy is a strugglebetween differentinterests,Kenyanversusintemational,for a greatersliceofthe economiccake. Thesupremeironyofthis situationis that both sidesshoulddresstheir strugglein the languageof communitydevelopment,when,if it takesplaceat all, suchdevelopmentis no morethana by-productofthe widercontest.

Thisis not to saythat thesearethe conscious intentionsof all or evenmostofthe individualactorsinvolved. As a descriptionofthe outcomeof individualactionsit does, however,providean alternativeto the orthodoxaccountwhich acceptspolicy statements at theirfacevalue. Accordingto this account,whichis particularlywell articulatedin the donor community,the currentfashionfor communitydevelopmentandlocal participation

in the development processhasevolvedin response to the widespread failureof technically-oriented aid projectswhich havepaid relativelyattentionto the statedneeds andwishesofthe peopletheseprojectsaffect,whetherasbeneficiaries or asthe subjects of other kinds of impact. This provides,for example,the statedrationalefor recent in the directionofBritish aid (proposedaswell aspast)to Embuand changes neighbouring districts.Therestofthis sectionis devotedto a discussion ofthis case, which affectsMbeeredirectly.

In 1982the British government,working throughODd beganto implementa large projectin Embu,Meru.andIsioloDistrictswhichwasdesigned asa pilot technical approachto naturalresourcedevelopmentin arid and semi-aridlandareas. the EMI ASAL Project. The projectconsistedof five maincomponents(includingsoil andwater conservation,smallstockbreedingandtree planting)which requiredreportingto four differentministriesin the Kenyangoveflrment.EMI wasmanagedby expatriatestaff working togetherwith Kenyancounterpartsandlastedthroughtill l99l .

Despitehavingsomelocalisedimpacts,EMI wasgenerally judged,at leastby ODd to havebeena failure. The termsofreferencedrawnup for a subsequent projectpreparation missionlist a wholeseriesof problemsandthe reasons for these: "Whilehighprioritywas givento achievingbetterunderstanding ofthe socio-economicfactorsin the ASAL areas, this waslargelyneglected duringimplementation...At best,the beneficialimpacton ASAL households wasmodest...[the] objectiveof strengthening the institutionsresponsible for planning,implementingandfacilitatingthe socialandeconomicdevelopmentof ASAL peoplewasnot realised... wasexacerbated by the failureto perceive [poor coordination] the overallneedsoffarmersandtheirproductionsystems in an integratedmanner..."(ODI 1992:L 1-2).

25

EMI was originallydesignedto work from the provinciallevel downwardsand partly asa resultof this failedto adaptto the newlydecentralised system:"SinceEMI wasoriginally conceived,therehasbeena major shift in Govemmentpolicy affectinglocal level development, enshrined in theDistrictFocusfor RuralDevelopment Policy...and adopted in 1983. Thispolicyenvisages muchmorelocalparticipation in the development process with local communitiesbeinglargelyresponsiblefor their own development.It is intended thatthe role ofthe Government shouldbeto facilitatethe processofsocialandeconomic developmentby creatingan enablingenvironmentwhereconstraintsto developmentare removedandopportunitiesarecreated.EMI projectagreements were not revisedto take properaccountofthis policy changenor for the creationofnational research institutions...which weregivenmandates (ODI of directrelevance to the Programme" 1992:L.2\.

In linewith this critiqueEMI waswounddownin mid-1991.In late l99l a teamof consultants wasdespatched to IsioloDistrictto drawup, amongotherthings,proposals for a newdistrictsupportprogramme there. In 1992a secondteamwascommissioned to do the samefor Embu,Meru andthe newly-createdTharaka-NithiDistricts. The termsof referencefor this secondconsultancymarkeda definiteshift ofapproachconsistentwith the statedobjectivesofthe DistrictFocusstrategy:"Thestudywill designapproaches to strengthenDistrict and sub-Districtcapacitiesin resourceplanningandmanagement in EmbuandMeru. The aim will be to developreplicablemodelswhich focus on increasing the efficiencyofuse ofthe resources for economicandsocialdevelopment available in the Districts. Proposalswill be madein the contextof a process-projectwith participatory planningasa key objective. To promoteandencourageinitiativesaimedat achievingthe foregoingmodestcapitalaid maybe madeavailable.Particularattentionwill be givento meetingthe needsofthe poorest,the specialrole ofwomen,andthe conservation ofthe naturalenvironment"(ODI 1992:L l).

Theresultingproposalsmarkaninteresting returnto the policiesadoptedby theBritish colonialadministration morethan40 yearsago,althoughthis fact is not madeexplicit,and perhapsnot evenrecognisedby manyofthe consultantsandODA personnelinvolved (thereis, however,a real link, in that somesenioradvisorswere onceemployedby the colonialadministrationin Kenya,andalsoretainedby the govemmentafter independence in 1963). Alongwith variousrecommendations for improvingthe planningprocessat district level, particularattentionis paidto strengtheningandincreasingthe involvementof thelocalauthorities- theMunicipal,Town andCountyCouncils(CCs)- andespecially the CommunityDevelopmentDepartmentswithin the latter. At present,andasa legacy from the colonialperiod,the CCsareresponsible for palng the salaries of Social DevelopmentAssistants(SDAs), but not for their management andsupervision,which are the tasksof the Departmentof SocialServicesin the Ministry of Cultureand Social Services.It is proposedto shiftall of theseresponsibilities to the Community DevelopmentDepartmentsfor an experimentalperiod,aswell asmakingother provisions (suchasa grantto the CCsfor the purchase ofbicyclesandmotorcycles)to improvethe efficiencyoftheir work. It is alsoproposedto givea grantto the CCsto enablethemto assistwomen'sand self-helpgroupstkough a SmallProjectsFund.

Th€ SDASare normally recruitedfrom the communitiesin which they work- They are expectedto work closely with local leadersin the identification, planning, monitoring and evaluationof developmentprojectsfrom the sub-locationalup to divisional levels. They are also expectedto act as secretariesto the different local development committeesand sub-committees(specificallyWomen's and Social DevelopmentSubcommittees)and one of their major tasksin this contextis to ensurethe coordinationof developmentactivities to avoid duplication. However, SDAs rarely havethe resources or the necessaryfraining to carry out all of theseactivities effectively. The ODI report thereforealsoproposedthat they shouldbe given training appropriateto their roles and

new responsibilities,including assistingin the administrationof the Small projects Fund.

Theseproposalsreproducemanyaspects ofcolonialcommunitydevelopment policyand practice. In commonwith existingpracticethe basicapproachrecommended is still from the top down, the major differencebeinga proposedshift in institutionalresponsibilityfor communitydevelopmentawayfrom the administrationandgovemmentministries (especially theMinistryof culture andSocialServices) andtowardsthe electedcouncilsA similarshift took placein colonialpracticewhenthe responsibilityfor community developmentwastransferredfrom District Officersto officialsappointedand employedby the DistrictCouncils(thesewere,of course,the forerunners of today'sSDAs). Fromthis point of view, communitydevelopmentin Kenyawould seemto be movingin a circle. Moreover,the new proposalsseemhardlyto recognisewhat haschangedin the intervening40 years: not only havepolitical agendasandthe distributionofpower changed,but alsotherehavebeenmanydevelopments at grassrootslevel, includingthe proliferationofinitiatives which the stateandits agentshasonly partiallybeenableto control. Theseinitiativesarethe subjectofthe remainder ofthis paper.

