7/28/2019 Comparative Decentralization Lessons http://slidepdf.com/reader/full/comparative-decentralization-lessons 1/15 Comparative Decentralization Lessons from Pakistan, Indonesia, and the Philippines Author(s): George M. Guess Reviewed work(s): Source: Public Administration Review, Vol. 65, No. 2 (Mar. - Apr., 2005), pp. 217-230 Published by: Blackwell Publishing on behalf of the American Society for Public Administration Stable URL: http://www.jstor.org/stable/3542555 . Accessed: 07/03/2012 04:53 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Blackwell Publishing and American Society for Public Administration are collaborating with JSTOR to digitize, preserve and extend access to Public Administration Review.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Comparative Decentralization Lessons from Pakistan, Indonesia, and the Philippines
Author(s): George M. GuessReviewed work(s):Source: Public Administration Review, Vol. 65, No. 2 (Mar. - Apr., 2005), pp. 217-230Published by: Blackwell Publishing on behalf of the American Society for Public AdministrationStable URL: http://www.jstor.org/stable/3542555 .
Accessed: 07/03/2012 04:53
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].
Blackwell Publishing and American Society for Public Administration are collaborating with JSTOR to
digitize, preserve and extend access to Public Administration Review.
Thisarticleprovidesan analytic rameworkoguide regimes hatare designingor implementingdecentralizationrograms. t s based on a comparison f threeAsian cases of fast-track ecen-
tralization.The rameworkuggests hatregimescontemplatingevolutionmust ace fundamental
issues of (1) background upport, 2) cultureand institutions,nd (3) technicaldesign and se-
quencing. tcanbe usedby regimes ocompare herelative ifficultyf fundamentalchallenges o
decentralizationwiththeirown capacityand potential or effectiveresponse.The hreeregimes
ent responsesof thatregimeare significant.More research s required o explaindifferencesn
technicalperformancenthePhilippinesnd othersimilarprogramsand to attributemeasuresof
decentralizationuccessto thesedifferences.
IntroductionRegimes that are contemplatingdecentralizationpro-
gramsare often reluctant o take the political risk of un-
knowntechnicalconsequences n designand mplementa-tion.Regimesknowthatpolicies are often based on myth,and decisions on imperfect information.Policy makers
would like to eliminatemythsand reduceuncertaintybe-
fore proceedingwith programssuch as decentralization.
In 1996, policy makersat the AlbanianMinistryof Inte-
rior asked me to provide comparativeregional informa-
tion on how to preventfiscal transfersfrom acting as a
disincentive to local revenue mobilization. Without this
information,which was unavailableat the time, the min-
istry's state secretary or local governmentwas reluctant
topushfor devolution.Other egimeshavealso found hem-selves at this point, and they need an empiricallybased
guide or framework or the successful design andimple-mentationof decentralization.Considerable advice has
been provided by internationaldonors, consultants,and
universitypersonnelon decentralization.But at this stage,
regimesneed morethantechnical advice on single issues
such as fiscal transfersor tax policies.This article provides a preliminary but empirically
groundedframework or regimes that are contemplatingdecentralization.This framework,which is based on the
experiencesof Pakistan, Indonesia,and the Philippines,
predictsthatregimes will face both generalbackground
challenges (top support,decision capacity,and political
atedby citizen groups throughsurveysand social audits.This should lead to greateraccountability ndresponsive-ness of elected andappointedofficials to local citizens.Fi-
nally,decentralizationhould ead togreaterocalauthorityto hire andfireappointed fficials,meaninggreater espon-siveness of local governments o citizen needs. Officials
would no longerbe appointedrom thecenterandpaidfor
by the localgovernmentsWinklerandHatfield2002).
