Spreading Power to the Periphery--An Assessment of Democratic Local Governance

USAID Program and OperationsAssessment Report No. 21

Spreading Powerto the Periphery

An Assessment ofDemocratic Local Governance

by

Harry Blair

Center for Development Information and Evaluation

September 1998

PN�ACA�904

PB Program and Operations Assessment Report No. 21

Spreading Power to the Periphery: An Assessment of Democratic Local Governance iii

Page

Table of Contents

PageAcknowledgments ................................v

Summary ................................................vi

Central Findings .......................................... viConclusions ............................................... viii

1. Introduction .......................................1

Definitions ....................................................1Current Donor Strategy ...............................3Assessment Methodology .............................4

2. Background: PromiseAnd Experience ................................. 6

The Promise .................................................6Decentralization�s Track Record ...................7

Success in decentralization.......................8Major problems in decentralization .........9Subsidiary problems .............................. 10

Reasons for Decentralizing ........................ 11Traditional central government

reasons for decentralizing ............... 11Democratic reasons for

decentralizing ................................. 11

3. Findings ........................................... 13

The Country Context ................................ 13Incentives for decentralizing.................. 14The legal enabling environment ............ 14Basic provisions .................................... 15

Strategic implications ........................... 16Donor and Host-country Roles ................. 16

Previous aid .......................................... 16Strategies adopted ................................. 17Balancing democracy and

public administration ..................... 18Strategic implications ........................... 19

Resistance and Political Will ..................... 20The importance of political will ........... 22Strategic implications ........................... 23

Representation, Empowerment,And Benefits ............................................. 23

Representation ...................................... 23Empowerment ....................................... 25Distribution of benefits ......................... 26Poverty reduction .................................. 27Strategic implications ........................... 27

Fiscal Autonomy and Regional Equity ...... 29Fiscal autonomy .................................... 30Allocations from

central governments ....................... 30Local revenue mobilization ................... 31Regional disparities ............................... 32Strategic implications ........................... 34

Public Accountability ............................... 34Bureaucratic accountability

to elected officials ........................... 34Elected officials� accountability

to the public ................................... 35Patterns of accountability ..................... 40

iv Program and Operations Assessment Report No. 21

Strategic implications ........................... 41Performance and Accountability .............. 42

Strategic implications ........................... 43National Advocacy .................................... 43

Strategic implications ........................... 44

4. Conclusions, Lessons Learned,And Recommendations ................. 45

Conclusions ............................................... 45Lessons Learned ........................................ 45Recommendations .................................... 48

References ........................................... 49

Tables

Table 1. The Two Sides ofDemocratic Local Governance .................. 18

Table 2. Sources of Opposition toDLG From Politicians and Bureaucrats ..... 21

Table 3. Mechanisms for Accountability .. 41

PagePage

Spreading Power to the Periphery: An Assessment of Democratic Local Governance v

CONDUCTED BY THE CENTER FOR DEVELOPMENT

Information and Evaluation (CDIE) of theBureau for Policy and Program Coordination(PPC), this assessment began in the summer of1995, with research for a concept paper. Manypeople contributed generously at USAID/Washing-ton and the Agency Missions in Bamako, Kyiv,Manila, La Paz, and Tegucigalpa, as did contrac-tors involved in implementing decentralizationprojects. Host-country officials, members of non-governmental organizations, and representativesof other donors were uniformly and graciouslyhelpful.

Several individuals deserve particular mention.As chief of CDIE�s Program and Operations As-sessment (POA) division at the time, MichaelCalavan worked diligently to get the study approvedand launched before leaving for the Philippines inJuly 1996. In 1996, Hal Lippman became assess-ment manager for this evaluation and has guidedit throughout, as overall manager, team leader,and principal author for four of the impact evalu-ations. In addition, he contributed much time andeffort in helping edit the final version of this syn-thesis report. Susan Merrill, who succeeded MikeCalavan as POA chief, saw the assessment throughto this report. Patricia Jordan and Joseph Liebersonsupplied logistical backup in managing the CDIEcontract with Development Associates, Inc., con-tractor for the assessment. At Development Asso-

ciates, Joel Jutkowitz and Russell Stout were ma-jor participants in the assessments, leading fieldteams and writing impact evaluations, and DianaDavis provided logistical support for work in Wash-ington and the field.