2.2 Grassrootslnitiatives

one striking featureofthe official andorthodoxrhetoricof communitydwelopmentis its failure to recognisethe existenceofgrassrootsinitiativesexceptin so far asthesehave beencapturedby the stateandincorporatedinto the developmentprocess.Women's groupsare recognisedif only becauseofthe strenuousefforts by the govemmentandlocal politiciansto usethemfor their own purposes,thoughtheseeforts havenot entirely succeeded andwomen'sgroupsare currentlyaccordedlesssignificancethan they once

were. Othergrassroots initiativesarealmostentirelyignored:this appliesin particularto rotatingsavingsandcreditassociations andsomeofthe activitieswhicharelinkedto the localchurches andschools.Indeedthereis somedistrustoftheseactivitiesbecause of politicalconnections, their presumed andthis is especially the casewhereNGOsare involved.

2.2.1 Women'sGroups

The women's group movementin Kenyatracesa variety of origins, but its true history beganin the mid- 1960swith the formation of large numbersof groupsby Kikuyu womenin Cenfal Province. Many of thesebeganas mabai groups,functioning like rotating savingsand credit associations(ROSCAs)with the aim of enablingtheir membersto buy nmbai, iron roofing sheets,or to afford other homeimprovements. The local contextin which groupswere formed placeda high premium upon mutual assistanceamongwomen: land and malelabour (becauseof labour migration to the towns) were becomingincreasinglyscarce,while women's agricultural and domestic responsibilitieshad increasedand their accessto cashincomeremainedrestricted. This was againsta backgroundof political supportfor self-helpinitiatives (Harambee)in building the newly-independentstate(seesection1.2.2 above).

with official encouragement,similar groupsbeganto appearelsewherein the country. The governmentfust declaredits commitmentto a women's group programmein 1966. In 1975, at the start of the United Nations' InternationalDecadefor Women, it establisheda Women's Bureauto coordinatethe activities of a nation-wideprogramme. In many respectsthe state'srelationshipto the women's group movementis an ambivalentone, and its assistancecan alsobe interpretedas an attemptto control and

makebestuseof what are, essentially,grassrootsorganisations. This was particularly evidentin the late 1980s,when the governmentintervenedto take over Maendeleoya Wanawake("Women'sProgress')andincorporateit within the women'swing of KANU, then the only political party (Maendeleohad hitherto beenan NG0, especially active in women'sdevelopmentduring the late colonial period, thoughrather less important for tlte women's group movementby the time it was takenover). In different parts of the country women's group memberswere then told that they would haveto join and subscribeto the new Maendeleoif they were to receiveany support llom the government. Ultimately, however,this attemptto gain further control over the women's group movementand its resources- including the funds allocatedto groupsby outsideagenciesand NGOs - failed when KANU was forced to diseneage itself from Maendeleofollowing tlle legalisationof other political parties.

The majority of women's groupsreceivelittle or no assistanceexceptthat provided by the government. They are requiredto registerwith the Departmentof Social Services and are subjectto the various attentionsof its extensionagents,along with agricultural ofFrcers,chiefs and other officials in the local administration. Women's groupsare usedin a numberof ways to promotethe government'sdevelopmentpolicies. Registeredgroupsare eligible to receivegrantstowardstheir projects, though the demandfor theseis much greaterthan the supply. Groupsare also encouragedto hold fund-raisingharambeesfor the samepurpose. Otherwisethey typically raise subscriptionsor sharesfrom their members,and engagein a wide rangeof economic activities, including collective farm labour for payment,in order to gatler seed-money for their projects. As a nrle it is difficult for groupsto obtain commercialcredit or establishlarger enterpriseswithout externalassistance.While most groupsaim to developcommunity servicesor profitable enterprises,they alsoperform a variety of

30

welfare and other functionsfor their own members: for example' by exchanging labour or by operatingROSCAS'

Registeredwomen,sgroupshaveto possessan electedchairwoman,tfeasufer, secretary and committee. within this framework actualdistributionsof authority and organisationalproceduresmay vary considerably. Groupsalso differ considerablyin size and composition,both from one anotherand as they developover time' Men may also belongto women's groups,though in the majority of casesthey are excluded. Groupswith more than five male membershaveto registeras self-help, not women's, groups. In generalzuchgroups,including thoseformed exclusivelyby men, are few and far between.

It is difficult to obtain accuratefigures for numbefsof women's goups and their membership. Thereare a numberof reasonsfor this. while socialdevelopment for assistantsrecord the numberof registeredgtoupsin their reports, it is lesscommon are them (and not really in their interests)to deleteor draw attentionto groupswhich nolongeroperative.Tothisextentofficialfigurestendtobeinflated,thoughthe existenceof unregisteredand unreportedgroupsmay redressthe balancesomewhat. figures Meanwhile, for thosegroupswhich are registered,it is unusualfor membership to be updatedfrom thosereportedat the time of registration. In somecases membershipwill havegrown considefably,in othersthe active membershipwill only be be a portion of the total recorded. For thesereasonsthe reportedfigures haveto treatedwith somecaution.

AccordingtotheWomen'sBureau,in1988therewere26,92|women,sgloupsinthe for country with a total of over one million members(1,053,391). Mbeereaccounts just a small fraction of this total. The information compiledin the following table 3l

showsthat in 1982there were 140registefedwomen's groupsin the whole of lower Embu (Siakagoand Gachokadivisions), I 11 in the Mbeerearea(lower Embu excluding Mwea). The registeredmembershipof 91 of thesegroupstotalled 2,692 persons,a meanof 29.6 memb€rsper group. By 1990the numberof groupsin lower Embu had risen to an estimated200 with a total of 7,517 membersand an overall mean of 37.6 membersper group. While thesefigures offer no more than a rough approximation,they suggestthat as many as one third, possibly more, of all adult women(aged20andover)inlowerEmbubelongtowomen'sgroups.Givensucha level of involvementit is probablethat the majority of womenhavebelongedto a registeredwomen's group at sometime in their lives, evenif they do not do so now.

