DISCUSSION PAPER SERIES - discussion paper series paper no. 70 where does development success come from? explanations and practical implications anthony bebbington and ...

Download DISCUSSION PAPER SERIES -   discussion paper series paper no. 70 where does development success come from? explanations and practical implications anthony bebbington and ...

Post on 02-Mar-2018

213 views

Category:

Documents

1 download

TRANSCRIPT

IDPMDISCUSSIONPAPERSERIESPaperNo.70WHEREDOESDEVELOPMENTSUCCESSCOMEFROM?EXPLANATIONSANDPRACTICALIMPLICATIONSAnthonyBebbingtonandWillyMcCourtUniversityofManchesterOctober2006ISBN: 1904143849Furtherdetails:Publishedby:InstituteforDevelopmentPolicyandManagementUniversityofManchesterExternalAffairsOfficeHaroldHankinsBuilding,PrecinctCentre,OxfordRoad,ManchesterM139QH,UKTel:+44(0)1612752814Email:idpm@man.ac.ukWeb:http://idpm.man.ac.uk1WHEREDOESDEVELOPMENTSUCCESSCOMEFROM?EXPLANATIONSANDPRACTICALIMPLICATIONSAnthonyBebbingtonandWillyMcCourt(NOTE:Thispaper isaslightlyreworkedversionof the finalchapterofA.BebbingtonandW.McCourt(eds)Developmentsuccess:StatecraftintheSouth,tobepublishedbyPalgrave in early 2007. Chapter references are to the relevant chapters of thatcollection.)In this paper we draw lessons from the seven cases that constitute the core of ouredited collection, and whosemain elements are summarized in Table 1. We discusswhattheysuggestabout thenatureandexplanationofdevelopmentpolicysuccess intermsoftheframeworkwhichwedevelopedinChapter1andwithreferencetobroaderdebates in the literature. As noted in Chapter 1, our research strategy and choice ofcasesmean that our conclusionswill be tentative ones.We will openup avenues forfurtherresearchatthesametimethatweclosedownourpresentenquiry.2Table1: SummaryofcasefindingsCHAPTER COUNTRY/REGIONNATUREOFPOLICYDURATION(allpoliciesextantexceptwherestated)EVIDENCEOFSUCCESS SUCCESSFACTORS2:MeloBrazil CashtransfertopoorfamiliesFirstmunicipalprogrammestarted1994federalprogrammestarted1997andsurvivedchangeofgovernment95%ofmunicipalitiesparticipating34millionpeoplebenefitingfromannualtransferofUS$2.2billionFunctionalelectoralcompetitionDesignandpoliticalincentivesEarlysocialmobilizationandleadership3.Hofmanetal.Indonesia Macroeconomicpolicy196797,coterminouswithSoehartoregimeAverageGDPgrowthof7%p.a.196797povertydownfrom60%to11%ofpopulationPositiveinvestmentclimatecompetent,insulatedtechnocratspolicydesignpragmatismandflexibilitydonors4.GrindleLatinAmerica IndustryandeducationMid1980sonwards(industry)early1990sonwards(education)4.49%annualgrowth196580,illiteracy4215%195090successofcurrentpoliciesremainsunclearPolicydesignLeadershiptodislodgevestedinterests5.HulmeandMooreBangladesh Microfinance GovernmentordinanceforGrameenBankin1984survivedchangesofgovernment1200+MFIs,withgoodrepaymentrates13millionpoorhouseholdsbenefitinggenderorientationempoweringwomenInnovationPolicyspecificationImplementationfactorsLeadershipandsocialenergy6.MitlinChile,Philippines,SouthAfricaHousing Scaleoperationbeganinmid1980s(Chile)19828(Mexico)programmessurvivedregimechangeinChileandPhilippinesOperationatscaleupto1.4millionbeneficiaries(SouthAfrica)enhancedabilityofpoorpeopletocontestforpowerBuildingapolicyalliancewithoutelitecaptureEmpowermentofurbanpoorgroupsandallies7.Shankland&CornwallBrazil Health SUSenshrinedin1988constitutionsurvivedchangesofgovernmentUniversal,publiclyfundedhealthprovisiondramaticimprovementinbasichealthindicatorsPopularparticipationpreventselitecaptureofpolicyallianceProrightsbasedhealthsystempolicynetworks/epistemiccommunities8.JacksonMozambique ParticipatoryplanningPPFDstartedin1996survivedtwoelectionsandkeypersonnelchangesSomeeconomicimprovementincreasedstatelegitimacyPolicycoalition3THENATUREANDDURABILITYOFSUCCESSWhatkindofsuccess?In our introductory chapter we suggested that how we define development successdepends on how we define development itself. Our normative definition was theenhancementofhumancapabilities, inparticularfor thepeoplewhohavethegreatestcapability deficits. Such capability enhancement, we said, could occur through directinvestment in financial, physical, social or humancapital, or through improvements inthe environment in which these assets are developed and used improvements thatcouldoccurthroughinitiativesasdiverseaspeacebuilding,macroeconomicreformandgood governance programmes. We were interested in both economic and socialdevelopment, but with the proviso that the development should be (in the jargoncurrentatthetimeofwriting)propoor.Althoughwewentontorefertothetraditionaland contrasting public policyassumption that developmentmight also bewhatever alegitimatepublicactorsaiditwas,infactallourcontributorshavewritteninthespiritofournormativedefinition.That ismarkedly the casewith a chapter like Alex Shankland and Andrea Cornwalls,whichcontrastsBrazilshealthreformswiththeneoliberalpoliciesprevalentelsewherein Latin America, and implicitly also with Hulme and Moore, Mitlin and Jacksonschapters.ButitiseventrueofMerileeGrindleschapter,whichdiscussespreciselythoseneoliberal policies in the industry and education spheres. Her critique of the statecentredpolicieswhichtheychallengedisthatwhiletheyincreasedaccesstoeducationintheearlierperiod,theydidsomorefortherichthanforthepoor,andfortheruralpoor least of all. Likewise, if she ultimately reserves judgement on whether the newpolicieshavebeenasuccess,thatislargelybecausewelackevidencethattheyhaveledtobroadbasedgrowthorimprovedtheschoolingofpoorchildren.ThenormativeviewalsocharacterizestheanalysisofBertHofman,EllaR.GudwinandKianWieThee,whichagainatfirstglancemightappearindifferenttothedistributionaleffects of growth. Earlier accounts of goodeconomic performance in Eastand SouthEastAsiadepartedfromprevailingfreemarketorthodoxymainlyinstressingthepositiverolethatgovernmentandgovernment institutionshadplayedin regulatingthemarket4(Wade,1990WorldBank,1993a).Hofmanandhiscolleagues,however,saythatwhatgovernmentgot rightmore thananythingelse intheperiodbeforethe crashof1997wasthatitsecuredhighgrowththatwasalsohighlypropoor.Thusournormativeviewofdevelopmentsuccessseemstohavesurvivedtheshockofcontact with reality as represented by our seven cases. But of course even if ourcontributorsagree roughly onwhat constitutessuccesswhichmight reflectanearlytwentyfirst century consensus on the purpose of development that is no morepermanent than the late twentieth century Washingtonconsensus thatpreceded it the policies that they have chosen as their examples of success differ widely, as weintended thattheywould.Letusseewhethertheevidencethatourcontributorsoffergivesusanygroundsforcomparingexplanationsforthesuccessoftheseverydifferentpolicies.EvidenceofSuccessInordertogeneralizeaboutsuccess,wemustfirstshowthatsuccessdidindeedoccur.InChapter1wesaidthatwewerelookingforcaseswhichhadthefollowingfeatures: Theywouldtargettheenhancementofhumancapabilities, inparticular for thepeoplewhohavethegreatestcapabilitydeficits They should do so on a large scale: this might entail scaleup from an initialpolicyexperiment The policies would have been implemented over at least ten years, andpreferably across at least one change of government: policy duration wasimportant Theywouldpreferablyhavesucceededagainsttheoddsthatis,atthepointofinceptionareasonableobserverwouldhavepredictedthatsuccesswasunlikelyWe list our contributors evidence in the above order, starting with two forms ofevidence of enhancement of human capabilities (readers may want to refer again toTable1,wheremostofthisevidenceissummarized).51. Impact on income or other human development indicators, arguably themostimportantformofevidence.WhilewemustofcourseallowfortheusualproblemofdemonstratingacausalrelationshipbetweenapolicyandaparticularoutcomesomethingthatDavidHulmeandKarenMooredealwiththoroughly,andthatDavid Jackson also discusses our authors, experts in their respective policydomains, are confident that some such causal link exists. Some of themmayhave hoped for greater impacts (DianaMitlin'sdiscussioncomes tomind), butthe chapters do suggest that improvements in indicators of health, shelter,income, nutrition and other outcomes can be attributed to the policies theydiscuss.2. Socialandpolitical impact.AlthoughHulmeandMoorepresentevidenceof theeconomic impact of microfinance, they also say that it often seemsas if thisfundamentally economic approach has performed best in the social domain,particularly inwomens empowerment. Similarly, Jacksonargues that themostimportantsuccessoftheNampulaexperimenthasbeeninthepoliticalsphere,intheway inwhich theexperimenthasenhancedthe legitimacyofthestate inacountrywhichisstillrecoveringfromalongcivilwarandinaprovincewhichhadbeenbypassedbyboththe independencestruggleandthe initialpoliciesoftheFRELIMOgovernment.Theimplicitargumentsherearethatthereareimportantindividual social benefits such as increased confidence andparticipationwhichhumandevelopmentindicatorsfailtocapture,andthattherearealsocollectivepoliticalbenefitswhicharedifferentfromthesumofthebenefitstoindividuals.3. Scaleup:anindicationthatinitiativesthatbeganlifeinapolicytesttubewererolled out on a large scale. Cash transfers in Brazil existed only in SenatorEduardoSuplicysfertileimagination inthelate1980s,butby200295percentof municipalities had such a scheme (Chapter 2). The Grameen Bank inBangladesh began as a student action research project in 1976, but by 2002therewere1200microfinance institutions, reaching13millionpoorhouseholds(Chapter 5). Only Nampulas participatory planning experiment in Mozambiquehas still not been successfully scaled up to national level, despite donor6sponsored attempts, and despite its embedding in Nampula province itself(Chapter8).4. Policyduration.WestipulatedinChapter1thatasuccessfulpolicywouldbeonethatsurvivedforat leasttenyearsand,inacompetitivedemocracy,preferablyalso a change of government. This was the