James Vermillion of PPC, coordinator and im-presario of the �Tuesday Group,� which meetsweekly at USAID/Washington to analyze democ-racy issues, facilitated presentations at variousstages of this assessment, offering valuable feed-back and critical insight. Many people commentedon early drafts; critiques by Todd Amani ofUSAID�s Center for Democracy and Governance,Mike Calavan of USAID/Manila, and MargaretSarles of the Agency�s Bureau for Latin Americaand the Caribbean were especially detailed anduseful. And last, Laurie Denton Conly of Conwal,Inc., served as principal editor for the final stagesof moving the draft manuscript toward publica-tion; in addition, she digested the report for pub-lication in the CDIE Highlights series.

This report is the final component of CDIE�ssix-country study of democratic local governance,but the interpretations and conclusions herein donot necessarily reflect USAID policy. They are theresponsibility of the author, as are any errors orother shortcomings.

AS PART OF THE GROWING PROFILE DEMOCRACY

has assumed in the international donorcommunity in recent years, decentralization has

Acknowledgments

PB Program and Operations Assessment Report No. 21

taken on significant strategic importance. Demo-cratic decentralization, or democratic local gover-nance (DLG) as it is called in this report, is thedevolution of meaningful authority to local bod-ies accountable and accessible to their citizens,who enjoy full human and legal rights to politicalliberty. Donors have used decentralization initia-tives to improve public administration for decades.As a democratization strategy, however, it is rela-tively new and deserves evaluation.

This report is based on studies of democraticlocal governance conducted in 1996�97 in Bo-livia, Honduras, Mali, the Philippines, Ukraine, andthe Indian state of Karnataka. All have introducednew democracy initiatives in the 1990s. For thefirst five studies, CDIE field teams visited the coun-tries for about three weeks each and publishedimpact evaluation reports. CDIE commissioned agroup of social scientists in India to look atKarnataka state as a basis for assessing issuesemerging from the otherstudies.

The report incorporates the experience ofUSAID and other donors, and a review of the lit-erature in the DLG area. Its central objective is toassess development community experience in pro-moting democratic local governance and to de-rive lessons from that experience to inform futuresupport of DLG.

Historically, decentralization initiatives have notenjoyed great success, largely for two reasons: alltoo often, despite their rhetoric, central govern-ments do not truly want to devolve real power tothe local level; and when significant authority isdevolved, a disproportionate share of the benefitsis often captured by local elites. The new demo-cratic variant of decentralization, however, mayovercome these problems by introducing greaterparticipation, accountability, and transparency inlocal governance, and by empowering marginalgroups. It also offers more scope for local revenuegeneration by linking services to local paymentfor them.

Central FindingsThe democratic local governance context.

Prior to initiating DLG, all five countries had cen-tralized, top-down governance structures, but mo-tivations for embarking on decentralization var-ied greatly. They ranged from idealism to the cen-tral government�s failure to deliver adequate ser-vices to a desire of those in power to build up aparty base. DLG initiatives also had various provi-sions. Some made representation for women andminorities mandatory, others set up a system ofelected bodies to monitor municipal governments,and one country had local citizens participate inthe redistricting process that formed new localgovernment units. Which sectors of responsibil-ity devolved to local governments also varied. In

Summary

Spreading Power to the Periphery: An Assessment of Democratic Local Governance vii

some cases, almost everything has been turnedover to local government; in others only someareas�typically health and education�have beenturned over.

Donor and host-country strategies. All fivecountry cases have some USAID assistance to de-mocracy in place. USAID and other donors havecommonly conducted pilot projects, choosingpromising communities to develop local gover-nance approaches that can be replicated on alarger scale. Countries typically launch DLG ini-tiatives nationwide, simultaneously. There is somerisk in �building on the best,� because what worksfor the �best� may be less effective for the rest ofthe country. In addition, donors can exacerbateregional differences by selecting the �best� com-munities for pilot programs. Donors and coun-tries must also decide how much to emphasizethe democracy (input) side of DLG, as opposed tothe public administration (output) side.