Welbourn(1990)speculatesthatsocialatrdeconomicchange,includingtheimpactsof landreform,haveplacedalargerburdenuponwomenandthuscfeatedfavourable "More wives havehusbands conditionsfor the spreadof women's groupsin Mbeere: workingaway,theirchildrenareatschool,andtheynolongerhavethemutual farm and traditional sup'portof closekin living nearbyto help them. All the effort of the householdlabour now falls on their own shouldersrather than being sharedwith and rest of the family...This I believeis an importantreasonfor the rapid emergence growthofwomen'sgfoupsinMbeereaselsewhere"(Welbourn1990:38).Reference shouldalsobe madeto the role playedby ROSCAsin the formation of women's groups: indicative, in part, of a demandfor new conzumergods on the part of This women and theif limited accessto incomeand/or other forms of saving. groupsin Central explanationmirrors that given abovefor the rapid qrad of mabati Provinceafterindependence,thoughitdoesnotmentiontheroleofofficialsupport' given to women's groupsas part of the SpecialRural including the encouragement DevelopmentProgramme(SRDP) in Mbeere'

)L

Women's Groups and Their Membership in I-ower Embu' l9E2 and 1990 214) and ODI (1992: l'3) sources: Mwaniki 1d1986:

Year

Administrative Area

t9E2

1990

Siakago Division

53

84

numberof members

1430

2542

meanper group

27.0

30.3

87

116

numberof members

N/A

4975

meanper group

N/A

42.9

numberof groups

Gachoka Division numberof groups

TOTAL LOWEREMBU numberof groups

140

2N

numberof members

N/A

7517

meanper group

N/A

37.6

lll

N/A

numberof members

N/A

N/A

meanper group

N/A

N/A

TOTALMBEERE numbef of groups

JJ

Despitethe evidentimportanceof women's gtoupsin terms of the large numbersof of the women'sgroup movementand the women who join them, many assessments groupsthemselveshavebeennegative. This conclusionhasbeenreachedin studiesof women's groupsin Mbeereas well as elsewherein Kenya. Following a surveyof 25 women's groupsconductedin 1982in three locationsin Mbeere, Mwaniki (1986:22G 225) listed a whole catalogueof internal and externalconstraintsupon their effectiveness. Theseincludedwomen'sheavydomesticand agricultural responsibilities (restricting their contributionsof time and moneyto groups); food shortages,water scarcity and inadequatenutrition (especiallyduring periodsof drought); poor organisationand weak leadership(eading to internal disputesand allegationsof the misappropriationof funds); lack of zupportfrom members'husbands;the failure to identifu viable projects; lack of adequatenumagementpractices,including bookkeepingskills; lack of capital; lack of goodroadsand isolation from markets;the lack of marketsin any eventfor someproducts;lack of trained extensionworkers (Social DevelopmentAssisans) and tlerefore appropriateadvice; and, last but not least, the generalproblem of women's subordinationto men (which is at the root of someof the constraintsmentionedabove). Given sucha long fist it is surprisingthat anyonewould want to belongto a woman'sgroup at all. Mwaniki assertsthat he only saw one project that had demonstratedsornepotentialfor generatingincome: a multi-purpose hall built and ownedby the Union of Kithunthiri Women's Groups, one of two zuch unions includedin his survey.

Welbourn(1990:45-a8)is equallypessimistic,andalsoassertsthat mostwomen's groupsare not successful. The reasonsshegives do not add muchto Mwaniki's list, and include resistanceby husbandsand the exclusionof singleor otherwise troublesomewomen; lack of educationamonggroup membersand a consequentlack of stronggfoup leadersand groupswith clear objectives;failure to identiry viable

projects; and women's lack of time and money. According to Welbourn tlis last constraintsuggeststhat in somecasespoorer womenwill not be able to join groups, althoughMwaniki doesnot provide any supportfor this conclusion(which in some ways contradictsWelbourn's own thesisaboutthe formation of groupsand the type of women likely to form them).

Similar analyseshavebeenproducedin virtually every study of women's groupsin Kenya (seeMcCormacket al. 1986:.10-17for an overview). The pattern is: a catalogueof constraints,longer or shorteras the casemay be, and the conclusionthat the potential for further developingwomen's groups,and especiallytheir capacityfor running successfulincome-generatingenterprises,is limited. Mwaniki concludes: '...there are no easyor immediatesolutionsfor manyproblemsfacing Mbeere women's groupsbecauseunderlying them are broad structuralfactors...In light of these problemsMbeerewomen's groupswould perhapsbe better off if they concentratedon their mutual assistanceactivities. . . " (1986: 225), althoughhe concededthat some income-generatingprojectsmight work. Welbourn is similarly cautiousin recommendingthat the (former) EMI Programmeshouldwork with women's groups: 'Working with women's groupsaloneis not necessarilygoing to give womenaccessto long-term, sustainableincome; it is not necessarilygoing to lighten women's physical and psychologicalburdens;nor will it ensurewomenan opportunity to speakout about their problemsand needsto a sympatheticresponsiveaudience' (1990: 48). These conclusions arealsoechoedin the ODI reporrQ992: 1.2-3).

It is not difficult to reachsuchconclusionsafter scanningthe historiesof individual groups. We collecteddetailedhistoriesof 2l women's groupsin Mavuria Location, formed from 1974onwards(and only a sampleof all the groupsfonned during this period: Mwaniki, for example,cites the existenceof 38 women's groupsin this 35

locationin 1982(1986:214)). 15 of the 2l goups in our samplehavesincecollapsed or haveotherwisebecomedormant, for reasonswhich can all be found in Mwaniki's list. KamacaciWomen's Group, for example,which was formed in Gatakavillage in 1979, disintegratedin 1984following the failure of its plansto constructa rental building / businesspremises(the building stoneshad alreadybeenpurchased). According to someformer membersthis wasbecausethe group's officials had misappropriatedthe funds for the project: on anotheraccountownershipof the plot on which the premiseswere to be built was assertedby the sonsof the man who had donatedit, following which mernbersaccusedthe group's officials of incompetencefor failing to ensurethat no suchproblemswould arise.

What suchaccountslack, however, is a broaderperspectiveon group formation and memb€rshipover time in any one area, as well as an understandingof the underlying reasonsfor group successas well as failure. The formation, developmentand decline of groupsis a dynamicprocess. The fact that groupscontinueto be formed, and that many women, despitehaving belongedto failed groups,continueto experimentwith others, suggeststhat womendo gain somethingfrom group membership. One of these gains may well be the experiencewhich leadsto later success:and it is not unusualto find that the membersand officials of successfulgroupshavecomefrom groupswhich havenot survived. ln any eventthe survival rate of women's groupsis probably not much worse than that of other small enterprises.

A rather different picture emergeswhen successfulgroupsare analysed. In their study of women's groupsin CoastProvince, McCormacket al.(1986) found that the initial successof groupsrestson the extentof their accessto the labour of membersand to the cashprovided by membersor their households.The amountof incomewhich these households(and womenas householdmembers)are preparedto invest in groupsis

conditionedby the sum of demandsupon them, their ability to meetthesedemands, and the returnsthey can expectfrom this as opposedto other investments. Under these circumstancesit is not easyfor groupsworking aloneto establishviable enterprises. In order to overcometheseobstacles,groupsneedallies, and the major allies availableto them are the government(in the shapeof the Departmentof Social Services)and various NGOs. Although there is often a price to be paid for its assisrrnce,the governmentcan provide grantsto capitalisenew enterprises,while the NGOs can also provide valuabletechnicalassistancein the planning and operationof enterprises. Considerableimportancealso attachesto the choiceof enterprise: thosewhich are new to a community, particularly thosewhich are capital and labour intensive, are diffrcult for groupsto operatesuccessfully.