Table1 ComparativeDecentralization erformance
Shareof subnational SNG sets tax
own-revenuesntotal base and raterevenues**percent)
Philippines 42.8
Pakistan
Yes
Yes
No
5.0
Indonesia 25.4
SNGauthorityo
setbudgetpriorities
Yes,butcentral
governmentstill armarks
Yes
No
SNGmayor/
council elected
Yes
Nazim bycouncil
Mayor bycouncil
SNG candesign/
implementrg-staffingorservicereqs
Yes
No
No
Dedicated nd
rule-basediscaltransfer/tax
Yes
Yes
No
"*Source:overnmentinance tatistics earbook001 (Washington, C: nternationalonetary und).With heexception f Pakistanorwhich ocalstatistics avenotyetbeenreported,hemeasures f local revenue ufficiency ereobtainedby dividing otalrevenues nd grantsby revenuesromother evelsof national overnmentnd
multiple measuresfor decentralization.They know it isefficient to decentralize,but majorinefficiencies can re-sultfrompoorly designedand mplementedprograms.For
instance,they could lose theirelected or appointed obs,andthecountry oulddescend nto(perhaps reater) haos.What shouldtheydo?
ComparativeethodologyTo provide applied lessons to policy makerswho are
deciding whether or how to engage in decentralization,
comparative ield lessons areneeded from similarcases.The matched-case method of comparison that Xavier
(1998) uses to compareMalaysianandAustralianbudgetreforms is useful for this purpose.Thatmethod was usedin this researchto select countries with similarintergov-ernmentalstructures hat have institutedsimilar reform
multiple causation,which otherwisebesets comparativeanalysis. By picking decentralizationcases with similar
key variables,this studyfocused on the measures,condi-
tions,andvariables hat aredifferent.Focusingon the dif-ferences of matchedcases allows for a morerigorousex-amination of factors that may account for variationindecentralizationprogramresults.2
BackgroundndContext nMatched
Program asesThethreeprograms resimilarnthat heyare"fast rack."None of them was donordriven,andinternational onorssuchas theAsianDevelopmentBankandtheU.S. AgencyforInternational evelopment ontinue oplaycritical up-portingroles. Forthepurposesof matchingcases,these aresimilar egional xamplesof once-centralizedtates hathaveevolvedintofederalgovernance ystems.Throughhis evo-
ployment,andbudgetdeficits(Guess,Loehr,andVazquez1997, 39). Stabilitycan be used as a resourceby central
regimes oreallocate"policy atitudeor"agencyautonomy"to its counterpartsn lower-tiergovernments Peters1978,
168).Inpoliticalandeconomicstability,hePhilippineswasrankedenthmoststableandIndonesia ighteenth yEurasia/LehmanBrothersEconomist 003, 100).Economicgrowthremains trong nbothcountries,withgrossdomesticprod-uct increasesof 5.8 percentand 3.9 percent,respectively,from the same quartern 2002. In the contextof relative
Despite recent terrorist-control roblemsand internalcriticismfroma vibrant ocal press, the Pakistaniregimecontinues to governthroughcivil serviceandmilitary n-
stitutions,with reasonablesuccess. Recent changes and
securitythreatssince September11 have kept PresidentMusharraf's egimein powerandhave increasedsupportfor devolution.Fearof regionalfragmentationn light of
growingpovertyand nequality-partlyattributableopoorlocal services-has been high among all three regimes.Pakistan'sprogram bandoned econcentratedubnational
governance onsistingof provinces,divisions,anddistrictsandcreatedthreenew levels of countrywideelected gov-ernment:districts(96), tehsils, or towns (337), andunioncouncils (6,022). The program mplementsthe intent ofthe May2000 local governmentplan.The operatingprin-ciple is to transferpowerdownward o citizengroupsanddistrict-levelgovernments o provideandfinanceneededbasic services thatwere not being providedby eitherthe
provincesor the centralgovernment(NRB 2000, 9). An
impetus orthedevolutionprograms that hecentralgov-ernment-the National ReconstructionBureau-rightlyfeared hatpopulardissatisfactionwithitsperformancewas
growingandthatthis threatened he state'slegitimacytohold the nation, four large and dissimilarprovinces,to-
gether.Powerwas formallytransferredo thenew districttier of governmentnAugustof 2001. In 2002, it was esti-matedthe districtsgeneratedan averageof 5.0 percentoftheirown revenues(table 1). Thatis, 95 percentof theirfunds were still providedfrom transfersand shared rev-enue sources.3
The Philippinedecentralizationprogramhas been de-scribedas one of the most farreachingin the developingworld. A long traditionof political-administrativecen-tralismexisted in the Philippinesthatwas initially chal-
lengedby theconceptof local autonomy,enshrined n theConstitutionof 1987 and laterin the Local GovernmentCode of 1991 (RepublicAct 7160). This was the fifth at-
temptsince 1946 to empowerthe fourlevels of local gov-ernmentwith political andadministrative uthority YapandSator2001, 1).