Resistance to decentralization. Implicit in anysubstantial DLG initiative is a shift in politicalpower, which inevitably creates apprehension,jealousy, and opposition from both national po-litical leaders and civil servants. Some hostilityreflects self-interest, but much opposition may bewell founded. Political leaders rightly want to ad-vocate national visions, for example, while civilservants have a legitimate interest in maintain-ing national standards. Donor strategists have tosafeguard this legitimate political space and ad-ministrative integrity, while opposing resistancethat undermines DLG initiatives. The challengefor donors is to convince both politicians and bu-reaucrats that democratic local governance is awin�win situation. Much political will at the topis needed.

Representation, empowerment, and benefits.Much of decentralization�s appeal lies in an ex-pected chain of positive change: as marginalgroups�women, ethnic minorities, and thepoor�increase participation in campaigning andvoting, they will achieve better representation on

local decision-making bodies. These groups willthen become empowered�able to influence pub-lic decisions and actions that affect the welfare ofthe group. This in turn will produce more ben-efits for them in terms of service delivery, whichwill ultimately alleviate poverty. The CDIE assess-ment found that participation and representationdid increase significantly for these groups, but thataside from geographically concentrated minori-ties who have won control of local councils, turn-ing representation into empowerment, benefits,and poverty reduction has proven difficult.

But this doesn�t mean these efforts have beenwasted. Members of marginal groups have oftengained valuable political experience and have be-come potential role models for children in theircommunities. Geographically concentrated mi-norities have moved further along the chain be-cause their numbers allow them to win controlof local councils and move them in new direc-tions. More progress has been made in the LatinAmerican countries studied, where local councilsappear more willing to undertake public projectsin fields such as health and sanitation that benefiteveryone, regardless of gender, ethnicity, or wealth.

Fiscal autonomy and regional equity. The ideaof fiscal autonomy is simple: allowing local con-trol over revenue generation lets people decide forthemselves what services they want and howmuch they are willing to pay for them. But mostlocalities have a low tax base and need allocationsfrom the national government for local governanceto succeed. In some systems these grants havebeen generous (up to 40 percent of national rev-enue); in others, considerably less so. Local abil-ity to generate revenue also varies. Advancedmunicipalities generate more revenue, providemore services, and become even more advancedthan poor communities, which have trouble rais-ing even minimal revenues. Central governmentscan compensate for such inequities with specialallocations, but excess subsidies can weaken local

viii Program and Operations Assessment Report No. 21

incentives to tax and can be politically difficult tosustain nationally.

Public accountability. Two types of account-ability are essential for democracies: governmentemployees must be accountable to elected repre-sentatives, and representatives must be account-able to the voting public. Bureaucratic account-ability suffers because central governments givein to civil servants� reluctance to be placed underlocal control. Elected officials are accountable atthe time of elections (if elections are free and fair),but elections are blunt instruments of popularcontrol and typically occur only at widespreadintervals.

However, there are a number of ways citizenscan hold local government officials accountablebetween elections. They include political parties(especially opposition parties), civil society, themedia, public meetings, formal grievance proce-dures, and opinion surveys. Each mechanism (ex-cept the surveys) has been instituted in at leasttwo of the six countries studied, but none hasbeen tried in all of them. No two systems havetried the same combination of mechanisms, andno single mechanism has been effective every-where. Some systems appear to be doing well withseveral mechanisms; others have fared less well.The implications are that the package of mecha-nisms should be crafted to suit the country.

Performance. Most of this assessment looks atthe democracy (input) side of DLG, but what lo-cal government achieves (output) is equally im-portant. For governance cannot just be something;it must also do something�it must deliver ser-vices citizens find useful. The initiatives studied inthis assessment are too recent for a good readingon performance, but service delivery seems to haveimproved in a couple of countries and increasedpublic accountability seems to explain much of

this improvement.