only element of our view ofdevelopmentsuccesstowhichallourcontributorsresponded.Alloftheirchosenpolicieshaddeliveredbenefitsforoveradecade.EvenGrindlescritiqueofLatinAmericasexhausted importsubstitutionpoliciesispremisedontherecognitionthat they generated sustained growth for at least the 19651980 period thepolicieswere good before theywent bad. In someof thecases policiesalsooutlasted the government that introduced them (income transfer programmesandhealthpolicyinBrazil,housingprogrammesinChileandthePhilippines).5. Successagainst theodds. (Wecouldnot findaneconomicalwayofpresentingthisevidenceinTable1.) Wedonothaveevidencehereinallourcases.Growthin Indonesia up to 1997 was not more impressive than in the other NewlyIndustrializing Countries of East Asia, and, from the vantage point of thepresent, less impressive than in South Korea or even neighbouring Malaysia.From the samevantagepoint, industrial development inArgentina, oneof theLatinAmericancountriesthatGrindlediscusses,hasproceededinfitsandstarts,withthecrisisthatreacheditsclimaxinDecember2001stillarecentandpainfulmemory.Butwedohavefairlyclearevidenceinthreeofourcases.WhenHenryKissingerdismissed Bangladesh as a basket case in December 1971, would anyoneoutside thecountryhaveexpectedthat thissupposedlysupine recipientof richworld handouts would be the source of that burst of social energy whichcreatedthemicrofinancemovementonlyfiveyearslater?Similarly,thoughlessdramatically, Brazils National Health System, the SUS, swam against theneoliberal current in Latin America which saw neighbouring Chile introduceprivate insurancebased health provision, and it came out of the neoliberal7shock of the Collor government stronger thanwhen it went in. On a smallerscale, participatory planning took root in a relatively remote province ofMozambiquewhichhadbeenneglectedbythecapitalbothbeforeandjustafterindependence.PolicyDurability:RentSeekingorSocialLearning?The way in which our case policies endured is as interesting as the sheer fact ofendurance.Our initialnave assumptionwas that itwouldnotbepossibletohave toomuchofagoodthingthataslongasagoodpolicylasted,itwouldcontinuetoproducereliablebenefits.Wedid indeedhavethreefairlyunambiguousexamplesofthat,inthecontinuingdevelopmentofmicrofinance inBangladesh,and themaintenanceofpublichealth provision inBrazil and of a participatory planning model in Nampula province,Mozambique(Chapters5,7and8).Eventhere,however,weunderestimatedtheeffortthatgoesintomaintainingagoodpolicy,andalsothepotentialforgoodpoliciestogobad.Microfinancewassustainedbyinnovationinmicrofinanceproducts,theresultofavirtuouscircleofsociallearninginwhichimplementationgeneratedfeedback,whichwasused inturntorefinethepolicy. InBrazilandMozambique,policiesweresustainedbypolicy coalitions. Even when the Brazilian coalition was placed on a firm institutionalfooting, itstillhadtobenurtured inhealthcouncilsandconferences. InMozambique,wherethecoalitionsinstitutionalfootingwasstillwobblyatthetimeofwriting,internalbondingactivitiesandcultivatinglocalandnationalrelationshipshavesteadiedit.Thus health policy in Brazil and participatory planning in Nampula are examples ofpoliciesthathavestayedoncourse.Butthatisnottrueofallourcases.Mitlin(Chapter6) shows that the nature of the coalitions that assembled around propoor housingpolicyinChile,thePhilippinesandSouthAfricacompromisedpolicyobjectivesfromtheoutset. The private construction companies which it was felt necessary to involveskewedhousingpolicytowardstheprovisionoffinishedhousingunits,andawayfromasites and services approach which would have served poor people better, becausesupportingresidentsowneffortsisfarmorelikelytomeetbroaderneeds,includingtheenhancementof residentscollectivecapacity, thanproviding finishedunitsonaplate,as it were. This is not only because of the inherently higher cost of finished units,8sometimes inflatedby the construction industry lobbying for rules requiringminimumstandards that increase the unit cost. Politicians, for their part, were prone to usinghousing for political advantage, at theexpenseof allocation decisionsbased onneedratherthanclientelism.Innovation and social learning through feedback, and the positive and negative rolesthat policy coalitionsmight play: these were all envisaged in our original framework,although the case chapters have fleshed out their ramifications in important ways.Jacksonspersonalaccount,forexample,makesitpainfullyclearwhatatortuousaffairassembling and maintaining a coalition can be. But we failed to anticipate Grindlesinsight,whichHofmanetal.sexpositionofIndonesiasmacroeconomicpolicyreinforces,thatpolicies thatstartedout goodmightendupgoing bad. Thishappenswhen thepolicydiverges from the policy need it wasmeant tomeet. It is asure sign that thepolicy has been captured, whether by the original beneficiaries who are no longerworthyones(suchasindustrialistswhoseinfantindustriesshouldbewellabletostandontheirowntwofeet),orwhatGrindlehighlightsbyinterestgroupsthatthepolicyitselfhascreated(teachers'unionswhichresistareorientationofeducationpolicyfromaccess to quality), or alternatively by frankly parasitical interests (Indonesias cronycapitalists).UptothispointGrindles insight isconsistent,asshe recognizes,withthe rentseekingliteraturespessimisticviewthatpolicycaptureisinevitable,asitiseconomicallyrationalforinterestedpartiestocaptureit.Therentseekingliteraturedealswiththeproblembycalling for a reduction in state activity in order to diminish the potential for rents(Krueger, 1974 Tullock, 1967 and 2005). However, Grindle reports that it has beenpossiblefordeterminedleaderslikeCarlosSalinasinMexicotodealwiththeproblemina different way, by introducing a new policy which dislodges a preceding policy andwhichmeets theneedsofsociety ratherthanofhostile interests, initiallyat least.TheLatinAmericanleaderscapitalizedonthewidespreadperceptionofeconomiccrisis,usedpatronagepowertoplacesupportersinkeyappointments,sidelinedhostilepartyleadersandtradeunions inshort,theypulledalltheleversthattheirformalpositionsgrantedtheminordertobreakstalematesandfacedownrentseekers.9GrindlesanalysisimpliesapolicyversionoftheSecondLawofThermodynamicswherepoliciesare inevitablysubjecttoentropy.Everypolicy,sheeffectivelysuggests,comeswithasellbydate(Hofmanetal.sexpirationdate).Wearemeanttoresignourselvesto the expectation that the same heroic forces that are currently defending cashtransfersandhealthpolicyinBrazilthepoliticalactorswhoestablishedthepolicy,andalsothenewinterestgroupsthatthepolicyitselfhascreated(familiesbenefitingfromthe Bolsa Escola programme,health service users) will degenerate into reactionaryobstaclestoanewandmoreappropriatepolicy.Further,wecandraw fromGrindlesanalysis the inference thatwhen thesellbydatearrives, the best way to bring about the termination of an exhausted policy is bysupersedingitwithanewonethatissupportedbynewideasandnewsupportersbytrumpingtheoldpolicyratherthanbyafrontalassaultonit,thelatteradrasticoptionthat few leaders will undertake (DeLeon, 1978 Frantz, 2002 Sato, 2002). Thusterminatingapolicybecomes,paradoxicallyenough,anissueforthebeginningandnottheendofthepolicycycle,sinceadvocatinganewpolicywillgeneratemorereformingenergythanopposinganoldone.However, the experience of microfinance suggests that there are ways of buckingGrindleslawofentropy,foratimeatleast.Thesameinterestgroupsdiggingtheirheelsintoresistanydiminutionoftheirprivilegesare,fromadifferentpointofview,sourcesof feedback that policymakers canuse tomaintain or rescue the congruenceof theirpolicies.Institutions,afterall,dosometimesrenewthemselves:thecounterreformationintheCatholicChurchisonecelebratedhistoricalexample.Andforpolicymakersjustasforprelates, the incentive is thatadaptationbasedon feedbackpreservescongruencewith the changingpolicy environment,withoutwhich the old policy risks being sweptawaybyanewandmorecongruentone.Where does this leave us? Durability turns out to be more important, and moreinteresting,thanwethoughtitwasinChapter1.Wesawthatapolicyneedstohavea(preferablybroad)coalitionbehindittosurviveforanylengthoftime.However,wealsosaw that durability turns out to be a twoedged sword: policies have the same10propensity to go bad as to stay good. To prevent policy decay, at the very outsetpolicymakers need to minimize the influence of potential coalition members whoseinterests are hostile to policy objectives. Following that, policymakers need to ensurethat policy coalitions become feedback mechanisms that facilitate adaptation, ratherthanossifyingintoentrenchedassociationsofrentseekers.Figure1presentsthepolicytrajectoriesthatwehavediscussedinthissection.11Figure1 Policyduration:AtypicalimplementationtrajectoryPolicycoalitionassemblesImplementationconsolidation adaptation ossification(Shankland&Cornwall) (Hulme&Moore) termination supersession stagnation(DeLeon,1978etc.) (Grindle)12DESIGNMATTERS:LESSONSONPOLICYCONTENTPolicyandthePoliticalEnvironmentOurcaseshave toldus relatively little aboutthenutsandboltsofpolicyat themicrolevel. The detailed discussion of microfinancemodels in Hulme andMoores chapter,where design has mattered critically to financial sustainability, viability and women'sempowerment, and likewise the brief but significant discussion of customs reform inHofman et al.s chapter, throw into relief the decision of our other contributors toconcentrateonthenonprofessional,elementsofpolicysuccess.Perhapstheirabsenceis inevitable in a collection that spans the gamut of development policy, militatingagainsttheprofessionalscrutinythatafocusonasinglepolicyareamighthaveallowed.However,wethinktherearestillsomeusefullessonsaboutthewaythatpolicydesignatthemacrolevelinteractswiththepoliticalenvironment,includingaboutthewaythatthepoliticaleconomyofpolicyisitselfamenabletodesign.WhereShouldPolicyAutonomybeEmbedded?The first lesson is a refinement of what Peter