Theprogrambeganwiththeimplementation f theLo-cal GovernmentCode in 1991
(Galang2001, i)and ex-
pandedthe responsibilitiesand authorityof the 76 prov-inces, 1,540municipalities, 6 cities,and42,000 barangays(precinctsorwards).Themajorchangewastheexpansionof city, municipality,and barangay urisdictionover the
norscontinue o support heprogram,withpositiveresults.
The Pakistaniand Indonesianprogramsare similar n that
both devolvedauthority o local governments, argelyby-
passing the provinces and shifting large groups of civil
servantsto the districts to provide them with technical
capability.To summarizeprogramsimilarities,all three countries
devolved authority or programsthat were formerlythe
responsibilityof higher-tiergovernments.The three re-
gimes (Philippines,Pakistan,andIndonesia)decided to go
full speed aheadon reformsby (1) devolvingmajorfunc-tions, (2) assigningrevenueauthorityandblock grantfi-
nancing,(3) transferring adres of central andprovincialofficials andpositions to administer he decentralization,and(4) organizing ocal electionsfor councilsandmayorsto hold local administrations ccountable.All are financ-
ing the devolution and new responsibilitieswith entitle-
ment sharesof higher-tierrevenues(Pakistan's"divisible
departments,downsizingandrefocusingthem accordingto a new distribution f functions.This enablesprovincestomonitor, egulate,andsupervise ocal districtoperationswithoutcontrolling heir affairs(WorldBank2002a, 10).
Third,initial resistanceto decentralizationwas strongin all three countries,based on the argument hat locals
were not to be trustedwith central funds and should bemonitoredclosely. Pakistanhas initiated at least four re-
forms n the name of empowering ocalgovernments ince
the 1950s.All were incompletereforms and were largelyreversedby laterregimes,whichargued ora lack of local
gumentshave been raised aboutthe local capacityto ab-
sorb fiscal and technical resources. Several international
evaluationsconcluded herewaslittleadvancepreparationfor decentralizationnd hat ocalgovernmentswere argely
unpreparedor theirnew tasks(Alm,Aten,andBahl2001,
5). Coupledwith the traditionof "waitingfor the centralgovernment o act"(Alm, Aten, andBahl, 2001, 9), top-level supportwas neededto overcome theperception hat
capacity problemswere so severe as to jeopardize he en-
Culturemay be turnedfrom a static obstacle to the dy-namicfoundationon whicha decentralization rogram an
be built.5
Civil Society
Culturalandcivil society issues are relatedin that re-
gimes governing n centralizedculturesgenerallyrepress
thedevelopment f intermediaryivil societyorganizations,such as themedia,unions,citizenassociations,andpoliti-cal parties.Opponentsof decentralization rgue hat f the
goal is democraticdecentralization nd the clustersof in-
terestsaremissing or of the wrongkind, it is difficult to
developrepresentative oliticalparties n anybutthe nar-
rowest sense, that is, corporatistpartiesrepresenting he
greaterparticipationand associational activity at lower
levels (1999, 57). In his view, decentralizationprogramscan makeentirely adequateachievements n the absence
of lively civil society and can make up for them in the
shortterm. The absence of full civil society need not be
an obstacletoprogrammplementation.6 ecentralization
programsneed to identify these useful culturalpracticesandbuildonthem,rather hanrelyingon systemsimposedfrom the outside,often from a combinationof poorlyde-
signed donorprogramsand importedconsultants. Evenin EasternEuropeand the former Soviet Union, where
civil society did not formallyexist, local partyorganiza-tions provided eadershipskills to members thatare now
beingputto use in thedevelopmentof civic institutions n
thatregion(Jacobsen2000). Based on these conclusions,the Pakistani egimeincludedcivil societycreationaspartof its devolutionprogram.