National advocacy. Like other new undertak-ings, DLG requires leadership and financial sup-port at the national level. International donorsand national political leaders can provide supportinitially, but this backing will wane over time. Inits place, local governments will need championsfor their cause in the capital. Associations ofmayors or municipalities have become effectiveadvocates for local interests in several of the coun-tries studied.

ConclusionsThe potential gains from a decentralization ini-

tiative that is well designed and implemented morethan justify the effort. But local governance ini-tiatives cannot do everything. From this assess-ment of USAID programs, host-country govern-ment initiatives, and to the extent possible, otherdonor experience, it is possible to draw a numberof conclusions about the strengths and limitationsof DLG.

■ Participation and empowerment. DLG can sig-nificantly increase political participation ofmarginal groups and can empower geographi-cally concentrated minorities; it appears un-able to empower marginal groups that are notgeographically concentrated, at least in theshort run.

■ Poverty reduction. DLG can help alleviate pov-erty by strengthening the capacity of localgovernment to deliver services that benefit thewhole population, in sectors such as educa-tion, health, and water supply; it shows lesspromise in reducing poverty through effortsdirected specifically at marginal groups.

■ Fiscal sustainability. DLG can partly sustain it-self through local revenue generation, but this

Spreading Power to the Periphery: An Assessment of Democratic Local Governance ix

will tend to exacerbate regional imbalancesunless supplemented by central subsidies topoorer areas.

■ Accountability. When a variety of mechanisms,such as civil society, media, and political par-ties are used together, DLG can improve ac-countability of local government bodies to thecitizenry; used in isolation, these instrumentsappear much less effective.

Spreading Power to the Periphery: An Assessment of Democratic Local Governance 1

Introduction

globe graphic

1

A S DEMOCRACY ASSUMED MORE IMPORTANCE

in the international donor community inthe 1990s, it made sense that decentralizationwould too, since donors cannot support nationaldemocratization for very long without recogniz-ing a similar need at the local level. So demo-cratic decentralization, or democratic local gover-nance (DLG), as it is called here, has been an aimin strategies pursued by USAID and other donors.Based on studies of democratic local governancein six countries, this assessment of donor sup-port is one of a series of assessments of democ-racy undertaken by the Agency�s Center for De-velopment Information and Evaluation (CDIE) inrecent years.1 The report incorporates the expe-rience of USAID and other international donorsand a review of the relevant literature.

International development groups began pro-moting decentralization after World War II, butuntil the late 1980s these efforts primarily em-phasized public administration. The goal was tomake government programs more effective bybringing them closer to the people. Only late inthe 1980s did democracy become a central com-

ponent of donor-supported decentralization.2 Evenso, decentralization has been promoted longenough in its democratic incarnation to justifyan assessment of its track record and suggest broadstrategies for supporting DLG. This assessmentlooks in particular at local participation and pub-lic accountability. It also discusses fiscal autonomyand performance. The central objective is to as-sess recent donor community experience in pro-moting DLG and draw lessons that can informfuture efforts to support DLG. The goal is to ad-vance, analytically and substantively, our under-standing of democratic local governance.

DefinitionsDecentralization is arguably as old as gover-

nance itself. Ancient empires, medieval kingdoms,and modern states have all had to decide howmuch tax authority or discretionary power to keepcentralized and how much to apportion to localofficials. It is not surprising, therefore, that do-nors and host countries have often engaged indecentralization initiatives in the five decades ofpostwar international aid.3 In recent years, de-

1 Earlier assessments include analyses of judicial reform (Blair and Hansen 1994), civil society (Hansen 1996), andlegislatures (Lippman and Emmert 1998).2 For an excellent brief historical analysis of decentralization as a development theme, see Cohen and Peterson (1996,esp. 1�5).

2 Program and Operations Assessment Report No. 21

centralization has assumed a new incarnation asa central part of the democratization strategiesUSAID and other donors are pursuing. The �demo-cratic� in democratic decentralization can bethought of as value added to the older traditionof decentralization in development.4

At least since the early 1980s, analysts havedistinguished three types of administrative decen-tralization:5 deconcentration, delegation, and devo-lution.

Deconcentration: assigning specific functionsor duties to locations outside the home office, suchas a line ministry�s field offices. Generally, in thismost common form of decentralization, the cen-tral agency continues to exercise as much initia-tive and authority as possible.