Evans (1995) has called embeddedautonomy.Policydesign in thetraditionthatstems fromWeber isassumedto requiresome insulation from politicaland social influences for the technocratswhodesign it.Hofman et al. argue that insulation along these lines contributed to Indonesiaseconomicsuccessup to1997.Intheespeciallycomplexareaofmacroeconomics,thatinsulation is symbolizedby the traditional purdah, or retreat from daytodaypolitics,whichtheUKFinanceministerandhisofficialsarepermittedintherunuptotheannualBudget. At theotherextremeare thosequickanddirty responses tosome immediatepoliticalneedthattendtobestigmatizedaspolicyonthehoof.However, in another instance of good policy going bad, the same insulation thatfacilitatedIndonesiassuccesslaiditopentofailure,Hofmanetal.tellus,asthecronycapitalists captured economic policy unconstrained by any civil society checks andbalances.Butthatdoesnotmean thatweshould feelnostalgic fortheearlyphaseofSoehartosauthoritarianregime.For,admittedlywiththewisdomofhindsight,Hofmanet al. castigate Soeharto for having bad policies in good times: for having failed toestablish sound institutions, including the institutions of democracy, at a time when13windfall oil revenues would have made this relatively easy to do. Moreover, inneighbouring, moreorless democratic Malaysia1 and under the roughlycontemporaneousMahathir, the technocratsenjoyed thesame insulation frompolitics.That alsogoes for their counterparts elsewhere in the region, asHofmanet al. note.Thosecountriesareindeedexamplesofembeddedautonomy,wheretechnocratshaverealautonomythatisqualifiedbybeingembeddedinadensenetworkofsocialtiesandwhichis,amongotherthings,asourceofintelligenceorofwhatwehavecalledsociallearning.Arguably therefore,ourcasesshowthatnot justanyold formof embeddedautonomywilldo.Socialties,yesbutwhichones? Afterall,macroeconomicpolicyinIndonesia did become embedded after a fashion, but in a small sector of cronycapitalists, while its insulation from other groups and political processes fatallyweakenedpolicyadaptability.Thus the what of policy cannot be separated from the who whom? questions ofpoliticaleconomy.Wehavetodistinguishbetweengroupswhichsupportpropoorpolicyandgroupsthatopposeit, inverymuchthewaythatMitlindistinguishes inChapter6between the private companies and politicians that distorted propoor housing policyand the residents groups that tried to keep it on track. In a similar way, becauseindustrial policy in Latin America was not designed to enforce slow phasing out ofindustrialprotection,andtherewerenocountervailinggroupstopointtothatomission,itwaspossibleforindustrialintereststolobbyforprotectiontoincreasebeyondreason.That in turn contributed to the replacement of the import substitutionpolicy by neoliberalpolicy,asGrindleoutlinesinherchapter.DesigningParticipationGrasping the importance of distinguishing between groups that support and opposepolicyobjectivespreparesustolearnfromthehistoryofBrazilsNationalHealthSystem,perhaps the most significant of all the experiences which our cases record. Theinstitutionalization of support has contributed to its consolidation, as Shankland andCornwall explain. Enshrined in the Constitution and the Basic Health Law, the HealthCouncils with their entrenched representation from organized civil society and theregular cycle of Health Conferences at municipal, state and federal levels have been14crucial.Itmaybethatpolicymakersunderestimatetheextenttowhichthewhowhomissuesofthedistributionofpowercanbedesignedinthesamewayasthewhatissuesof policy content. We will expand on this point when we discuss the importance ofinstitutionslateron.PolicyDesignandPolicyLearningDesignissuesarealsoimportant inensuringadaptivecapacityinpolicy.Whilemostofour contributors do not use the language of policy learning in their chapters, thesubstanceoflearningispresentinmostofthem.Thechannelsforlearningaremultiple,andmanyarenotdirectlyrelated topolicydesign:thefreedomofacriticalpress, thescopeforpoliticaldebate,theexistenceof independent researchbodiesandahostofotherattributesofahealthypublicsphereareallcriticalforpolicylearning(seebelow).But learning can also be designed for. Brazils health councils and conferences, andNampula's district planning processes (Jackson), are as much sources of learning aswaysofembeddingpolicyandmakingitlegitimate.IdeasandCoalitionsWhetheranascentpolicyappealstoanaudiencethat isbroadenoughto lift itoff thegroundisamatterof ideasaswellas interests(John,1998).ThiswasclearinMarcusMelo'sdiscussionofSenatorEduardoSuplicy'seffortstogainpoliticalsupportforabasicincomegrantpolicyinBrazil.Wehaveseenalreadythatinmostofourcasestherewasabroadandloosecoalitionofpolicychampionswhichshepherded thepoliciesthroughthenecessaryparliamentaryandotherprocesses.InBrazilitwasthefertilesuggestionofaneconomistcalledJosMrcioCamargothatincomegrantstopoorfamiliesshouldbe conditional on children going to school which attracted the broad political supportthatgavethepolicythenecessarytraction.AfurtherelementofdesignthatreceivessomewhatlessattentionisclearinHofmanetal.'s chapter. This relates to that aspect of design that enhances synergies amongdifferentpolicyinstruments(forinstance,betweensectoralpoliciesandmacroeconomicpolicies), thus increasing the overall human development effects not just of eachinstrumentbutofthepolicypackageasawhole.Thisthemeappearsascentraltothe15storyofIndonesianeconomicsuccess,andwhileHofmanetal.donotnecessarilysaysoexplicitly,thesuggestionisthattheabilitytodesignwithsuchsynergyinmindwasitselfhelped by the existence of a strong, protected and professionally competent team oftechnocrats.ThePoliticalEconomyMustSupportPolicyDesignIn this sectionwehope to haveshownhowpolicydesignhas interactedwithpoliticaleconomyinourcases.Certainlyweareobligedtorepeatthatouremphasisonpoliticaleconomicand institutionaldimensionsofsuccessoughtnot toobscurethebrass tacksof planning and design which matter greatly to development success theWeberianinsistenceonpolitical insulationforthe technocraticplannersiscorrect.However,eventhe best design in the world will only succeed if the surrounding political economyprovides,orisdesignedtoprovide,acoalitionthatwillsupportthepolicyandthatwillgivefeedbackthatwillallowthepolicytorespondtochangingcircumstances,andhencetoendure.POLITICALECONOMIESOFSUCCESSThatpoliticaleconomywhichwehavejusthighlightedinrelationtopolicydesigncanbeanallencompassingconcept.Indeedinchapter1weuseittocovereverythingfromleadership to statesociety and power relationships. Underlying this broad use of theterm, though,are foursimple ideas.First,policiescannotbeanalyzedseparately fromtheirsocial,politicalandeconomiccontexts.Second,heroic leaders,entrepreneursandbrokers, so often the darlings of policy analysis,must be understood in terms of thenetworks of which they are a part, and in turn the institutions and social structureswithinwhich thosenetworksareembedded.Third,theforms,functionsandeffectsofpoliciesareasmuchdeterminedbywidersocioeconomic relationshipsas they arebythe internal dynamics of the policy design process. And finally, for our particularpurposes in this collection, success isaquestionofpolitical economy:not only in itsdetermination,butalsointhepowerrelationshipsthatultimatelydecidewhatcountsassuccessinasociety.16StateSocietyRelationsandPolicyCyclesThe cases have reinforced our initial editorial conviction that policy was far fromsynonymouswithstatepolicy,andthatsuccessfulpublicpolicydependsonmuchmorethan state action. They depict interaction between the state and a wide range oforganized nonstate actors which includes social movements, unions, NGOs andbusinesslobbies,andwhichtakesdifferentformsatdifferentstagesofthepolicycycle.Herewefocusontwowaysinwhichstatesocietyinteractionscontributetosuccess.First, they have stimulated a vigorous public debate that has made certain policiesimaginableandpoliticallyviable.Socialmovementshavebeenkey.TheBraziliancasesareclearon this.TheprogressivedemocratizationofBraziliansociety itself largelyaproductofsocialmovementpressure (Dagninoetal.,2006)alsomadepossiblenewandincreasingly inventiveformsofmobilization.Oneof thesewas thecivicmovementto impeach President Collor which gave rise to a countrywide mobilization (Melosphrase)todistributefoodtopoorpeopleinwhichNGOsplayedanimportantrole.TheAction in Defense of Citizenship for the Fight against Hunger spawned thousands oflocalcommitteesacrossthecountry.Publicawarenessofhungerintensified,influencingdiscussions in Parliament. Melo is convinced that these antecedents were critical increating an environment in which a basic income transfer programme could becomepolitically attractive. Shanklandand Cornwall suggest something similar about Brazilsmovimento sanitarista (the movement for public health) which brought activists andprofessionals together and paved the way for a publicly funded, rightsbased healthsystem.This process ofpolicydevelopment is very different from the basically topdownonethatHofmanetal.andGrindleoutline,butnotwhollydifferentfromtheoneoutlinedbyHulme and Moore. The similarity between Bangladesh and Brazil is that incometransfers,microfinanceandpublichealthprovisionallalikearosefromaninitialburstofwhat Hulme andMoore, quotingUphoff who in turn got the idea from Hirschman,(1984) call social energy: the creative popular ferment which accompanieddemocratization inBraziland independence inBangladesh.Thedifference,however, isthat Brazils activists channeled their energy into the state, whereas Bangladeshs17channeled it into the voluntary sector: Brazils activists were as optimistic aboutgovernmentasBangladeshswerepessimistic.Suchsocialmovementprocessesnotonlygivemorevisibilitytoissuesthatlaterbecomepolicies.Theyalsocreateexpandingnetworksofrelationshipsandsharedcommitmentsthat help sustain the policies that are grounded in those commitments. Moreover,through their