Institutional Systems and Practices
Institutionsare the formal andinformalrules, the sys-
ineffectiveprovincialgovernmentsandto give powerand
fundingto the new local tier.
Nevertheless,the overallapproachof the NationalRe-
constructionBureauhasbeentop-downplanning,withan
emphasison rules, laws, andregulations romthe bureau
and the provincial levels. Very little emphasis has been
given to the needed local managementflexibility to de-
liver theneeded services.By contrast, hePhilippine egalandregulatory pproach as beenmorefacilitativeand ess
commandandcontrol.Conflicts of law have been recog-nized and an institutionalmechanismproposed o remedythe problem.Indonesian aws havenot assignedclearex-
penditureand revenueroles for local governments.The
laws also excluded the provincesfrom the devolution of
authority.Of the threecases, the Philippineshave givenmost attention o the developmentof the legal andregula-
toryframework.
The thirdset of regimeactivitiesfocuses on theresolu-
tionof localcapacity
weaknesses.Regimescontemplatingdecentralizationhave to deal with this twice: initiallyas a
background ssue, andlateras an operational ssue to be
remediedby technicalassistance,training,andimprovedincentives.All threeregimes dealt with this issue by cor-
rectlydefiningit as a short-termssue thatcould be rem-
edied by a promise of supportand by deployinghigher-tier staff to local units. The Philippines and Indonesia
required ocal units to pay for salaries out of fiscal trans-
fers. None of the programsrequired ocal civil services,which createdtension with centralandprovinciallevels.
Only the Philippineshas madesignificantstrides n mod-
em performance-based apacity building.Pakistan's ca-pacity building is largely in-house and derived from ar-
chaic colonial-era institutions and practices. The
Philippineshas given themost attention o this issue frominitiationof theprogramn 1992. Forexample, texempted20 percentof the localdevelopment und fromcentralear-
marks.While Indonesiadeployedabout two million staffto localunits,thegovernmentof Indonesiadidnotprovidethemwith any incentivesfor capacitybuildingor the ap-
plicationof skills to new local systems.It was found that the fourthstep followed by the re-
gimes was toestablishfacilitationnetworks o supportandsustainthereform.Without ocal supportgroupsand citi-zen participation,decentralizationreform typically re-mains unimplemented egislation. In Pakistan,a militarygovernmentwith an initially shortoperating ife initiatedthe reform. It was directedby the NationalReconstruc-tion Bureau with little or no civil society or associational
backing.This was due to the rapidplanningrequirement,the knowledgethat the provinceswould attempt o derail
the reform,the recognitionthatmost districtswere tech-
nically deficient and could not make sound evaluations,
and the absence of any real civil society or independent
professional sector in Pakistan.The Local GovernmentOrdinance(SBNP 2001) establishedcitizen communityboardswith entitlements to 25 percentof local develop-mentbudgets.Thisinstitutionalmechanismwas anattemptto break he lock on local politicsby traditional lites and
partiesand to create a nascentcivil society organization
responsible for developmentprojects. Civil society has
beenrepressedand s still weakin Pakistan.Pakistanheldlocal elections (fornazims,ormayorsandcouncils) earlyin the program(2000 and 2001), and this generatedex-
cessive local demands hatthreatened hecapacityof new
districts to respond.