Delegation: a more serious transfer of respon-sibility to local officials. Broad policies are deter-mined centrally, but local agents have consider-able autonomy in implementing them and carrymuch of the responsibility for results.

Devolution: usually a statutory transfer of au-thority and responsibility for specific sectors tolocal bodies. Local school boards in the UnitedStates, county councils in the United Kingdom,and the new local government scheme in thePhilippines are examples of devolution, the most

ambitious form of decentralization. At a higherlevel, so is the federalism of North America andIndia.

Decentralization does not necessarily involvedemocracy, given that it was a development strat-egy long before democracy appeared on the inter-national development agenda in the mid- to late1980s.6 Nor have decentralization efforts histori-cally assumed that �local is necessarily better,�although that assumption is explicit or implicit inmuch of the literature. And while devolution im-plies a degree of democracy, it does not requirethat local bodies be democratically constituted.

For our purposes, democracy involves

■ Popular sovereignty, with subnational units ofgovernment accountable and accessible to citi-zens, both periodically, through elections, andcontinually, through the rights to advocate andlobby

■ Political equality, with citizens enjoying fullhuman rights, in the form of protection againststate abuse, and full legal rights with respect toeach other

■ Political liberty, or freedom of speech and as-sembly7

This definition applies to the national and lo-

3 A search of USAID data revealed 331 projects incorporating some degree of decentralization since the late 1970s with a�life of project� cost of $9.35 billion. The database goes back only to the late 1970s. Earlier projects included the CommunityDevelopment Program in India in the 1950s (see Bowles 1954, Blair 1982).4 Democratic local governance could also be seen as bringing ideas of decentralization to democracy.5 Dennis Rondinelli has greatly influenced the sorting out of terms discussed in the text (Rondinelli 1981, 1984). Forcomprehensive analyses of decentralization, see Conyers (1983, 1984, 1986) and Smith (1985).6 Although the development community may not have promoted democracy significantly in earlier decades, it was concernedwith the need for local participation and popular input. The 1961 United Nations Handbook of Public Administrationmentions �the gradual development of local self-government� and �citizen participation,� even mentioning �legislation toestablish elected councils� (UN 1961: 67�68). But the emphasis then, as in 1970s, was on public administration. In USAID,local participation was part of Title IX of the Foreign Assistance Act in 1966 and part of the �New Directions� mandate ofthe early 1970s (Esman and Uphoff 1984), but democracy remained subordinate to administration. A fine line divides�participation� from �democracy,� but the Title IX and New Directions initiatives generally emphasized inclusion in develop-ment activities, particularly the project cycle. Today there is more emphasis on decision-making in democratically electedlocal governments.


cal levels, but the responsibility for maintainingdemocracy is national.

Democratic local governance combines devo-lution with democracy, especially the popular sov-ereignty component of the definition. Democraticlocal governance is a system in which meaning-ful authority is devolved to local bodies that areaccountable and accessible to local citizens, whoin turn enjoy full human and legal rights and po-litical liberty.

CurrentDonor Strategy

USAID policy strongly emphasizes democraticdecentralization, as this 1994 statement explains:

Local involvement is important in any kind offoreign assistance, but it is essential in democ-racy building. Local forces must provide the

principal impetus for creating, nurturing, andsustaining an environment in which democracycan thrive. USAID’s role is to stimulate and re-

inforce democratic elements at the city andcommunity level. [USAID 1994,18]

Democratic decentralization is one of five pro-gram approaches to �more transparent and ac-countable government institutions,� and one offour strategic objectives under the Agency goal ofbuilding democracy.8 It is also one of three tar-gets of the Clinton administration�s New Partner-ships Initiative, along with nongovernmental or-

ganizations (NGOs) and small businesses. Thatinitiative stresses reforming the enabling environ-ment for democratic decentralization and build-ing local capacity�both in local government andlocal civil society (USAID 1997a).