subsequent involvementwith thepolicies, the networks develop furtherlinks with government reformists (cf. Fox, 1996), and with other activists andprofessionalsoutside thecorebureaucracy.ShanklandandCornwall argue that this ispreciselywhatoccurred inBrazilshealthcouncils,andanalyze theprocess in termsofPeter Haas's concept of epistemiccommunities in their case communities that runfromcouncilmembersthroughtoprofessionalsandadministrators.Theseepistemic communities playvarious roles.First, theyprovideavisionofwhatafully fledged policy would look like. The case of microfinance (Hulme and Moore) isobvioushere,but thisalsoapplies to participatory planning inMozambique (Jackson).Followingthat,theyhelptomaintainthepolicy,theyprovidefeedbackthatcanbeusedforpolicyadaptations,andtheyguardagainstthepossibilitythatpolicywillbecapturedor watered down by others. Mitlin's arguments about the susceptibility of housingprogrammes tocaptureby business or political interests reflectherconcern thatsuchnetworkbased communities are weak in the field of housing policy, leading to herclosingcallforpoliciesthatfostermoreempowerment.NetworksandtheLegitimacyofNonStateActorsMitlin'sdiscussionisalsohelpfulbecauseitmakesclearthatthereisalwaysarangeofnetworkbasedcommunities that surroundpolicy. Their relative power determines thedirectionofpolicy.Hofmanetal.saythesamethinginsomanywordswhentheypointto the multiple ways in which nonstate (or better, nonSoeharto) institutions wereweakened, preventing other networks fromdeveloping and steering policy away fromtheincreasinglydominant(andclosed)networksofthecronycapitalists.Assoofteninthiscollection,thequestionsuccessfulforwhom? iscrucial. Wesuggestthatitisthe18groupswhichareabletomobilizenetworksandinfluencepolicydevelopmentwhichwillbenefitfromthatpolicy.ItwouldseemtofollowthattheBrazilmodelofsocialengagement(developmentbythepeople)offersmorethantheBangladeshmodelofsocialaltruism(developmentforthepeople, ofnonpoor,forpoor,asJamesCopestake isquotedassayingbyHulmeandMoore).Thenatureoftheevidenceinthiscollectionmeansthatwecanonlyflagthisupasaninterestingpossibilityonwhichfurtherresearchwouldbeworthwhile.However,ifwecandevelopthatsuggestion,thendevelopmentsuccessbecomesawarofattritionwhich requires thesustained empowermentof propooractors. Thatempowerment inturnrequiresallianceswhich,howeverrivenbylatenttensions(Mitlin),arestrongtotheextent that that they aregrounded inprior networks of solidarity and forms of socialenergy thatbind likemindedactors together (Fox,1996Hirschman, 1984).Critically,and again asMitlin suggests, for such alliances to win, they must involve both civilsocietyandstate.Civil societyactorsneed toenroll state reformists in theirefforts toelicit and sustain particular policies, and reformists need to reach out to civil societyactors.Thus the state isnever thesoleactor in anyof ourcases. That isnota new insight:scholars have recognized it at least since Heclo and Wildavsky (1974) and, morerecently, Rhodes influential writing on networks proceeds from it (1997). But in ourcasesthestateisnotalwayseventheleadingactor.This ishumblingforbothelectedandappointedstateofficialsinoneway.Butinanotherway,itcreatesthepossibilityofharnessing the social energywhich socialmovements seem to find easier to tap intothanpublicbodies:thecivicmobilizationwhichleduptothesubstantialcommitmentstoAfrican aid at the G8 Summit of 2005 at Gleneagles in Scotland is a dramatic recentexample.Dethroningthestatewouldappeartoraisethecriticismof legitimacysooften leveledat socialmovements: bywhat right do youpresume to place yourselves on an equalfooting with a legitimate government, especially a democratically elected one? Wecannot hope to answer that criticism fully in this chapter. However, we are perhaps19entitled topoint toShanklandandCornwallsaccountoftheparticipatorystructuresofBrazils National Health Systemas adding to rather than supplanting the conventionallegitimacy of the state. Certainly those structures may over time degenerate into aglorifiedrentseekersclubwhereanobsessionwithuniversalaccesstohealthservicesprevents the service improvements that a marketbased system may be better atproviding:thatwouldbeGrindlesfear.Butforthemomenttheyappearonbalancetobeenhancingthehealthcareavailabletocitizensratherthandamagingit.VarietiesofPolicyandtheQuestionofDistributionIntheintroductorychapter,wesuggestedthatpoliciesthataimtoredistributeresourcesarelesslikelytobesuccessful, largelybecausetheyencountermoreresistance.Atonelevel,thismayseemanunhelpfulobservation,asalmostallpolicieshaveredistributiveeffectsovertime.However,ourcasesappeartoendorsetheargumentthatpoliciesthatareexplicitlyredistributivebethisbecausetheytransfertheresourcesofonegrouptoanother,becausetheytargetsignificantpublicspendingtoparticulargroupsorbecausetheyappeartotakeresourcesawayfromaparticularinterestgroupareverypronetoproblems.This is partly due to the institutionalcomplexities that derive fromproblems of scale.Any policy that attempts redistribution on a significant scale needs to involve largeinstitutions.Mitlinargues,forinstance,thatmasshousingpolicyneedstheinvolvementofthestateforbothresourceandlegalreasons.Inthecaseofmicrofinance,anyefforttoscaleup(to stepup fromaproject toa policy, so to speak)has likewise requiredNGOs to seek funds from new sources. In Indonesia, many of the sectoral andmacroeconomicpolicieshadtoberesourcedbyexternalloans.Inallthreecases,goingto scale drew new players into the institutional network through which policy wasdelivered,andthoseplayersbroughtwiththemanadditionalsetofinterests,potentialconflictsandconditionswhichcomplicatedpolicydynamics.MerileeGrindle'schapteraddressesanotherrelatedsenseinwhichpolicytypeinfluencestheprobabilityofsuccess.Arguingparticularlyfromtheexperienceofeducationreform,she suggests that policies of expansion are more likely to succeed than are policies20aimedatchangingthewayaserviceisprovided.This isbecauseexpansionarypolicies(e.g. thoseaimedat increasingeducationprovision),whilebenefitingsomemorethanothers, do at least bring somebenefit tomany different groups. Policies that aim tochange themechanisms throughwhichservicesare delivered andcontrolled are quitedistinct. They offer far less by way of immediate tangible benefits, and generallyconcentrate heavy costs in a few particular groups (who, furthermore, are often thesamegroupswhobenefitedfromearlierpoliciesofexpansion).Theygeneratefarmoreresistance.LeadersandLeadershipGrindle shows that the possibility of overcoming resistance depends greatly onleadershipskills,boththeabilitytonegotiateandtheabilitytoforcethroughchangeinthe faceof resistance.Likewise, itmaybe that inconditions of increasing institutionalcomplexity such as thosewe outlined above, forceful leadership becomesevenmoreimportantasawayofcuttingtheGordianknotof institutions.Certainlyleadersappearas important actors in the policy narrative in all our cases, and also as part of theexplanationofsuccessinGrindleandHulmeandMooreschapters.Giventhatpolicyisapublicmatter,leadersarethepublicfaceswhichthosecoalitionsmusthaveiftheirideasaretobeimplementedaspolicies.We havealready listedways inwhichGrindles leaders used the levers of power.Weneedonlylist,additionally,acoupleofthewaysinwhichMuhammadYunusshapedthedevelopment of microfinance in Bangladesh: through inventing it, primarily, but alsolateron throughusinghiseliteposition tonegotiate itsuniquestatutoryposition,andlaterstillthroughreinventingitintheformofGrameenII.Leadershipthusemergesasintegraltopolicysuccess,consistentwithotheraccountssuchasTendlers(1997).Yetitremainshardtotheorizeandworkwith,asMelopointsoutinChapter2.Perhapsitishelpfulifwebreaktheidentificationofleadershipwithasingleleaderandthink of it as a task, not a person. Individual power is not the exclusive property ofleaders who are formally designated as such. Leadershipmay be distributed (Barry,1991BrownandHosking,1986),sharedrelativelyequallyamongseveralmembersofa21coalition. That seems to have been the case with the coalition, as Jackson calls it,which drove participatory planning in Nampula (Chapter 8). Alternatively, leadershipmay be sequential, with different individuals having more or less power at differentstagesofpolicydevelopment,ratherinthewaythattheconchshellinWilliamGoldingsLord of the Flies passes from hand to hand to indicate which boy can speak in ameeting.MelosfinegrainedaccountofBrazilsincometransferpolicyillustratesthis.2 ItwasSenatorEduardoSuplicywhoputtheoriginalideaonthepolicytable,(hemodestlyattributes the idea itself to a colleague: see Chapter 2, Note 12), whereupon theeconomistCamargorefineditinawaythatmadeitattractivetothemunicipalities(theMayorofCampinas,MagalhesTeixeira,ranwiththeball forawhile),andthentothefederalgovernment,initiallyunderCardosoandlatterlyunderLula.Therearealsodifferentcategoriesofleader.Oneset,exemplifiedbySuplicyandYunus,includes those leaders who both help elaborate the initial policy idea and thenassiduouslypromote it. InBurns (1978)wellknown terms, theyare transformationalleaders, exercising power that derives from their determination and intrinsic skills aspolicy innovators (see also Rogaly, 1996). They are not necessarily elected or deeplyembeddedinbroadersociopoliticalstructuresthatgivethempowerandlegitimacy.Asecondcategorydoesderive itspower fromthepoliticalbase.Theseare theclassicpolitical leaders who populate Grindles chapter. They resolve conflicts, build policyconsensus, and force through policy adaptations which break down the politicalequilibria thatblockchange.However, it isnota foregoneconclusionthattheywilldoanyofthosethings.Melosuggeststhattheirwillingnessorotherwisedependsontheircalculationofthepoliticalbenefits.InBurnsterms,thesearethetransactionalleaders.