On the basisof thesefindings, the two NGOs working with thesegroups, Tototo Home Industies (of Mombasa)and World EducationInc. (of Boston), designeda programme of training and other assistancewhich firrther increasedthe profitability and sustainabilityof women's group enterprises(seeKane et al. l99l\.

The wider

significanceof this work was that it arguedthat a lot more could be doneto increase the viability of women's groupsand their enterprises. For the most part, however, women's groupshavereceivedlessattentionfrom NGOs and other agenciesthan they did in the nid-1980s. one reasonfor this hasbeenthe negativeassessment offered in most reportsaboutwomen's groups, many of them, it mustbe said, basedupon quick and impressionisticresearch. A secondreasonhasbeenthe generalshift in favour of micro-enterprisecredit, especiallyon a group-lendingmodel (the loansbeing madeto individuals), as a more cost-effectiveway of increasingincomesand employment, fostering economicdevelopment,and alleviating poverty in both rural and urban communities.

J I

A third reasonfor the relative lack of interestin women's groupson the part of developmentpractitionersin both the public and private sectorslies, in a sense,behind the two reasonsalreadygiven. While the stateand other agencieshavebeenkeento captureand usewomen'sgroupsfor their own purposes,both political and economic, they havebeenlessthan willing to engagegroupson their own terms. As a result most researchon women's groupshasbeensuperficial, and hasled to the conclusionthat groupsare not worth any significant investment. Micro-enterprisecredit, on the other hand, offers a numberof advantages.It requiresminimal interactionwith the clients, and developmentis, again, managedfrom the centre,b€ing reducedto the circulation of funds and the simple monitoring of repaymentrates.

2.2.2 Rotating Savingsand Credit Associetions

The nabai groupswhich gaverise to the women's group movementin Kenya were, in effect, rotating savingsand credit associations(ROSCAs)formed in part for the purposeof financing homeimprovements. Even today the majority of registered women's groupsin Mbeerehavegrown up aroundROSCAsof one kind or another, and it is unusualto find a group which doesnot continueto operatea ROSCA in addition to its other activities. ROSCAsare in fact ubiquitousin Mbeere: there are very many more ROSCAsthan registeredwomen's groups,including associations which are in transition betweenthe two and are sometimesdescribedds unregistered women's groups. From this point of view ROSCASand the "informal' groupswhich developout of them are part of an extensiveand fertile substratumof voluntaristic organisationin local communitieswhich the stateand other agencieshave not beenable to manageor otherwisetake advantageof, exceptin so far as they havecaptured,or

38

aftemptedto capture,the women's group movementat the most developedend of the ROSCA-women's groupcontinuum.

ROSCAsin Kenya are usually relegatedto a footnotein the developmentliterature, and we do not know of any concertedattemptsto mobilise their developmentpotential or evenunderstandtheir existing impacts. There is, however, somecomparativeliterature on the phenomenonof ROSCAsworld-wide, in which debatehasfocusedupon establishingtypologiesof ROSCAsand understandingtheir functionsin societyas a whole. In a seminalpaperGeertzdescribedthe ROSCA as "essentiallya deviceby meansof which traditionalistic forms of socialrelationshipare mobilized so as to fulfil non-traditionalisticeconomicfunctions...[t is] an 'intermediate'institution growing up within peasantsocial strucfure,to harmonizeagrarianeconomicpatternswith commercialones,to act as a bridge betweenpeasantand trader attitudestoward money and its uses"Q9A: 242). Following a more extensivesurveyof the literature, Geertz'sconclusionwas questionedby Ardener, pointing to the persistenceof ROSCAs alongsideformal financial institutionsand their absencein many societieswithout commercialinstitutions (19&:221-222). Ardener arguedthat ROSCAs,including someof the most "developed"ones,cannotbe understoodin termsof economicmotive alone,althoughtheir "mostobviousfunction...isthat they assistin small-scale capitalformation, or more simply, they createsavings"(196/:217).

Subsequentanalyseshavenot addedsignificantly to theseconclusions(see,for example,Kurtz 1973; Schraderl99l; and Brusleyet al. 1992). As hasoften turned out to be the cas€in the comparativestudy of institutions, the searchfor a simple typology and globally applicablefunction or functionsis probably doomedfrom the start. Within MbeerealoneROSCAsperform a variety of functions, both social and economic,and evenwhen they evolve into multi-activity women's groupswe cannot 39

say that one set of functions (for example,the economic)hasbecomemore important than the other. "An associationformed Ardener providesthe most succinctdefinition ofa ROSCA as upon a core of participantswho agreeto makeregular contributionsto a fund which is given, in whole or in part, to eachcontributor in rotation" (196/.:2Ol). The two essentialcriteria, accordingto Ardener, are rotation and regularity, to distinguish ROSCAsfrom other mutual benefit clubs and cooperativeundertakings. In Mbeere this definition might b€ stretchedto include cooperativeweedingby women (sometimes helpedby men) when it is performedon a rotational and regular basis(a common pattem is every Saturdayat one of the member'sfarms). This kind of cooperationis quite commonin Mbeereand also often performedas one of the activities of women's groups,though it may also be undertakenby groupsof peoplewho meetsolely for this purpose. The essentialdifference, of course,is that the input to thesework groupsis labour, whereasthe primary input to RoSCAs (as normally understoodand defined here) are cashcontributions. Needlessto say a full accountof RoscAs and their origins in Mbeerewould haveto considerboth typesof association.

ROSCAsin Mbeereare formed by women, men, and mixed groupsof men and women, though as in the caseof registeredgroupswomen-onlyRoSCAs seemto be by far the most common. sometimesthey are formed by groupsof closely relatedkin, but the vast majority of RoscAs are composedof small groupsof neighbours,someof whom nray be closely relatedbut othersnot. A few ROSCAsare formed on an occupationalrather than a neighbourhoodbasis: markettraders,for example,often form their own associationsin which the object is to assistin the purchaseof capital inputs (usuallypetty commodities)to their businesses.As notedabove,the proliferation and sheernumbersof RoscAs in Mbeereare striking. Just over half

(11/20)of a smallsampleof womenin Kamugu,nearSiakago,wereactivemembersof while ROSCAs,someof them(2i ll) beingmembersof two ROSCAssimultaneously, other women (3/20) who were not current membershad belongedto ROSCAsin the past: making a total ofjust underthree-quartersof the sample(14i20) who declared past or presentparticipationin ROSCAs. A much smallerproportion of thesewomen, one-fifth to be exact, were pastor presentmembersof regisGredwomen's groups. Thesewomenmay well haveunderreportedtheir pastmembershipof ROSCAs. A smaller sampleof 12 womenin the Ishiaraareashowedthat all of them had belonged to ROSCAsat one time or another(7 of them currently and the remainderin the past). Data collectedin the courseofa householdsurveyh 1992suggestthat comparable levels of involvementcan be extrapolatedfor Mbeereas a whole.