The Agency has expanded its decentralizationemphasis from public administration to include asubstantial democracy component as it has movedfrom a more rural to a somewhat more urbanview, at least in some regions.9 The shift may re-flect both increasing urbanization10 and the endof the Cold War, which has lessened concern aboutrural insurrections in developing countries.

As for other donors, the World Bank remainsmore oriented toward public administration thanUSAID in its decentralization strategy (World Bank1994, 73�79), but United Nations DevelopmentProgram (1997a, 1997b), United Nations Capi-tal Development Fund (1995), and Norway (1994)have moved toward supporting democracy. In arecent report, the Development Assistance Com-mittee (DAC) of the Organization for EconomicCooperation and Development (OECD) commendsdemocratic decentralization as a component ofstrategies for supporting democratization.11 Eventhe World Bank, whose charter proscribes involve-ment in political activities, has shown an increas-ing appreciation of the participatory aspects ofdecentralization.12

AssessmentMethodology

7 This definition of democracy is taken (with some liberties) from Greenberg and Page (1993: 24�28 ff).8 The other three program approaches are greater transparency (reducing corruption), increased civilian control of policeand military, improved legislatures, and conflict mitigation. At a higher level, governance is one of four Agency strategicobjectives, along with rule of law and human rights, elections, and civil society. See USAID (1997b, passim).9 This is true for Asia. In Latin America and the Caribbean, USAID support for decentralization has historically involved asignificant urban dimension.10 By the year 2000, the developing world�s population is expected to be about two fifths urban; by 2015, about half urban(UN 1989, 5).11 See DAC (1993), and OECD (1997 parts I, 11, and 1997, part II, 23�26).


This analysis, like earlier CDIE assessments,began with a concept paper that led to a series ofcountry field studies on DLG. The concept paper(Blair 1995) broadly outlined the course of theevaluation.

The countries chosen for field studies repre-sent a range of regions and experiences with de-centralization. So few decentralization initiativeshave succeeded, however, that we made a con-scious effort to find countries where DLG holdspromise. Negative lessons can be instructive, butwe could only conduct a few field studies, so wesought cases where USAID-supported efforts hadshown signs of working. Our first criterion wasthat a democratic political system be in place na-tionally, because the national context is criticalfor the success of DLG.13 Our second criterion wasthat the national government had launched a se-rious DLG initiative with a good chance of suc-cess. Our final requisite was the existence of sig-nificant USAID support for the host-country ini-tiative.14 All selections were made in close con-sultation with democracy specialists at Washing-ton regional bureaus and with the relevant USAIDMissions.

The clear choice in USAID�s Asia and Near Eastregion was the Philippines, where USAID sup-ported efforts to craft the Local Government Codebefore its inception in 1991 and to implement it.It was the site of our first CDIE field visit in April�

May 1996. In Europe and the new independentstates, Ukraine, site of our second field study inJuly�August 1996,15 has had a local governmentlaw in place since 1992. The Latin America andCaribbean region provided two countries for study:Bolivia, which was implementing the Popular Par-ticipation Law passed in 1992 (visited in Septem-ber�October 1996); and Honduras, with its Mu-nicipal Reform Law in place since 1990 (visitedin October�November 1996). In sub-Saharan Af-rica, no system met all our criteria, so we selectedMali, where a slow, painstaking effort is underway to fulfill the commitment to democratic gov-ernance so prominent in that country�s 1991 revo-lution and 1992 constitution (visited inJanuary 1997).

These proved good choices but none had a DLGprogram dating from before the 1990s. To includeone enterprise of longer standing, we undertooka study of Karnataka state in India, recognizingthat USAID had not had a role in DLG there. Butinstead of sending a CDIE team to Karnataka, wecommissioned local social scientists to analyzeissues there. Karnataka is regarded as having oneof India�s more successful Panchayati Raj sys-tems.16 The Panchayati Raj program, which es-tablished locally elected self-governing councils,has been in effect in this state since the beginningof the 1960s, albeit under various guises and withsome interruptions. In India�s federal system,Panchayati Raj is subject to state control. India�s