(Ofcoursesomeverygifted individualscanbeboth transactionaland transformationalleaders:MuhammadYunusinBangladeshisanexample.)A third category of leaders are the technocrats. They appear as important players inmacroeconomicpolicyinIndonesia,andinindustrialandevenmoreeducationpolicyinLatinAmerica.Theirs iswhatFrenchandRaven(1959)havecalled expertpower.3They may also have their own networks, particularly professional ones epistemic22communities insideandoutside their countrieswhich are also a power base of sorts.Technocratic leaders operate behind the scenes. While this gives them room formanoeuvre,itmeansthattheycannotplayarole inresolvingpoliticalconflictsaroundpolicychange,andsotheyneedtheclassicpoliticalleaderstodothatjobforthem,andtoinsulatethemfrompoliticalpressures.Ourcasessuggestthatleadershipisparticularlyimportantatthreepolicystages: the inceptionstage,whereanindividualor individualsget(s)the ideaontothepolicyagenda(forinstance,SuplicywithincometransfersinBrazil) the implementation stage where an individual or individuals overcome(s) theopposition of supporters of the old dispensation (Grindles Latin Americanleaders) an individual or individuals ensure(s) that the policy adapts to changingcircumstances,usingthecoalitionasafeedbackmechanism(MuhammadYunuswithmicrofinanceinBangladesh)LeadersandFollowersAllthisissomewayfromtherailwaybookstallpictureoftheleaderastheautonomous,lonely individual taking toughdecisionsanddragginghis followers inhisorherwake.Leaders of the kinds we depict may articulate their followers aspirations better thantheycan,andtheymayevenpersuadetheirfollowersthattheyseetheirinterestsmoreclearly than theydo themselves. But theirpolitical identity is inseparable from thatofthe coalitionswhose figureheads they are. Indeed their power evaporates when theybecome detached from their power bases and identify themselves with other elitefiguresoutsidetheircoalitions.ThatwasthefateofBishopAbelMuzorewaintherunupto Zimbabwean independence, of Mikhail Gorbachev in the dying days of the SovietUnion and of Margaret Thatcher after eleven years as the United Kingdoms primeminister.Wecancontrast that fatewith theway thatNelsonMandela in SouthAfricaresisted the apartheid governments attempt to drive a wedge between him and hisparty,theAfricanNationalCongress,justbeforehisreleasefromprisonin1990.23This is close to the view of leadership which Melo derives from Fiorina and Shepsle(1989) inChapter2. It isclearthatGrindles leadersarenotofthis type,but it isnotclearthatthepoliciestheyhaveforcedthroughhavebeensuccessful.Inthiscollectionatleastandonceagainwestressthelimitationsofoursevencasesitistheleadersofthekindwehavedepictedwhosecoalitionspolicieshavesucceeded.Perhapstheneedforleaderstostayclosetofollowersmitigatesthestrikingfactthatalltheleadersthatwehavenamedinthissectionarestillwithinthecharmedcircleofthepolicy elite: they are of nonpoor, for poor, in Copestakes words which we quotedearlier.Thereisnothinginourcasestosuggestthatindividualpoorpeoplehavedirectlyinfluenced propoorpolicy (HulmeandMoorehaveaninterestingdiscussionofmensleadershipand[poor]womensagencywhichbearsonthis).LackingallofFrenchandRavensbases of social power, their route to influence is through collectiveaction,asMitlinemphasizesinherchapterconclusion,andasShanklandandCornwallillustrateintheirdiscussionofBrazilsinstitutionalizedhealthcouncilsandconferences.IMPLEMENTATIONGiven that our definition of success hasbeen a timebased one wewanted to seepolicies that have endured and, preferably, survived a change of government theimplementationphase ispartandparcelofdevelopmentsuccess.Thechapterssuggestseverallessonsaboutit.ComplexPoliciesandEmpoweringthePowerlessThe first lesson concerns policy complexity. Generally, the simpler a policy'simplementationprocess, thegreater thechanceof success.Macroeconomic policiesofthe kind that Hofmanet al. address are easier to implement and sustain than socialdevelopment policies because the path from design to implementation is short andinvolvesfeweractors.4 Ashortpathmeanslessresistance,slippageorreworkingofthepolicy.This infactwasoneoftheargumentsusedtojustifysettingupLatinAmericassocialfunds.Theyweredesignedtominimizethestepsfromcentralgovernmenttofinalrecipient in order to speed up implementation and limit the diversion of funds (vanDomelen,2006).24However,formostsocialsectorpolicies,implementationisacomplexprocessinvolvingmultiple actors, intermediaries and interests. Our cases suggest a useful lesson here,one thatrelates toempowerment.DianaMitlinarguesforcefully inherchapterthatarangeofactorstrytotwistpolicyimplementationtotheirownadvantage.Sincesomeofthem naturally have more power than others, the extent to which a policy actuallyaddresses human development needs comes to depend on the extent to which theimplementationprocessempowersthosewholackpower,sincetheyaretheveryonesatwhom propoor policy is directed (inMitlins case, thosewho lack secure shelter).Shankland and Cornwalls chapter pulls in the same direction. Brazils health councilsenhancedhealthsystemusersability to insistthataccesstoservicesshouldbebasedoncitizenship,notthemarket.These observations, which echo those made by Guggenheim (2006) among others,highlight the fact that policy success means keeping implementation processes ontrack,orientedtowardshumandevelopmentgoalsratherthanclientelistones.Forthistooccur,theprogressiveempowermentofthegroupswhicharecommittedtoensuringthesehumandevelopmentgoals isofsignal importance to forestallpolicycaptureanddistortion.PolicySuccessandtheInstitutionsofDemocracyWe included institutions under the heading of implementation in our Chapter 1framework,soitisconvenienttoreviewherewhatourcaseshavetoldusaboutthem.At this stage we do not need to labour their basic importance, which most of ourcontributorsrecognize,notablyHofmanetal.andGrindle.However,ourcontributorsaredividedaboutthespecial caseofthe institutionsofdemocracysowewill spendsometime discussing them in this section. Itmaybeanadvantageofacrossnationalcasecollectionlikeoursthatthereisanopportunitytoreflectondivisionsthatmaybetakenforgrantedassumptionsatnationallevel.Our first group of contributors associates democracy with success. The association isstrongestinthetwoBrazilchapters,whicharguethattheelectoralpopularityofincometransfersandtheinstitutionalizationofcivicengagementinthehealthsectortherehave25doneagreatdealtokeeptheincometransferpolicyandtheNationalHealthSystemontrack.ButitisalsopresentintheIndonesiachapter.EarlieronwenotedHofmanetal.sreading that although prosperity was possible in Indonesia up to 1997 with weakinstitutions which included Soehartos guided democracy, the Crisis of 1997was thepointatwhichtheinstitutionalanddemocraticchickenscamehometoroost.However,othercontributorsbarelyraiseevenE.M.Forsterscelebratedtwocheersfordemocracy. All that Hulmeand Moores social entrepreneurs seem to need from theBangladeshgovernmentis for ittokeepoutofmicrofinancesway.TheynotethatthegrowthofNGOsthere,includingmicrofinanceNGOs,wasareactiontotheearlyfailureof government to respond to the problems that accompanied Bangladeshsindependencein1971.SonegativeistheviewofgovernmentthattheGrameenBank,they report, refuses toemployanyonewhohasbeen taintedbyworking in thepublicsector. Likewise, Mitlins heart is with community democracy rather than the partypolitical variety, which she sees as indelibly clientelistic while Grindle, for her part,makesnodistinctionbetweendemocraticandauthoritarianleaders.5Outsidethesometimeswishfulthinkingofinternationaldevelopmentagencies, itistheskeptical viewofdemocracyof the lattercontributors that isprobably themainstreamacademicone,asMelohighlightsatthestartofhischapter.MushtaqKhan(2005),forexample, has used the incontestable fact that rich countries are generally moredemocratic than poor ones to argue that the electoral competition that stems fromdemocratization indevelopingcountries isvery likely to increase thepower ofpatronclient factions,sincetheyareusuallybetterorganized toasserttheirpreferencesthantheatomizedmajorityof poor voters.Heexempts fromhis rule only thehighmiddleincomecountrieswhereheseesdemocracyapproximating to the versionprevailing inrichcountries.YetKhanremainsexplicitlyademocrat,asisGrindleamongourcontributorsdespiteheradmirationforsomeauthoritarianleaderspoliticalskills,andevenMitlinandHulmeandMoorescriticisms of democratic politics arecontingent rather than fundamental ones.Therefore the challenge is to make democracy work for development, and to see26whether the socialentrepreneurialism ofChapters 2,5 and7can be replicated in,ortransferredto,othercountries(fromBrazil),andequallytothepublicsector6 (fromtheNGOsector,wheremicrofinancehasbeenmostvibrant).ConstraintsonReplicatingBrazilsStateLedSuccessIn canvassing that possibility wemust be aware of four specific constraints that ourcasessuggest, inadditiontothegeneralonethatcontextmatterswithwhichmostofusarenowfamiliar.1. Both our positive examples of civic engagement come from a single country,Brazil. Relevant to Khans critique, one of Brazils characteristics is that it is alowermiddleincomeeconomy,whereasBangladesh,despiteencouragingrecentgrowth, remains low income. (However, so too does Mozambique, whosesuccess with embedding democratic participation in a disadvantaged regionJacksonhasrecorded.)2. Even if thedecisionofsocialactivists togodown theNGOpath inBangladeshandthegovernmentpathinBrazilwasfortuitous,ratherthanbasedonacleareyedcalculationabout thepotential for astateledapproach tosucceed in thetwo countries (and we have already noted the view that the growth ofBangladeshi NGOs reflected government failure following independence), theremay be some path dependence by now (Mahoney, 2003), restrictingBangladeshsfreedomtoswitchtothegovernmentpath.3. A replicationofthesocialenergythataccompanieddemocratization inBrazil inthelate1980swillprobablybenecessary,asthatwascrucialattheinceptionofthe income transfer policy and the National Health System. We