By contrast,the Philippineshad a large civil societystructure o drawupon;thegovernmentof thePhilippinesdevelopeda broadsupervisorygroupthatincludedrepre-sentatives romcivil society andprofessionalgroups.The
supervisorygroup ncluded heDepartment f Interior ndLocal Government,Departmentof Budget and Manage-
ment,Departmentf
Finance,Civil Service
Commission,the EconomicandDevelopmentAuthority,andthe Com-mission on Audit.Associations includedthe Association
of LocalBudgetOfficers.Indonesiareliedon several ocal
government associations, including APPSI-provincial,APEKSI-localand district,andAPKASI-mayors.Thesehavebeen important upporters,buttheyhavenot servedas devolutionprogramadvocatesso far.Partof this mayoccur in the shakeoutperiodwhich, unlike Pakistanand
Indonesia, he Philippineshas alreadyexperienced.In the fifth step, regimesfocused on the establishment
trydidthis;the othertwo madelargely symbolicmovestocreatemonitoringand evaluationsystems. The National
ReconstructionBureauin Pakistanplanned a full-scale
implementationwithoutpilotprograms. t madeno effortto establishanymonitoringsystem,perhapsout of the ex-cessive hubristhat it could simply dictateprogrameffec-
tiveness. By contrast, n 1992 the PhilippinesestablishedtheRapidFieldAppraisal ystemtogenerateneedsassess-
buildingapproach,which attempted o avoid failed sup-ply-driven rainingapproaches. ndonesiamadeno majorefforts in this area,otherthan the 1994-95 District Au-
tonomyPilotProgram,whichwas used to assess local ca-
pacity to execute newly devolved tasks priorto the pro-gram.Negativeconclusionsgenerated ythisprogramwerenot utilized by the governmentof Indonesia in programplanningor implementation.
In the sixth activity,all threeregimes recognizedthe
need to providestablefinancing f local autonomywas tobecome a reality.All threeattempted o deal with the fi-
224 Public dministrationeview March/April005,Vol. 5,No. 2
cation)to ensurefundingfor primaryandsecondaryedu-cationneeds at the local level.But there s still noagreementon theprovisionof stableannualamountsof fundingorits
basis (thatis, a performance ormula)from the National
and forecastingerrors,this percentagedoes not indicate
the actualyield for local governments.Indonesiancentral ransfersprovide25 percentof local
revenues.No othermajor
revenue ourcesareassigned
and,
giventhe absenceof anyreal tax base in most areas, ocal
own-source revenuesarerelatively high for the coverageof local needs (25.4 percent,see table 1).The assignmentof revenues has not been based on the local cost of ser-
vices estimates (that is, a target level of expenditures),which shouldhavebeen done first.This means the center
retainscontrolof revenues. In the Philippines,the earlyinternal evenueallotment,requiredby the 1987 Constitu-
tion,provided unds o localgovernmentunitsbutexcluded
the costs of the devolution.This deficiencywas remedied
by the 1994 "cost of devolved functions"mechanism in
health(WorldBank 1994).Workremains o be done in allthreecountriesonimproving hestabilityandperformancebasis of funding orlocal governments.Thelong gestation
periodin the Philippines suggests that fiscal stabilityis a
medium-termssue requiringa longershakeoutperiod.The seventh activity is restructuring nd streamlining
local offices to prepare hem for the performanceof their
newly devolved tasks. The NationalReconstructionBu-
reau has given considerablethoughtto this in Pakistan,andeffortshavebeen madeattheprovincialevelto stream-
els havebeen moreintermittent,which could impedeser-vice delivery.The bureauhas notprovidedmethodological
optionsfor functional review or reorganization,nor have
regulatoryrelationshipsbetween provincialdepartmentsand corresponding local group offices (such as publicworks)been worked out clearly.In the Philippines, ocal
to outputsandoutcomes. This is somewhatsurprisingbe-cause one would expect results budgetingto be first, as
core systems such as transfersanddevelopmentprojects
depend on their information.Instead,all regimes calcu-
lated thatexisting informationcould be used for results
purposes, orinstance, ine items to controloperationsand
maintenancexpendituresnd o gauge sufficiency orpres-ervationof capital stocks. As results-orientedbudgeting
requires mprovementsn analyticalskills and thegenera-tion of performance eporting nformation,t is more of a
medium-term ctivity,and hence last on the list of regime
priorities.In its short
programhistory,Pakistanhas done
much in this areaat theprovincial evel, and this servesas
a model for local reform.At the local level, efforts sup-
portedby the AsianDevelopmentBankto develop perfor-mance systems have been constrainedby the need to en-
sure consistency with past forms, accounting practices,manuals,and administrative ractices.There are no local
professionalpressuregroupsadvocating hangeshere(thecivil society constraint).The lead agency for reform,the
National ReconstructionBureau, s hamperedby inexpe-riencein this area,togetherwith a mostlyconsultantstaff
andtop-downchain of command.ThePhilippineprogram
utilized heAssociationof LocalBudgetOfficersandotherprofessionalassociations to changelocal practicesand to
upgradeanalyticalskills. The Philippinecentralgovern-menthad substantial xperienceat thecentral evel in per-formanceandprogrambudgeting.Despiteanemphasison
local borrowing consideredsophisticatedby most),Indo-
nesia has not yet focused on performancereportingand
budgeting.