12 See Parker (1997), Manor (1995), and the most recent World Development Report, which devotes considerable space todecentralization (World Bank 1997, 120�130).13 It is possible to initiate decentralization without democracy at the top, and it has occasionally been tried, (as in Pakistan�swell-known Basic Democracies program under Ayub Khan in the late 1950s ), and hinted about (in such countries as Chinaand Indonesia). In no cases known to the author, however, have such efforts been successful over time. Countries selected forthis sample had become largely democratic at the national level by the time of the CDIE assessment. All six countriesenjoyed a 1996 ranking of 7 or better on the 14-point Freedom House scale (2 is the best score) (Freedom House 1997).14 Ideally all of the cases would have had enough experience that a visiting CDIE team could get a clear picture of what hadworked and what had not. This was not possible, so we had to select some countries that had begun their initiatives not longbefore our visits.15 It would have been worthwhile to include an Eastern European country in the sample as well, but none met all threecriteria.


states plan and implement their local democraticgovernance programs fairly autonomously so theyare comparable to nation�states elsewhere.17

With just six cases, of which only five wereCDIE field studies, our sample is not scientificallyvalid, but constraints on time and funding lim-ited us. Moreover our sample included only coun-tries with some established base of democracy. Itdid not include, for example, transition countriesundertaking the passage to democracy, such asmany of the countries in sub-Saharan Africa. Nordid the sample include countries in the Near East.But the sample does broadly illustrate experiencewith DLG in a variety of settings, offering a more-t h a n -adequate canvas for assessment.

For each field study, we assembled a CDIE teamof three or four people, including at least one di-rect-hire officer from CDIE, one or two consult-ants with experience in democratic governance,and, in most cases, a direct-hire officer from theCenter for Democracy and Governance or the ap-propriate regional bureau in Washington. We alsohired one or two local experts for backgroundconsultation, translation, and logistical support.The teams spent about three weeks in each coun-try, both in the capital and visiting local govern-

ments. Data came mainly from interviews withkey personnel in central and local government andnongovernmental organizations, as well as USAIDMission personnel, contractors, and other donors.Documents from all these sources were also ex-amined. Focus groups proved useful in severalcases. In two countries small-scale opinion sur-veys were conducted.

Reports for the five field study cases are beingpublished in CDIE�s Impact Evaluation series, non-technical analyses for the general developmentaudience.18 This synthesis is based on the six coun-try analyses, but draws on some material not in-cluded in the impact evaluations and some dataacquired after the field studies were completed.

Background:Promise andExperience

THE TRACK RECORD OF EFFORTS TO PROMOTE DLGhas been mixed, yet decentralization still at-

tracts support among donors and host-countrygovernments.

16 Democratic local governance was the subject of an excellent study in the early 1990s (Crook and Manor 1994). AnandInbanathan, principal investigator on the Karnataka study, worked not only on the Crook and Manor study but also on anearlier one on democratic local governance there (Inbanathan 1992).17 With a population in 1991 of 48 million, Karnataka was a bit smaller than the Philippines, about the same size asUkraine, and considerably larger than the other three countries.18 The reports are Lippman and Jutkowitz (1997); Lippman and Blair (1997); Blair (1997); Lippman and Pranke (1998); andLippman and Lewis (1998). The Karnataka case (Inbanathan and others 1997) was not designed as an Impact Evaluationstudy, so it is not being published in the series.


The PromiseThe promise of DLG, in the political, economic,

and social realms, accounts for much of its at-traction to donors and host country governments.It would be hard to imagine another initiative of-fering so much to so many constituencies.19

Participation and accountability. The signalpromise of decentralizing government authorityis enhancing democratic participation by encour-aging more people to get involved in the politicsthat affect them, and making government moreaccountable by introducing citizen oversight andcontrol through elections. If democracy lies in ruleby the people, the promise of democratic decen-tralization is to make that rule more direct, im-mediate, and productive.

Transparency in government. An allied virtueis transparency in government. The idea is that ifit is easier to monitor what goes on in local gov-ernment, officials can be held to account for cor-rupt practices and malfeasance and will find it dif-ficult to hide misdeeds and improprieties. This couldbe a mixed blessing if, in learning more about theirgovernment, citizens become alienated from andcynical about it.