need hardlystresswhat a tall order that is, althoughweshouldnote that thereare thosewhobelievethatsomeinducementorcatalyizationispossible:seeforexampleUphoff(2000).274. Greatcarewillbeneededtodesignaninstitutionalstructureofparticipationthatwill empower the intendedbeneficiariesof thepolicy, andprevent thepoliticalpatronsandrentseekersfromcapturingit.Thisisalsoastringentconstraint,butatleastourBraziliancasesofferencouragingprecedents.It is beyond the scope of our enquiry to say if those formidable constraints can beovercome in any particular country. However, it seems reasonable to state that if inBangladesh (or elsewhere) a new generation of social entrepreneurs were to arise,shareKhanandGrindlesdemocratic instinctsanddecide toapply thesameenergytothe public sector that a previous generation applied to the voluntary sector inBangladesh, then the success of stateled social development in Brazil may offer amodelthatitcanadapt.POLICYTRANSFER:ENTREPRENEURSANDINCENTIVESVoluntaryandCoerciveTransfer:DonorsasBitPlayersIndiscussingpolicytransferinChapter1,wesuggestedthatitwasunlikelythatpoliciesimposedthroughcoercivetransfer(MarshandDolowitz,2000)wouldsucceed.Itmaybesignificant,therefore,thatnoneofthesuccessfulpoliciesinthiscollectionseemstohave been foisted on the policymakers. Indonesias macroeconomic policy underSoeharto and economic liberalization in Latin America were broadly orthodox, butaccording toHofmanetal. andGrindle theywere stilldeveloped locally. In Indonesiathe technocrats were more influenced, it seems, by their university teachers in theUnitedStatesthanbytheIMFortheWorldBank,whoseviewsalsocountedforlessinLatinAmericathanpoliticalleadersowndeepconvictionthatchangewasnecessary,asGrindlegoesoutofherwaytoemphasize.Where policies are transferred voluntarily rather than imposed, then politicalsalesmanshipwillbeatapremium:thesayingthatagoodproductsellsitselfdoesnotapply topolicies.Hereagain there isa role for thedomestic transformational leader.Eduardo Suplicy, Raul Prebisch and Muhammad Yunus in Chapters 2, 3 and 5respectivelyareexamples. Theypromotedaswell asdesigned thepolicieswithwhichtheyareassociated.28ItisstrikinghowmodestistherolethattheIMF,WorldBankandotherdonoragenciesplayinourcases,eitherforgoodorforill,consideringtheirubiquityinthedevelopmentliterature.7 Thismaybebecausewhentheseagenciespromoteapolicy,theirfinancialpowermeans that they inevitably dosoas coercers rather thansalesmen.But itmayalsobeandthisisahypothesisthatlargebureaucraciesarelessgoodatinnovatingor persuading than think tanks, research institutes, NGOs and public intellectuals. Ifthere is something to thehypothesis, then apart from the coldwater it pours on thelate90s ambition of the World Bank to turn itself into a knowledge bank, it hasimportant implications forknowledgegeneration tosupport policydevelopment.Multinodalsystemsofpolicyinnovationwhichbridgeresearchandpolicymaybemorefertilethanbureaucraticmonoliths(Stone,20002002).VarietiesofTransferEven in our cases of policies that havebeenadopted voluntarily rather than throughcoercion, there is still considerable variation in the nature of the transfer. Grindlesaccountof the transfer ofpolicy ideas in industrialdevelopment andeducationacrossLatinAmericangovernments isclosesttothetransferprocessthatMarshandDolowitzenvisage.Herethetransferisfromonenationalpolityandtechnocracytoanother,withotheractors,inparticularmultilateralones,oilingthewheels.Butwealsohave two formsof transfer that areoutsideMarshandDolowitzsmodel.Thefirstiswhatisprobablytheparadigmaticcaseofmoderninternationaldevelopmentpolicy transfer, that of microfinance. Its interesting feature is that it has largelybypassedgovernments,theconventionalagentsofpublicpolicy.TheprocessoftransferhasbeenfromitsinceptionandexpansionintheGrameenBanktoitsadoptionbyotherBangladeshiNGOs,andthensubsequentlybyinternationalpolicycircuits,includingtheWorldBank, at whichpoint it entered the pantheon of development orthodoxy. This,however,isalogicalconsequenceoftheviewofdevelopmentpolicywhichweespousedinChapter1, in termsof policy beneficiaries rather thanproviders,a view thatbringstheNGOswhichhave taken the leadonmicrofinancewithin thescopeofmainstreamdevelopmentpolicy.29Thesecondformoftransferwashintedatinthelastparagraphwhenwetalkedaboutthe expansion of the Grameen Bank and the replication of its activities by otherBangladeshiNGOs.Thatphaseofthedevelopmentofpolicyiswelldiscussedundertheheading of scaling up in the NGO literature (see for example Uvin, 1995), butsomethinglikeitalsooccursinthepublicsector.ThusMelodescribesthewithincountrytransferofBrazilsincometransferpolicy,initiallyacrossmunicipalitiesandsubsequentlyup to federal government, and Jackson describes the difficulties encountered inMozambiquesattemptstoscaleupNampulasparticipatoryplanningmodel.PolicyRefraction:IncentivesandCongruenceIn Chapter 1 we talked about policy refraction, where policies are adapted to localconditionsratherthanadoptedwholesale.Ourcaseshaveprovidedtwoinsightsintothisprocess. The first concerns incentives. When policymakers are free to adopt or notadopt, incentives matter as much in transferring policies as they do in transferringfootballers.8 MelosargumentinChapter2 isthatthe incometransferpolicybenefitedfromabandwagoneffect,withfirstonemunicipalityafteranotherandthenthefederalgovernmentscramblingonboardbecauseof theelectoral returnwhich theyexpectedthis popular policy to give them. Similarly, the opportunities tomakemoneyandwinvotes were incentives for private companies and politicians to attach themselves tohousingpoliciesinMitlinschapter.ArguablyitisbecausesuchincentivesaremissinginMozambique that the Nampula participatory planning experiment has not transferredsuccessfullytootherprovincesortothenationallevel.Theproblemhere is that the incentivemaybesomuchbaggageweighing thepolicydown.WithincometransferinBrazil,makingpaymentsconditionalonschoolattendanceaddedatransactioncost,assomeonepresumablyhastomonitorattendance,adjudicateonappealsfromfecklessparentsandsoon.Moreproblematically,wehaveseenalreadyhowbringingprivatecompaniesandpoliticiansonboardskewedhousingprovisionawayfromsupporttocommunityselfhelpactivities.Thesecompromisesmayhavebeenwhatmadethepoliciesviable,buttheycameatacost.30That said, one way or another policies have to be brought in line with the broaderincentives and values that structure policy making. The tremendous transfer anddiffusion of microfinance, for instance, has occurred partly because its emphasis onindividual repayment and individual entrepreneurship9 chimed with the broadly neoliberaldiscourse thatdominates international financial institutions,ministriesof financeandsocialresponsibilityfoundations.Theemphasisonwomenwashelpfultoo(thoughitwouldappearsecondary). The transfer of importsubstituting industrializationpolicy inLatinAmericawasalsoaidedby itscongruencewithnationalist ideologiesof stateleddevelopmentthatunderlaybothmilitaryandpopulistgovernmentsoftheperiod.Finally,and perhaps most interestingly among our cases, the argument for basic incomeguaranteetransferstiedtoeducationresonatedwiththestructureofpoliticalincentivesinBrazil'sconsolidatingrepresentativedemocracy.Itofferedapolicywithpotentialvotewinningdividends, and thus became an object of partisan competition as each partytriedto leapfrog theothersbyadoptingthepolicy toreachouttoa largerpartof theelectorate.Theconsumersofparticularpoliciesarethusatdifferentlevels,andsoitisnotonlytheimpact onbeneficiaries thatwill determinewhether they are transferred. Themore apolicyresonateswithotherpolicystakeholders,particularlypowerbrokers,thelikelieritstransfer into new domains will be. In this sense, once again, design matters. Thepolicies that travelarenot only thosewhosedesign increases the likelihoodofhumandevelopment effectsbutalsowhosedesignmakes thepolicyattractive to its differentaudiences. The design challenge, here again, is to structure the policy process toempower groups committed to ensuring that the policy stays faithful to its originalobjectives. Shankland and Cornwalls account of the development of Brazils NationalHealthSystemshowsthatthiscanbedone,eveninthefaceofconsiderablepressures.CONCLUSION:INSTITUTIONALIZEDSOCIALENERGYTheframeworkdeveloped inchapter1wasmorecomprehensivethanparsimonious.Itgaveourcontributorsagooddealofroomtomanoeuvre,butasanexplanatorytool itwasunwieldy.Inthelightofourcasesandouranalysisoftheminthischapter,canweidentify a shorter list of factors that has greater explanatory power? Based on our31cases,theitemsonthatlistwillallconcernthepoliticsofpolicy,aboutwhichourcaseshave toldusmore thanweexpected,andless thanweexpectedabout itsmanagerialandprofessionalaspects.Inpartthatemphasismustreflectourcontributors(andtheireditors)personalpreoccupations.Wewouldnotwantreaderstogoawaywiththeideathattheprofessionalcontentorthemanagementofpolicydontmatter.Thatis,foronething, because in individual cases they have mattered a good deal (the content ofcustoms reform in Indonesia, the Grameen Banks human resource system inBangladesh). But it is equally because in the actual development practice ofgovernments like Malaysias and of development agencies (to take an example fromoutside this collection), we observe the everincreasing use of the techniques ofstrategicmanagement inperformanceindicators,PovertyReductionStrategyPapers,logical frameworksand soon. However, it is thepolitical issues that ourcontributorshavegivenusamandatetohighlight,andwedosonow.Takenasawhole,thecasessuggestthreemainexplanatoryfactors:power,leadershipandinstitutionaldesign.Wediscussthemnowinturn.Power:OrganizationsandCoalitionsThebalanceofpoweriscentraltoallsevencases.Itcreatesenvironmentsmoreorlesspropitious to different policies, and determines which groups and which ideas willprevail.Thepowerofdifferentactorsebbsandflowsoverthelifetimeofapolicy.Poweroverpolicy