ROSCAsin Mbeere,as throughoutrural Kenya, conform to the most basicpattern describedin the literature (both Geertz(1962) and Ardener (1964) presentinformation on a rangeof associationswhich are organisationallycomplexand extremely sophisticatedfinancially). Most ROSCAson which we haYeinformation havebetween l0 and 20 members,usually closerto the lower figure. Membersme€tto contribute fixed amountsof cashweekly or monthly: contributions(calculatedon a montl y basis)uzually fall in the rangeof Kshs 10 to IShs 120per month, this last example coming from an associationof men and womenin Ishiara marketwhosemembersgive IShs 30 eachper week. The sum of contributionsin any we€k or month is usually presentedto just one of the members- In many casesthis moneyis alreadyearmarked for a particular purpose. Women's ROSCAsoften savein order to enablemembersto purchasecrockery and cutlery for their households.The exampleof market traders buying inputs for their enterpriseshasalreadybeenmentionedabove. Other recorded ROSCA objectivesinclude the purchaseof farm equipmentand (separately)goas.

4l

ROSCAssuffer from someof the sameconstraintsas women's groups,though perhaps not to the samedegree. When hard-pressedfor cashmembersmay haveno option but to drop out, therebyproducingcomplicationsin terms of contributionsowed. When more than one memberdropsout this usually sp€llsinstantdoom for the association, which may colapse in the midst of heatedrecriminations. However, the available evidencesuggeststhat this is not a commonoccurrence: peer group pressureand the mutual aid ethosengenderedby ROSCAstend to insulatethem againstmany stresses, while unlike women's groupsthey operatewith fewer membersand on a fxed and relatively short time cycle. ROSCAsare most likely to collapsein times of general stress,for exampleduring periodsof food shortageand famine, and we havesome evidencefor this happening. On the other hand, the facton which makeROSCAsquite resilient, including tieir modestsize, degreeof organisationrequiredand short time cycles, makeit easyfor them to reform following suchcrises, whetherwith a different membershipor not.

In the absenceof comprehensivestatisticsit is difficult to assessthe overall impact of ROSCAsin Mbeere, or anywhereelsein Kenya for that matter. We know for certain that they play a critical role in the developmentof women's groups. It is also evident that they play a very important role in the mobilisationof savings,as well as in tle creationof local networksof mutual support. The initial successelsewherein Kenya of micro-enterprisecredit programmesbasedupon group guarante€s(the Grame€nBank model imported by the Kenya Rural EnterpriseProgrammeand PRIDE) undoubtedly owes somethingto the samefactors. Ironically theseprogr:rmmeswere introduced without referenceto or evenan understandingof existing savingsand credit systems, including the ROSCAs,in rural and urban communities. In a sensecommunitieslike thosein Mbeerealreadyhavetheir own indigenouscredit programmes,the difference

being that they are not directedfrom aboveand that they are generallytailored to more modestdomesticrequirements.

2,2.3 Schoolsand Churches

Furtherindicationofthe dyramismofgrassrootsinitiativesis apparentin the development anddevelopmentactivitiesassociatedwith largerlocal institutions,particularlyschools andchurches.whereasthe historyof RoSCAsandotherformsof everydayassociation is almostinvisible,at leastfrom the state'sanda districtpointofview, the historyofthese largerandmore permanentinstitutionsis alsomoretransparent,andcanbe readilytraced backto the earlycolonialperiod. Theinitialestablishment of both schoolsandchurches was,ofcourse,a directconsequence of externalinterventions duringthe colonialperiod, interventions by boththe administration (in the caseofschools)andmissionary organisations (in the caseofboth churches andschools).Thereafter, however,andwith increasingpopulardemandfor education(to be translatedinto materialwelfare)andthe spiritualwelfareprovidedby the christian churches(the two were often intimatelylinked), manyofthese institutionshavetakenon a life oftheir own, with considerableinputsfrom the communities in whichtheyarelocated.

Primaryresponsibilityfor Kenya'seducationsystem- includingeducationpolicy, the schoolcurriculum,the provisionof educational facilities,the inspectionof schoolsand paymentofteachers'salaries- restswith the governmentandthe Ministry ofEducation in particular. However,the demandfor educationsinceindependence hasfar outstrippedthe state'scapacityto supplyall ofthe necessary facilities,especially schoolbuildings, classroomsandthe fittings andmaterialequipmentwhich they require. As a resultthe govemmenthasmorethan welcomedthe contributionsmadebv local communitiesto 43

buildingandequippingschools.Thisencouragement of self-helpinitiativeshasledto the construction of manyschoolssinceindependence, includingthe ',Harambee,' secondary schoolsaswell asmoremodestprimaryfacilities(seesection1.2.2above).Although MbeerehaslaggedbehindupperEmbuandother high potentialareasin the demandfor education, therehasbeena considerable proportionalincrease in the numberofschools constructedin the areaover the pasttwo decadesandmore (for an overviewofeducation in Mbeere,includingits earlierhistory,seeRileyandBrokensha1988:25,301-308)

The constructionanddevelopmentof GatakaPrimaryschool, nearKiritiri, providesan illustrationofthe kind oflocal initiativewhich hasbeenreproducedthroughoutMbeere. In 1973a groupof localresidents, all ofthem men,formeda committeeto pressfor permissionfrom the govemmentto constructa primaryschool. After muchlobbyingand waiting, a permitto do so was issuedthe following year,in 1974. Thereuponcommittee members beganthe difficulttaskof tryingto mobiliselocalpeopleto contributecashand their own labourtowardsthe constructionof the school. In the endcommitteemembers beganto build a nurseryclassroomthemselves, usinglocal materials(earthandtimbers). This work was intemrptedby shortrains,wheneveryoneturnedto the task of cultivating their farms. In early 1975the committeebeganto run a nurseryclassin a nearbychurch building,the teacherbeingpaid from contributionsby the children'sparents. This still left the problemofwhere they would be taughtwhena secondintakewas admittedandthey hadmoveduDto standardone.

Later on in 1975,anothergroupoflocal men,youngerthanthe committeemembers, seizedthe initiative andcompletedthe nurseryclassroomwhich hadbeenbegunthe year before. In order to raisethe fundsfor buildinga secondclassroomthey organiseddiscos everyevening,chargingan entrancefee. At the sametime they alsobeganto cut timber for the building mobilisingfriendsto helpthemwhenthey could. Incomefrom the disco

paidfor iron roofingsheetsandthe construction ofthe classroom aswell asa smaller buildingto serveasa staff-room.Thisgroupofyoung menthensteppedasideto let the committeecontinuewith its work.

In 1976theMinistryofEducationpostedtwo teachers to the school,whichnow consisted of a nurseryclassandstandardone. Thereafterat leastoneclassroomhadto be built everyyearto accommodate the school'sannualintakeofpupils. Fromthis point on the heaviestburdenfell upon the pupils'parents,who providedcommunallabour in addition to makingregularcontributionsofmoney to pay for materialsandthe input of professional builders.Not all of the classrooms werecompletedon time,in whichcase someclasseshadto be dividedinto morningandaftemoonshifts. The samesolutionwas alsoappliedduring a periodwhenthe schoolwasunderstaffedandwaiting for more teachers to be postedthere.