Differences
Severalconclusionscan bedrawn rom hedifferencesn
program mplementation.The differencesappearedess inthe sequenceof activities han n theiractualperformance.As indicated ntable2, sevendifferences an benoted nthe
performanceof technical activities at the organizationalframework nd mplementationtrategyevels.
First,thesupervisory tructure ordecentralization ar-ied amongthe threecases. The Philippinesbeganwith a
broad-basedstructure hatincluded multiple representa-tive interests.This extended to the use of civil society in-
stitutions for feedback and guidance on the progressofthe reform.By contrast,the supervisorystructureof the
Pakistanreformhas been limited to the militaryregime'sclient institution, the National ReconstructionBureau,whichrepresents ew interestsand s staffedby manycon-
sultants n a loose accountabilitystructure.Pakistanhas aweak civil society network;the reformwas intended to
projects,called citizen communityboards. There is still
fear these boardsmay entrench local moneyed interests
rather hanempowerthe poor,because it is the rich who
will have time todevelopthe technicalexpertiseto attract
district council funds (WorldBank 2002b, 23). As ex-
pected, citizen communityboards have not yet become
operationalandareexperiencingproblems n competitionwith otherlocally powerful groups.The scope of the In-
donesiansupervisory tructure alls betweentheother wo
programs-some representationat the top (the Office of
Decentralized Local Government),but more is needed
from civil society organizationssuch as the professional
municipalorganizationsAPEKSI andAPKASI.It is clear
that, after the establishment of macroeconomiccontrol,all threeregimesmade efforts to establishsupervisory n-
stitutions,but each variedin its emphasison representa-tion and the need for societal feedback.
Second,as indicated n table2, decisionsof supervisoryinstitutionsmustbe guidedby sensible lawsrelating o tax,
budget,debt,participationn governance,and the assign-ment of functionalroles. Decentralizationaws mustbe in-
formedby a full understanding f whatcan go wrong in
publicandmunicipaladministration,swell ashow topro-videspeedyremedies orconflictswithother awsandregu-lations.Many aws have been writtenby generalists rom a
top-downplanningperspectivethat contain few insightson theneed to deliver ocal services.In thedevelopmentof
the LocalGovernmentOrdinance,hePakistani egime n-
cluded few internationalxamplesand did not solicit
internationalexperience-despite the intent of the
Asian DevelopmentBank's Fiscal Decentralization
TechnicalAssistanceproject 2001-02). Pakistanas-
signedroles to districtsbut ignoredlower-tier nter-
ve governmental elations,and it is still uncertainhow
to proceedwith provincial-localrelations.Both the
ss PhilippinesandIndonesiahave reliedheavilyon in-
ternational ndcomparativeocal legalexperience odraftandmodifytheir egal frameworks.As a result,the institutionalbase of the lattertwo programs s
ffortsstronger han n Pakistan.Third,regimes that wish to sustaintheir decen-
est tralizationreformsconcentrateheavily on capacity
building.ThePhilippinesbeganearly oprovide rain-
ing througha varietyof institutions, ncludinguni-
versities,nongovernmentalrganizations, ndprivatefirms. Indonesia also stresses capacity building and is
stronglysupportedn this by fundsfrom the U.S. Agency
forInternationalDevelopment.However,Indonesia rans-ferredabout wo million staff withoutpriorcapacity-build-
ing support.Pakistanhasrespondedmoreslowly,restrict-
ing training to the existing colonial-based institutions,which areknownmore for theirlegal focus than for man-
but with little prior trainingother than from theirprovin-cial offices. The Asian DevelopmentBank will provide
support or trainingandcapacitybuilding.But it appearsthat much of this supportwill be locally drivenandmaynot be consistent with international estpractices.