hasmany sources, but two seem tomatter agooddeal:organizationandcoalitions.Actorsgainpoweras theybecomeorganized:unityasstrength,notonly inthe teachers and other unions where that motto was coined, but also in squattersassociations,businesslobbiesandusers'committees.Gettingorganizedincreasespolicyleverage.The ability to develop policy coalitions (which our contributors have variously labeledalliances,epistemiccommunities,socialmovementsornetworks) isalso important (cf.Keck and Sikkink, 1998). Our contributors make clear again and again how such32coalitions serve as sources of power in policy arguments. In policy tussles, it iseffectivelycoalitionsthatdothetussling.LeadershipSecond, one cannot read these caseswithout being impressedby the frequencywithwhichleadershipemerges.Policysuccess isnotimpersonal,andcannotbeunderstoodwithoutreferringtonamedindividuals.However,aswellastheheroicandlonelyfiguresofpopular legend (thoughtheyexist),ourcasesdepict twootherstylesof leadership,which we called distributed and sequential. We also noted three different types ofleadership, transformational, transactional and technocratic, variously important atdifferentpolicystages.Allthesestylesand typesarerootedin thecoalitionsthattheypersonifytheydonotfloatfreelyabovethem.InstitutionalDesignThird,whileourcaseshavetolduslittleabouttheprofessionalcontentofpolicydesign,they have given us an insight into its institutional architecture. We have seen theimportance of conducting propoor policy in a way that empowers those who arecommittedtoit(and,bythesametoken,thatrelegatesthosewhoareopposed).Oncemore we highlight Brazils health councils and conferences as the paradigm case ofempowermentthroughinstitutionaldesign.Likewise,thepolicyneedstobedesignedtoprovideincentivestoindividualsandgroupswhoseparticipation iscrucialtorealizingapolicy,aswesawhappeninginBrazilwhenincometransferswerelinkedtoschoolattendanceanexamplewhichillustratesthattheincentivesneednottaketheformofcrudepersonalinducements.InstitutionalizedSocialEnergy:AnAccountoftheStagesofPolicySuccessInthelightofourcasesandouranalysisofthem, it ispossibleto identifyaschematicaccountofthestagesofpolicysuccesswhichincorporatestheabovethreeexplanatoryfactors and which seems to map quite well on to the cases (Table 2). We suggestinstitutionalizedsocialenergyasalabelforthisaccount,aslongasitisrecognizedthatitisashorthandforaprocesswhosestages,intermsofTable2,arethese:33socialenergyideacoalitionleaderinstitutionalizationfeedback34Table2: Institutionalizedsocialenergy:AnaccountofthestagesofpolicysuccessPOLICYSTAGE EXAMPLE THREATTOSUCCESSAnupsurgeofsocialenergy DemocratizationinBrazil(Chapter2) Policylackspopularrootsorisimposed(structuraladjustmentpolicies)Generatesapolicyidea MicrofinanceinBangladesh(Chapter5)orhighlightsanexistingidea IncometransfersborrowedfromtheBasicIncomeEuropeanNetwork(Chapter2)Sterileoppositionism(Chapter6)aroundwhichacoalitionassembles thecoalitioninNampula,Mozambique(Chapter8)Sabotageorcompetingcoalitions(Chapter6)whichthrowsupaleaderwhogetstheideaonthepolicyagendaSuplicyinBrazil,SalinasandotherLatinAmericanleaders(Chapters2and4)Leadershipisweak(successionofpresidentsinVenezuela,Chapter4)andovercomesoppositionfromsupportersoftheolddispensation.DeliberateweakeningofteachersunionsinLatinAmerica(Chapter4)Oppositiontoeconomicreformistoostrong(KiraguandMukandala,2004)Thecoalitionisinstitutionalized,empoweringbeneficiariesanddeflectingpatronsandrentseekersHealthcouncilsandconferencesinBrazil(Chapter7)Patronsorrentseekerscapturethepolicy(partialcaptureofhousingpolicy,Chapter6)andconsolidatedthroughfeedbacktoadaptthepolicytochangingcircumstances.InnovationinmicrofinanceproductsinBangladesh(Chapter5)Policyentropy(macroeconomicpolicyinIndonesia,Chapter3universalaccesseducationpolicyinLatinAmerica,Chapter4)35Thereare three items listedunder threats tosuccesswhicharenotselfexplanatory.First,ourreferencetosterileoppositionismisinfluencedbyMitlinsaccountofhowtheurban housingmovement in Mexico shifted from negative resistance to evictions andland taxcharges towardsactivedemands.Theslogan thatencapsulated theshiftwasProtesta conpropuesta: protestwith proposal. Sterile oppositionismwouldmean notgoingbeyondresistance:protestasin(without)propuesta,asitwere.Second,whilewewould love tobeable tosaypithily thatpolicysuccesscomes frominsidewhile policy failurecomes fromoutside, thatwouldexclude thecaseof incometransfers inBrazil,whichhadaEuropeanorigin.However,wehavenoted that itwasadopted voluntarily, just like all our successful policies and in contrast to structuraladjustmentpolicies(ourchosenexampleofathreatatStageOne),whichnotoriouslyfailed partly because they were not owned (McCourt, 2003). Thus the Americanhistorian Barbara Tuchmans (1984: 411) rhetorical question What nation has everbeenbuiltfromoutside?retainsitsforce.Third,sinceallthepoliciesdiscussedherewereactually implemented,wehavehadtogooutsideourcollection foranexampleoffailure toovercomeoppositiontoapolicy.Our example (from a distressingly large set of possibilities) is the economic reformfailure and the election defeat of President Soglo of Benin in the mid1990s, at thehands of an informal alliance of trade unions and opposition parties (Kiragu andMukandala,2004).Wewoulddrawtheattentionofconservativemindedreaders,whomayfeeltheyhavehad little to show for our enquiry up to now, that despite the propoor definition ofdevelopment which the authors of this volume have espoused, our institutionalizedsocial energy account turns out to be ideologically neutral, with equal application topoliciesofanypoliticalcomplexionincludingantipoorones.The distance we have traveled to arrive at this account becomes clear when wecompare itwith a standardmodel of thepublicpolicyprocess suchas thatof Jenkins(1978),withitsstagesofinitiation,information,consideration,decision,implementation,36evaluation and termination.10 In a nutshell, where Jenkinsmodel is rational, ours ispolitical(whichisnottosaythatpoliticalpolicyisnecessarilyirrational).However,wehope that our account is not idiosyncratic. In its stress on thesocial origins ofpolicyideasandcoalitions,itsharesmuchwiththoseaccountsthatemphasizethesociologyofknowledge and politics of ideas in policy processes (for example, Bebbington et al.,2006)andinitsstressonthemultifariousmembershipofpolicycoalitions,inwhichthestateisnotnecessarilytheleadingmember,italignsitselfwithHoutzagers(2003)polityapproach. This is apart from those writers we have referenced from whom we havedrawninspirationandnotleastfromthecontributorstothisvolume.FurtherResearchResearch ishydraheaded:we resolve researchproblemsonly tosee others sprout intheir place. The limitations of our research strategy are a specific reason for thathappening here. If the results of our enquiry have vindicated our decision to look atdevelopment policy through the lens of success, then first and foremost we call forfurther analyses of success. There are frequently cited examples of success aboutwhichwe know surprisingly little. Economicgrowth inMalaysiaandMauritius are twowhichhavebeeninvoked inthiscollection(thereareofcoursemanymoreexamples).We saw inChapter 1 that it isnotclearhowmuch the formerowes toaWorldBankstructuraladjustment loan in theearly1980sandhowmuch to theearlier indigenousdecisiontolaunchanallislandFreeTradeZone.NorisitclearhowweshouldcomparethelatterwithanaccountlikeHofmanetal.sinthisvolumeofgrowthinneighbouringIndonesia.Suchanalysesmaywishtousethecriteriaforsuccess(durabilityetc.)andatleast the standard of evidence that we have established here, but theywill take ourenquirytoadeeperleveliftheyarebasedonfreshprimaryresearchratherthanlookingsidewaysattheirauthorspreviousresearchasourcontributorshavedone(andasweintendedthattheywould).Paraphrasing Mahatma Gandhis famous remark, our contributors tend to agree thatdemocraticdevelopment wouldbeagoodidea: they remaindemocratsatheartevenwhereitisauthoritariansuccessthattheydiscuss(Hofmanetal.andGrindle).Thereiseven the hope in Jacksons chapter that policy success will strengthen democratic37legitimacy. While there may be no statistical correlation between democracy andgrowth, as Mushtaq Khan has pointed out, we can still seek examples beyond thepossibly exceptional case of Brazil where democracy has gone hand in hand withdevelopment success, bucking the patronage and rentseeking trend about whichauthors like Khan and Anne Krueger are so pessimistic. Providing an answer to thissecond researchquestionofourswill throw lighton the very largebutvery importantissueoftherelativeviabilityofstateledandNGOled(nottomentionprivatesectorled)development.Athirdresearchquestionwhichouremphasisondevelopmentpolicyinthelongrunhascreated somewhat to our editorial surprise concerns the trajectory ofimplementation. Arepolicies doomed toatrophy, as Grindle implies theymust, or aretheyable to renew themselves,asHulmeandMoore imply that theycan? Again,weneed examples to supplement Hulme and Moores account of microfinance inBangladesh.Inthiscontextitwillbeinterestingtoseeifempoweringthosewhosupporttheobjectivesofpropoorreformandusingtheirfeedbacktoadaptpolicytochangingcircumstances is as important as our casessuggest. Theremaywell be other factorsthat we have overlooked, possibly those factors of professional content or policymanagementaboutwhichwehavehadsolittletosay.The fourthand finalquestionconcernstheoriginofpolicy ideas. Thecreativefermentfromwhichtheyemergetakesplacebeneaththepolicyanalysisradar.Forexample, ittookagoodnumberofyearsformicrofinanceinBangladeshtocometotheattentionofscholars,bywhichtime thegestationwas longover. In thesocialcapital tradition,weknow quite a lot about thesocialmovements fromwhichpolicy emerges, andat