By 1985a completesa of eightearth-and-timber-walled classrooms hadbeenbuilt. The schoolcommitteethendecidedto embarkon the construction of stone-wallclassrooms to replacethem. Parentsthereforecontinuedto contributeto a building fund everytime, and still do so today. In 1988the Embu-based NGO plan Intemationalbuilt two new classrooms for the school: parentscontributedKshs18,000to paythebuilderandalso dug the foundationsandcarriedwater to the site. plan alsoerecteda largewater tank in 1990,againassisted by the parents.In 1991the schoolwasgivenKshs 175,000by plan to build anotherpair of classrooms. The schoolcommitteerequestedthis assistance andparentsraisedKshs24,850to pay the builder. constructionofthe classrooms is still underway,and somearein usealthoughthey do not haveanywindowsyet.

In 1991the schoolcommitteedecidedto introducecompulsoryboardingfor standard eight students,reasoningthat they would thereforebe ableto spendmoretime studying 45

andperformbetterin their final KCPE (Kenyacertificate of primary Education)exams. Boardingbeganin March 1991andthe boarderswereaskedto contributesetquantitiesof maizeandlegumes,to providetheir own beds,mattresses, blanketsandboxes,andto pay a feeof overKshs100to coverthe salaryofa cookandthe costofvariousingredients. This went well until the next year, 1992,whenthe failure of manyparents'to pay these feesledto the suspension ofthe boardingfacility. our latestinformationis that no decisionhasyet beenmadeaboutits future.

Thishistory,whichcouldbe repeated for manyotherschoolsin Mbeere,indicatesthe strengthandversatilityofgrassrootsinitiative.All ofthe majordecisions aboutthe developmentofthe schoolhavebeentakenby communitymembersworking, after they had startedit, togetherwith the headmaster andteachers(manyofwhom areor have becomecommunitymembersthemselves).The schoolcommitteehasalsochangedand developed overtime: whileit beganassmallgroupofmen andelders,in late 1992it had twelve members,a third of themwomen,andmost of themlocal farmersand,of course, parents.

Many churchesin Mbeereare similarlyfoundedin local initiative andeffort. Unlike schools,though,they areformally separatefrom the state,which guaranteesfreedomof worship in its constitution. In fact this separationis so sharpthat the nationalchurches, especiallythoseaffiliatedto the NCCK (Nationalchristian council of Kenya),areoften perceivedasposinga threatto the government.This hasbeenparticularlyclearin the past decade,with the NCCK spearheading the campaignfor the introductionof multi-party democracyandactingin somewaysasthe unofficialoppositionto the KANU govemm€nt. This doesnot meanthat the local churchesareactivein nationalparty politics,thoughlike all communityinstitutionsthey arefrequentlysubjectto local political conflictsandmachinations.The separationof churchfrom statedoesmearLhowever.that

they presenta muchmorevariedanddyramic picturethanthe schools,which all conform to a basicpattern.

Kenyais hometo a bewilderingvarietyof Christiandenominations andsects,someof themgrown on its own soil. Mbeeresharesin someof this diversity. Active churchesin theKiritiri area,for example,includethe CPK (Churchofthe provinceofKenya,or Anglicanchurch), the Romancatholic church, the East African pentecostalchurch, the Full GospelChurch,theFull GospelAnglicanChurch(sic),andthe SeventhDay Adventists. The mostprominentchurchesin Mbeere,at leastto an outsideobserver,are the Anglican(cPK) andRomancatholic churches.This is becausethey were the first to becomeestablished, are part of wider national(andinternational)congregations,andas a resultare ableto supporta numberof developmentactivities. The cpK doesthis in part througha sisterorganisatioqcompassionIntemational,which hasassisteda numberof schoolsin SiakagoDMsion. TheRomancatholicchurchis particularlyactivethroughits separate missions, whichincludetheDon BoscoSalesians in Gachokaandthe consolata Fathers(and Sisters)in Ishiara. The missionin Ishiar4 to give but one example,has sponsored manylocalprojects,includingthe construction of primaryschoolsand classrooms,a hospitaland(by bringingin Italian governmentaid) the Nationalcerealsand ProduceBoard (NCPB) depotandgrain storagefacility at Ishiara.

At grassrootslevel, andawayfrom the influenceof foreignmissions,local initiativesare as evidentin churchorganisationandactivitiesasthey arein the schools. Therearemany parallelswith the buildinganddevelopment ofschools,asdescribed in our earlierexample. church committeesand(in the caseofthe cPK) parishcouncils canbe asactiveand innovativeasthe schoolcommittees,ifnot more so. Theymayalsobe evenmore representativeof local interests:the committeeof Nguru cpK church in Gatakavillage, nearKiritiri, has,for example,I I members,5 of themwomen,all of themmembersof the 47

local farmingcommunity(a degreeof representation far more impressivethan that on any committees).Thelocalchurches alsocoordinateother ofthe localdevelopment communityactMties:theseincludeyouthclubs,localbranches of theMothers'Union (which function muchlike women'sgroups)andvariousforms of assistance to the poor andneedy.

2.3 The Role of NGOs

The largerchurchdevelopmentorganisationsandprojectsreferredto abovefall into the categoryofNGOs or non-government organisations.The term NGO is a catch-allfor privatevoluntaryorganisationsof all shapesandsizesabovethe local communitylevel, rangingin a continuumfrom small,indigenous,bodiesup to large,international,agencies like Oxfam. Kenyahasproveda very fertile groundfor NGOs over the pasttwo decades, andthey haveproliferatedacrossthe country,the vastmajority ofthem operatingfrom headofficesin Nairobi. TheNGO communityasa whole standsin an ambivalentrelation in ODI I 992:G.l). On the onehandNGOshave to the state(comparethe discussion beenwelcomedby the govemmentandlocal politiciansfor the innumerableprojectsthey undertakeandthe resourcesthey attract from donor agencies.On the other handthey havebeensuspectedfrom time to time of working to underminethe govemment,a suspicionenhancedby knowledgeof the closerelationswhich existbetweensomeNGOs andthe churches in theNCCK. Thegovernment hasthereforeactedon a numberof occasionsto strengthenits control over both local andforeignNGOs, andthis, aswe have seen,wasoneofthe objectivesofthe DistrictFocusstratery(seesection2.1.2.I above). Judgingby the government'scontinueduneaseaboutNGO activitiesit would appearnot to haveachievedthe degreeof control which it wishes. At the sametime it would be fair

48

to saythat manyNGos, indigenous andinternational, aretreadingverywarilyin orderto avoidfallingfoul ofthe government andlocaladministration.