The finalkey element of the organizationalrameworkfor decentralization s monitoringand evaluation.This is
the fourthareaof majordifference ntechnicalperformance
and institutions, (4) strong interest in monitoring andevalu-
ation of the devolution program, (5) provision of sufficient
local fiscal autonomy, (6) efforts to streamline local gov-ernment operations, and (7) efforts to replace input bud-
geting and legal management with systems driven by per-formance incentives and
targets. Regimesthat focus on
these areas in sequence are likely to be more successful
than those that do not, illustrated by Pakistan and Indone-
sia at present.More research needs to be done on the determinants of
effective decentralization implementation. This three-stageframework offers a modest start. In the future, to provide
applied policy guidance to regimes contemplating designand implementation, research should focus on differences
in the performance of technical activities by similar coun-
tries and programs. Specifically, research questions should
ask, what factors are associated with variation in perfor-
mance of these activities? Armed with more precise data
and further lessons learned, it should then be possible to
measure the comparative program impact and link this to
differences in performance of technical activities. With this
applied information, regimes should have a better assess-
ment of the potential risks of their decentralization deci-
sions and how to minimize them.
Acknowledgments
Theauthor hanksRachelQuero,directorof DAI-Philippinesin Manila,for her supportandprovisionof materialson the de-centralizationprogram here.An earlierversion of this article
was presentedat the Association for Budgetingand Financial
regional equity (see Prud'homme1994; Guess, Loehr,and
Martinez1997, 1).
2. Differencesin decentralization efer to how well the activi-ties areperformed,hesequence nwhichtheyareperformed,andoverallprogramresults.Here,differencescan be noted
in how well the technicalactivities wereperformed.The se-
quenceof activityperformancewas similar n all threecases.
Dataon overallprogram esults are not yet available.
3. As the decentralizationprogramadvances and statisticsare
reportedmore systematically, he level of local own-source
revenuesshould ncrease.Thenew district ier has thepowerto set tax ratesandbases of such dedicated axes as theurban
immovableproperty ax. Districts also receive a dedicated
annualportionof the sales tax to increase local fiscal au-
tonomy.The NationalReconstructionBureauandMinistryof Finance have largely agreed upon the conceptof a divis-
ible pool of provincialresources(provincial inanceaward)to be sharedwith the districtson the basis of needs andper-formance.To a largeextent,districtscanestablishtheir own
within hebudget-executionystem.Thesmallchange nwhat
was viewed as a hopelessly centralistsystem produced m-
provements n budget management Guess 1993).
6. Onthe otherhand,the "wrong"kind of intermediary rgani-zations may exist. The threeregimes studied are gamblingthat in the medium-term,decentralizationwill reducesup-
portfor
unsupportive, egative ntermediary roupshat still
exist, to someextent, neachcountry.Warlords ndremnants
of al-Qaeda n the Pakistani ribalareas are unlikelyto en-
courage centrally sponsored government decentralization
programsthat could deprive them of indigenous support.Disruptionof establishedcorruptionnetworksby the devo-lution of formalgovernmentalcontrolsystems (such as in-
ternalaudit) n Indonesiamet the sameresistance.Regimes
recognize the threatof such organizations o sound gover-nance and the rule of law andattempt o reducetheirinflu-
ence by political co-option strategies.The appropriatemixofpoliciestoco-optormodifythebehaviorof negativegroups
nationalBankfor ReconstructionandDevelopment.. 2002a.StructuralAdjustmentCreditforGovernment f
the NorthwestFrontierProvince.Washington,DC: Interna-
tionalBank for ReconstructionandDevelopment.
.2002b. Pakistan:DevelopmentPolicy Review: A NewDawn?Washington,DC:InternationalBankforReconstruc-
tion andDevelopment.Xavier,J.A. 1998. Budget Reform in Malaysiaand Australia
Compared.PublicBudgetingand Finance 18(1):99-119.
Yap, MariaEufemia, and VanessaSator.2001. The Status ofDecentralization fHealthServices:ADecadeofBirthPains,Transition and Transformation.Manila, Philippines: U.S.