theotherendofthepolicyprocessweknowquitealotaboutpoliciesoncetheyareinthepublicarena.Butbetweenthosetwoextremesweknowlittle.ItwouldbeveryusefultohavemoreaccountsthatareasclosetothedataasMelosandJacksonsarehere.SuggestionsforPolicymakersWecomenowtothepracticalimplicationsofourfindings,eventhoughthesuggestionswe will make come with a health warning because of the nature of our research38strategy.Ourexpandedviewofthepolicycommunitymeansthatwewilladdresssocialactivists as well as the government and development agency officials to whomrecommendationsinvolumeslikethisareusuallyconfined.State officials, both elected and appointed, should develop antennae that will enablethem to scan for winners onwhich to base their policies in the everbubblingpot ofsocial ideas and inventions: outside agencies may be more fertile sources thangovernment ministries and departments. Officials will increase the likelihood of theirpolicies materializing and sticking by making common cause with groups outsidegovernment who share their commitments, and by giving them a formal institutionalrole.Suchgroupscanbeasoundingboardfortheappropriatenessofpolicies,allowingrunningrepairsthatwillkeepthepoliciesoncourse.Itiswisetogiveasfewhostagestofortuneintheformofconcessionstohostilegroupsaspossible.Social activists should think whether counterparts success in influencing governmentpolicy inBrazil,Mozambiqueand (for housing policy) Chile, thePhilippines andSouthAfrica is a precedent formoving from opposition to constructive engagement fromprotestatopropuestaandunitinginapolicycoalitionwithlikemindedstateofficials.The frustrations ofworkingwith thestatewill besubstantial,butsowill thepotentialrewards.Activistsshouldchannelat leastsomeoftheirsocialenergy ifatallpossibleinto state structures, and avoid the vicious circle where dissatisfaction with statecorruption and inefficiency leads to setting up parallel structureswhich further erodestatecapacity,andsoroundagain.Theyshouldstayengagedthroughimplementation,firstlybylobbyingtobeincludedinaformalinstitutionalstructurethatwillhelpensurethatthepolicy isnothijackedbypatronsandfreeloaders,and thenbygivingfeedbackthatwillcontributetothepolicystayingontrackratherthanatrophying.Development agencies should consolidate the view to which many of them alreadysubscribethatsuccessfulpolicycomesfrominside,andinnormalconditionscannotbeinducedbyanoutsideagencyhoweverbitterapill thatmaybeforsomeagenciestoswallow. Since they are committed to supporting whatworks, they should use their39purchasingpowertocommissionstudiesthatwillhelpthemunderstandlongrunpolicysuccessandpickwinners.Forallthreealike,itisimportanttoidentifythepolicyleaderwhoatthepolicyinceptionstagewillbethebatteringramthatgetsapolicyontothestatutebook.Andlastly,theleaders themselvesshould recognizethattheirrole is topersonify thepolicy,andstayclosetothepeoplewhoputthemwhere theyare.TheyshouldhaveMandelaandnotMuzorewaastheirrolemodel.CelebratingSuccessInhiseulogyfortheIrishpoetWilliamButlerYeats,W.H.Auden(1966:143)saidthatitwaspartofapoets job to teach thefreemanhowtopraise.11 Sobersidedpolicyanalystsaswellasecstaticpoetsdowelltorememberthatgivingcreditwherecreditisdue is part of our job. The seven cases in this collection, despite their scrupulousauthorsmanyandproperqualifications,provideseriousevidenceofpolicysuccessonalarge scalewhichhasbeensustained for a decadeor longer, often against theodds,evidencethattakestheformof impactonhumandevelopmentandsocialandpoliticallife. Even professional students of development are not immune from the prevailingimageoftheThirdWorldasaplaceoffamine,fireandflood,anditissalutarytoremindourselvesoccasionallyofwhat hasbeenachieved indevelopment policy,and to learnanddrawinspirationfromthereminder.Thatisthesimplethoughtwithwhichwebeganthisenquiry,andwithwhichwenowbringittoaclose.40ReferencesBarry, D. (1991) Managing the bossless team: Lessons in distributed leadership,OrganizationalDynamics,Vol.21,pp.3147.Bebbington, M.Woolcock, S. Guggenheim and E. Olson (eds) (2006) The Search forEmpowerment: Social Capital as Idea andPracticeat theWorldBank,Bloomfield:Kumarian.Brown, M. andD. Hosking (1986) Distributed leadership and skilled performance assuccessfulorganizationinsocialmovements,HumanRelations,Vol.39,pp.6579.Burns,J.(1978)Leadership.NewYork:HarperandRow.Case,W. (1995) Malaysia:Aspectsandaudiences of legitimacy, inM.Alagappa (ed.)Political Legitimacy in South EastAsia, Stanford: Stanford University Press, pp.69107.Crouch,H.(1996)GovernmentandSocietyinMalaysia,Ithaca:CornellUniversityPress.Dagnino, E.,A.OlveraandA. Panfichi (2006) Paraumaoutra leituradadisputapelaconstruodemocrticanaAmricaLatina,inE.Dagnino,A.OlveraandA.Panfichi(eds)Adisputa pelaconstruo democrtica na Amrica Latina, SaoPaulo: Paz eTerra.DeLeon,P.(1978)Publicpolicytermination:Anendandabeginning,PolicyAnalysis,4:pp.36992.Evans, P. (Ed.) (1996) StateSociety Synergy: Government and Social Capital inDevelopment,Berkeley,CA:InstituteforInternationalStudies.41Evans,P.(1995)EmbeddedAutonomy:StatesandIndustrialTransformation,Princeton:PrincetonUniversityPress.Frank,J.(2002)Politicalresourcesforpolicyterminators,PolicyStudiesJournal,Vol.30,pp.1128.French,J.andB.Raven(1959)BasesofSocialPower,inD.Cartwright(Ed.)StudiesinSocialPower,AnnArbor:UniversityofMichigan.Guggenheim, S. (2006) Crises and contradictions: Understanding the origins of acommunity development project in Indonesia, in A. Bebbington,M.Woolcock, S.GuggenheimandE.Olson(eds)(2006)TheSearchforEmpowerment:SocialCapitalasIdeaandPracticeattheWorldBank,Bloomfield:Kumarian,pp.10944.Fox,J.(1996)Howdoescivilsocietythicken?ThepoliticalconstructionofsocialcapitalinMexico,WorldDevelopment,Vol.24,pp.10891103.Haas, P. (1992) Introduction: Epistemic communities and international policycoordination,InternationalOrganization,Vol.46,No.1,pp.136.Heclo,H.andA.Wildavsky(1974)ThePrivateGovernmentofPublicMoney:CommunityandPolicyInsideBritishPolitics,London:Macmillan.Hill,M.(2005)ThePublicPolicyProcess,Harlow:PearsonLongman.Hirschman,A.(1984)GettingAheadCollectively,NewYork:Pergamon.Jenkins,W.(1978)PolicyAnalysis,London:MartinRobertson.John,P.(1998)AnalysingPublicPolicy,London:Pinter.42Khan,M. (2005) Markets, statesanddemocracy:Patronclient networksand thecasefordemocracyindevelopingcountries,Democratization,Vol.12,pp.70424.Kiragu, K. and R. Mukandala (2004) Pay Reform and Policies Report, Paris: OECDDevelopmentAssistanceCommittee.Krueger, A. (1974) The political economy of the rent seeking society, AmericanEconomicReview,Vol.64,pp.291303.Mahoney, J. (2003) Longrun development and the legacy of colonialism in SpanishAmerica,AmericanJournalofSociology,Vol.109,pp.50106.Rogaly,B.(1996)Microfinanceevangelism,destitutewomenandthehardsellingofanewantipovertyformula,DevelopmentinPractice,Vol.6,No.2,pp.10012.Sato,H.(2002)AbolitionofleprosyisolationpolicyinJapan:Policyterminationthroughleadership,PolicyStudiesJournal,Vol.30,pp.2946.Stone, D. (2000) Nongovernmental policy transfer: The strategies of independentpolicyinstitutes,Governance,Vol.13,No.1,pp.4562.Stone,D. (2002) Using knowledge: The dilemmas of bridging research and policy,Compare,Vol.32,No.3,pp.28596.Tendler,J.(1997)GoodGovernmentintheTropics,Baltimore:JohnsHopkinsUniversityPress.Tuchman,B.(1984)TheMarchofFolly:FromTroytoVietnam,London:MichaelJoseph.Tullock,G.(2005)PublicGoods,RedistributionandRentSeeking,Cheltenham:EdwardElgar.43Tullock,G.(1967)Thewelfarecostsofmonopolies,tariffsandtheft,WesternEconomicJournal,Vol.5,pp.22432.Uphoff,N.(2000)Demonstratedbenefitsfromsocialcapital:TheproductivityoffarmerorganizationsinGalOya,SriLanka,WorldDevelopment,Vol.28,pp.18751900.VanDomelen,J.(2006)Socialcapitalintheoperationsandimpactsofsocialinvestmentfunds, inA.Bebbington,M.Woolcock,S.GuggenheimandE.Olson(eds.) (2006)TheSearchforEmpowerment:SocialCapitalasIdeaandPracticeattheWorldBank,Bloomfield:Kumarian,pp.177203.Wade,R.(1990)Governingthemarket:EconomicTheoryandtheRoleofGovernmentinEastAsianIndustrialization,PrincetonN.J:PrincetonUniversityPress.WorldBank(1993)TheEastAsianMiracle,WashingtonDC:WorldBank.Zakaria,H. (1989) Malaysia:Quasidemocracy ina divided society, in L.Diamond, J.Linz and S. Lipset (eds) Democracy in Developing Countries: Asia, Boulder, CO:LynneRienner,pp.34781.44NOTES1Malaysiaduringtheperiodinquestionhasbeenvariouslylabeledaquasidemocracy(Zakaria,1989),semidemocracy(Case,1995)andperhapsmostexpressivelyarepressiveresponsiveregime(Crouch,1996)2AlthoughMeloalsowarnsagainstthequasipsychologicaldiscussionofleadershipinwhichweareabouttoindulge.3ItmaybeusefultolisttheitemsofFrenchandRavensclassictaxonomyofindividualpower.Aswellasexpertpower,theyarelegitimatepower,rewardpower,coercivepowerandreferentpower(anindividualsabilitytogeneraterespectoraffection).4Theproblemmayalsobecompoundedbyalackofwidelyacceptedmodelsofsocialreform(seeChapter2onthispoint).5Jacksonschapterisinathirdcategory,asanexampleofreversecausation,whereitishischosenpolicythatstrengthensdemocracyratherthanviceversa.6 ItisworthnotingthatHulmeandMooreareascriticalofBangladeshsprivatebanksastheyareofthegovernment.7Hofmanetal.schapter,withitsaccountofaidamountingtoover30%ofgovernmentspendingin1967,mightseemanexception.Butevenhere,Hofmanetal.saythatdonorinfluenceoneconomicpolicyinthatperiodcamepartlyintheobliqueformofFordFoundationscholarshipsthatallowedthetechnocratstostudyintheUSandofcoursetheIMFsroleinthe1997Crisisremainscontroversial.8WearegratefultoAnneMarieGoetzforsuggestingthispoint(thoughnotthefootballanalogy)atthefirstoftheseminarsonwhichthiscollectionisbased.9Notwithstandingtheemphasisongroupsolidarityandgroupbasedguarantees,microfinanceschemesproduceindividualentrepreneurs,notgroupbasedenterprise,andfarlesssocialmovements.10MorerecentaccountshavechallengedsystemsmodelslikeJenkins,butmostlybecausetheyareseenasmechanisticandoversimple,ratherthanbecausetheyareapolitical:seeHill(2005).11AswiththereferencetoPaulSimonatthestartofthisvolume,AudenspoemwasreadilyavailableinfullontheInternetatthetimeofwriting.

Recommended

View more >