In additionto the churches andmissionprojects,a numberofNGOs operateor have operatedin Mbeere.TheseincludePlanInternational, CAREKenya,theKenyaFreedom From Hungercampaign,the BelleriveFoundation,andHeifer International,to namebut a few. The mostextensiveprograrnmeis run by PlanInternational(Fosterparentsplan Internationalin full), an AmericanNGo which hasits Kenyanheadquarters in Embutown. Thedecisionto siteits officesin Embu,andnot Nairobi,is generallyascribedto successful lobblng on the part ofEmbu's principalpower-brokerandgovernmentminister_ whatever the case,the rangeofactivities pursuedby plan andits staff sincethe early 1980sis very impressive.In lowerEmbutheseactivitieshavebeenconcentrated in GachokaDivision, thoughthereareplansto expandinto SiakagoDivision. plan'soriginal andcoreactivityis childsponsorship, on the basisofwhich it hasprovidedassistance of differentkindsto schools(hencethe constructionof classroomsandwater storage facilitiesreferredto earlier)andworked in numerousother fields: the improvementof water supplies,nutrition, healthcare,agriculturalpractices(includingpestcontrol and composting), andthe provisionof faminereliefinthe form ofa "food-for-work', programme(diggingterracesandbunds).

Many NGo interventions,includingsomeof thosementionedabove,adoptthe technical, top-dow4 approachwhich characterises aid projectsin general. In other respects, however,their approachis moregenuinelyparticipatory. This is achievedin different ways. NGos typically employstaff andfieldworkersfrom the communitiesin which they work: indeedthe bestindigenousNGOs arerun by local people(thoughexamplesof this arefew andfar between,andtherearecertainlynonein Mbeere: for a caseelsewherein Kenyaseewalsh 1989) Anotherstrategyis to setup localcommittees aspart ofthe 49

frameworkof the NGos operations:Planhasdonethis with the creationof its own sublocationaldevelopment committees, aswell asin the contextofparticularprojects(for exampleby institutinglocal boreholecommitteesto supervisetheir maintenance). MoreoverNGOs like Planareusuallyvery responsiveto requestscomingfrom the communities in whichtheywork: hencePlan'spositiveresponse to the requestsofthe GatakaPrimarySchoolcommitteefor assistance in buildingnewclassrooms.The scale ofNGo activitiesalsomakesthemmoreflexible andcapableof both experimentingwith newapproaches andadaptingto changingcircumstances. Thediversification ofplan,s programmesin GachokaDMsion andtheir rapidresponseto local famineconditions provideillustrations of this.

while it wouldnot be difficultto producea list of NGo mistakes, the pointis that these arelesslikely to occur andmore likely to be correctedin the contextofNGo prograrnmes thanin that of otherkindsofaid project. Fromthis pointofview NGOsareideally situatedto mediatebetweenthe grassrootsandother agencies,includingthe govemment and donor organisations.certainly they arein a muchbetterpositionto foster initiatives from the grassrootsthanthe formal structuresfavouredby the governmentand,in modifiedfonn, in recentproposalsto the ODA. The ODI report, while acknowledgrng the role that NGos play in promotingsmall-scaleprojectsat villagelevel, criticisedthe lack of coordinationbetweenNGos andpointedto the presumedlimitationsof NGos in handlinglargerprojects: "They do not havethe resourcesor the mandatesto tackle projectswhich crosslocationboundaries,for example,majorwater projects,roads,etc. Engineering resources andbusiness adviceareareasof weakness" (1992:G.9). Thisis clearlynonsense:manyNGos, includingPlanInternational,arealreadyworking across location (anddivision,anddistrict) boundaries,manyNGos havewell-developedbusiness programmes;and,giventhat suchjobsareusuallycarriedout by privatecontractors,there is no needto involveNGOs in large-scaleengineeringworks 50

Having presentedthis rather deficient assessment of NGO capacity,the ODI report concludesas follows: 'Apart ilom somehelp if the NGOs decideto form a district associationand network, it is not recommendedthat ODA channelproject funds, aimed at improvedplanning and management,to the NGOs" (192: G.9). The alternative suggestedby ODI, as we haveseen,providesevenlessopportunity for the promotion of grassrootsinitiatives.

3 Conclusions

The conclusionsreachedin this paperarerelativelystraightforward,thoughthey do providea challengeto a largebody of receivedwisdom. Communitydevelopmentand its analogues,ascurrentlypromotedby developmentpractitionersboth insideandoutsidethe aid agencies,is at bestpoorly conceivedandat worst a sham. The claimthat a community-based approachhasevolvedsimplyin response to the failureoftop-down projectsis underminedby a considerationofthe very real continuities technically-oriented that exist betweencontemporaryprescriptionsandthe communitydevelopmentpolicy of the colonialperiod. In both casesthe rhetoricworks to maskanothervariationon the topdown approachandthe political agendasof communitydevelopment's different proponents.One ofthe governmentrs agendas,andthat ofits regionalpower-brokers,is to capturegrassroots initiativesfor its (andtheir)own politicalpurposes.Manylocallevel initiatives,however,evadecapture,aremisunderstood,or escapeattention altogether. At the sametime the presentandpotentialrole of NGOs in promotingthese initiativesis eithertreatedwith suspicion(by the government)or undervalued(in current proposalsto revive communitydevelopmentin Embuandneighbouringdistricts) The resultpromisesto be evenworsethanthe messwhich communitydevelopment- the 5l

orthodoxversion- is supposed to clearup. We can,however,takesomepleasurein the probabilitythat grassroots initiativesofone kind or anotherwill continueto flourish, whethersupportedor not.

Bibliography

Ardener, Shirley 1964 'The ComparativeStudyof RotatingCredit Associations', Joumal of the RoyalAnthropologicalInstirute, 94 (l l2), 210-229. Bahemuka,Judith Mbula and Mary Tiffen 1992 'Akamba Institutionsand Development,1930-90' in Mary Tiffen (ed) Instintional profile @nvironmmtal Changeand Drylnnd Managementin MachakosDistrict, KerryaI93GI9W, working paper). London: OverseasDevelopmentInstitute. Bardhan,P. 1993 'Analytics of the Institutionsof Informal Cooperationin Rural Development',WorldDevelopment, 2l (4), 633-ffi . Barkan, Joel D. 1982 Self-helpOrganizations,Stateand Society: Linkage and Developmcntin Rural Kenya, Occasionalpaper No.2l, Comparativet-egislative ResearchCenter, University of Iowa, Iowa City. Barkan, Joel D. and M. Chege 1989 'Decentralizingthe State: District Focusand the Politicsof Reallocationin Kenya',Journntof ModemAfican Sndies,27 (3),

43r-453.

Barkan,JoelD., FrankHolmquis, David GachukiandShemMigot-Adholla lgTg Is Small Beautiful? The Organiutiorml Conditionsfor Efieaive Small-scaleSelfhelp DevelopmentProjec.tsin Rural Kenya, Occasionalpaper No.15, Comparativel-egislativeResearchCenter, University of Iowa, Iowa City. Besley, Timotly, Stephencoate and Glenn L,oury 1992 TheEconomicsof Rotaing Saings and Credit Associations(revisedversion), Discussionpaper No. 157,

52

ResearchProgramin DevelopmentStudies,Woodrow Wilson Schoolof Public and InternationalAffairs, PrincetonUniversiW. Feldman,Rayah 1983 'Women'sGroupsandWomen'sSubordination:An Analysis of Policies TowardsRural Women in Kenya', Reiew of Afican Political Econow. 27128.67-85. Freire, Paulo

1972 Pedngogyof the Oppressed.Harmondsworth: Penguin.

Gachuki, David 1982 Harambeein Kenya: A Casefor its Regulntion,Occasional PaperNo. 19, Comparativel,egislative ResearchCenter, University of Iowa, Iowa City. Geertz, Clifford 1962 'The RotatingCredit Association: A "Middle Rung" in Development',EconomicDevelopment and Cultural Change,l0 (3),241-263. Hasseft,Donald 1994 Thc Community-based Approachto Natural Resource Management:A Review,working paperpreparedfor the Traditional Management of Artisanal FisheriesProject, CEMARE, University of Portsmouth. Hill, Martin J. D. l99l TheHarambeeMoyementin Kenya: Self-help, Developmentand Educuion amongthe Kanba of Kirui Distnct. I-ondon and Atlantic Highlands: The Athlone Press. Holmquist, Frank 1979 'Class Structure,PeasantParticipation,and Rural Selfhelp', in JoelBarkanandJohnOkumu(eds)Politicsand hthlic Policy in Kerrya and Tanzania. New York: Praeger. Holmquist, Frank 1982 Stue, Class,Peasantsand the Initiation of KenyanSelfhclp, OccasionalPaperNo.18, Comparativekgislative ResearchCenter, University of lowa, Iowa City. Holmquist, Frank 1984 'Self-help: The Stateand Peasantl:verage in Kenya', Africa, 54, 72-91. Kaluli, J. W. 1992 'NGOsand Technological Change'in Mary Tiffen (ed) Instintional Profile @nironmenml Changeand Dryland Matwgementin MachnkosDistrict , Kenya I 930-I 90 , working paper). London: Overseas DevelopmentInstitute.

53

Kane,Keviq MartinWalshandCandace Nelson 1991 'A Casefor Business Training with Women'sGroups',SmallEnterpriseDevelopment,2(l), 13-19. McCormack,Jeanne,Martin WalshandCandaceNelson 1986 Women'sGroup Enlerplises: A Studyof the Structureof Opportunityon theKenyd Coast. Boston: World EducationInc. Mayer,Philip 1951 'AgriculturalCo-operation by Neighbourhood Groupsamongthe Gusii in SouthNyanza',in P. Mayer, TwoStudiesin Applied Anthropologt in Kenya. London: Her Majesty'sStationeryOfficeMazingira Institute 1992 'The Women's Group Movementin Kenya: An overview' in Womenand Development:A KerryaGuide. Natrobi: Mazingira Institute- 6-19. Mbithi, P. M. and R. Rasmussen 1977 Self-reliancein Kenya: The Caseof Harambee. Uppsala: ScandinavianInstitute of African Studies. Mwagiru, Wanjiku 1985 Rural Womcn'sMutual Aid Estems in Kenya: A Case Sndy of Central Proince; Finnl Sndy Reparr, Departmentof Geography, University of Nairobi. Mwaniki, Nyaga 1986 'AgainstMany Odds: The Dilemmasof Women'sSelf-help Groupsin Mbeere,Kenya',Africa, 56 (2),210-227. Ng'ethe, Njuguna 1979 Harambeeand DevelopmentPanicipaion in Kenya, unpublishedPh.D dissertation,Carlton University. Njonjo, Apollo L., Anne Kairu, Wariara Mbugua, David Kimenya and Fred Chege 1985 Sndy on the Organizarionof Women'sGroupsin Kerrya: Main Repon (Vol.l). Nairobi: BusinessandEconomicResearch Co. Ltd. Nugent, J. B. 1993 'BetweenState,Marketsand Households: A Neoinstitutional Analysis of lncal Organizationsand Institutions', World Development,2l (4), 623-632. Odera, Millicent 1980 ltlyeri Women'sDevelopmentProject: BaselineSumey Repon. Nairobi: Family PlanningAssociationof Kenya. OverseasDevelopmentInstitute (ODD 1992 Embu, Meru, TharakaNithi District SupponPreparaion Mission: Firwl Repon @raft), repofi to the Overseas DevelopmentAdministration (ODA), l,ondon, July.

54

Pala,A. O., M. H. Wallis andJ. E. Reynolds 1978 'The Women'sGroup Programmein the SpecialRural DevelopmentProgramme(SRDP)', in A. O. Pala, T. Awori and A. Krystall (ds) ThePanicipuion of l4romenin Kerryan .tocrery. Nahobi: Kenya Literature Bureau. Republicof Kenya 1987 District Focusfor Rural Developmenr.Nairobi: Office of the PresidentRiley, BernardW. and David Brokensha 1988 TheMbeerein Kenya(Yol.l: ChangingRural Ecology). Lanham: University Pressof America. Shepherd,Andrew 1992 Putting Communitiesin Charge: TheRerurnrc CommunityDevelopmentin Africa: A CaseSndy and Discussion(Papersin the Administration of DevelopmentNo.zl6), DevelopmentAdministration Group, Schoolof Public Policy, University of Birmingham. Thomas,BarbaraP. 1981 Local Organization,Politics and Panicipation in Rural Develapment:A Studyof HarambeeSelf-helpin Kenya, unpublishedPh.D dissertation,BrandeisUniversity. Thomas,BarbaraP. 1985 Cash, Credit, Cooperationand Communiry: Responses of Womm in ktral Kenya, HIID DiscussionPaperNo.l8l, Harvard Institute for InternationalDevelopment,Cambridge,Mass. Uphoff, N. 1993 'GrassrootsOrganizationsand NGOs in Rural Development: Opportunitieswith Diminishing Statesand ExpandingMarkets', WorA Development,2l (4), ffil -622. Wallis, M. 1982 Bureaucra$, Politicians, and Rural Communitiesin Kenya, ManchesterPaperson DevelopmentNo.6, Departmentof Administrative Studies, University of Manchester,Manchester. Walsh, Martin 1992 'Community Participationin Isiolo District: PastInitiatives and Options for the Future', Annex 4 in TheIsioln District SupponProgrammc, a report subnitted by MASDAR Ltd. to the ODA, London. Wals[ Martin 1989 'Tototo Home Industries: AssistanceStrategiesfor the Future',in C.Mann,M.Grindle andP.Shipton(eds)SeekngSolutions: Fromeworkand Cases for SmallEnterpriseDevelopmentPrograms. West Flartford: KumarianPress.

t)

Welbourn,Alice 1990 lhe Governmentof Kenya/ODAEMI Programme: Consultancyin GenderResponsivePlamingfor Phase,I{ report to the EMI Programe,Embu. .Winans,Edgar V. and AngeliqueHdugerud 1977 'Rural Self-helpin Kenya: The HarambeeMovement', Hunan Organization,36 (4), 334-351.

56