Empowerment. Many believe that by makingparticipation easier, democratic decentralizationmakes empowerment more feasible locally thanit would be nationally, especially for marginalgroups. With empowerment, groups have a sig-nificant voice in public policy decisions that af-fect their futures. The idea is that it is simpler foreveryone to get involved locally, where politics is

2

more understandable and access to governmentis easier. Local empowerment is more possible inthat minorities are more likely to enjoy a criticalmass in small areas than in large ones. In addi-tion, organizational problems are fewer and costslower. Accordingly, many believe, women, thepoor, and ethnic or religious groups that are de-nied participation nationally can more easily be-come politically influential at the local level.

Responsiveness. Just as it is hoped that thoseon the periphery can gain political power throughdemocratic decentralization, so it is believed thata democratically decentralized government will bemore responsive to local demands. A governmentthat sees itself as genuinely accountable to its citi-zens is more likely to pay attention to their wantsand needs.

Political education. As they participate in localpolitics, citizens come to understand the demo-cratic process that allows them to manage theirown affairs. This education will help them under-stand and participate in national politics.

Leadership recruitment. Democratic decentrali-zation provides education through experience,training leaders for higher levels of government.It creates a new recruitment channel for national

19 For a list of benefits, see Manor (1995, esp. 32�33). Wunsch and Olowu (1996�97) provide a good analysis for Africa.


leaders, and an alternative to more familiar pathsthrough national political parties or military or-ganizations.

A political free market. The most carefully ar-ticulated rationale for democratic decentralizationis that it offers the closest possible political ap-proximation of a free market, which many be-lieve is the ideal structure for all social relation-ships. In this �public choice� view, democratic de-centralization is seen as the best way to bring �buy-ers� (citizens) and �sellers� (local governments)together, by providing an arena in which buyerscan signal what they want and sellers can signalwhat they are willing to offer. Buyers can alsodiscipline unsatisfactory sellers by monitoring andpublicizing their activities and by voting to replacethem. Sellers who displease too many buyers goout of business by losing the next election; thosewho perform satisfactorily are rewarded by beingreturned to office.20

Joining services and payment. Democratic de-centralization helps resolve one of the most de-bilitating fiscal problems of democracy: that po-litical demand for a service can be separated bymeans of subsidies from the need to pay for it.Urban citizens, for example, can demand cheaphospitals subsidized by taxing other sectors, suchas agriculture, or by government deficits (whichall citizens fund through higher taxes). Democraticdecentralization forces a convergence of politicaldemand for a service (putting political pressureon the system to deliver it) and effective demand(the ability to pay for it by mobilizing local rev-enue).

Less ethnic conflict. Decentralization is some-times seen as a way to reduce ethnic conflict by

allowing geographically concentrated ethnicgroups to dominate in their own regions in re-turn for accepting less (or no) power nationally.Switzerland is an obvious example of a countrywhere such a strategy has succeeded, but it hasalso been tried on North American Indian reser-vations, in autonomous regions in the Philippines,in the Taureg areas of Mali and Niger, and in someregions of China, among other places.

Decentralization�sTrack Record

Decades of experience with decentralization asa development initiative have produced many stud-ies and evaluations. Understandably, most of theearlier assessments studied decentralization largelyfrom the perspective of public administration orpublic finance�theareas on which donor involvement was focusedat the time. By contrast, a number of recent evalu-ations study it from the perspective of democracyand participation.21

The findings of the studies completed so far, in-cluding those mentioned here, provide some evidenceof success at both administrative and democraticdecentralization. Most of the evidence, however, hasbeen rather less encouraging.

Success in decentralizationDecentralization efforts have enjoyed some suc-

cess largely in two areas, participation and re-sponsiveness. In at least a few cases, they havealso led to an increased ability to generate rev-enue locally.

20 For an excellent synopsis of the public choice approach to decentralization, see Smith (1985, 32�35), from which much ofthe discussion here is drawn. Public choice theory, together with public finance theory and �new institutional economics,�contributed to the �institutional analysis� approach to public policy that guided the USAID Decentralization, Finance, andManagement project (1988�94), which has generated much experience and literature. See ARD (1994), especially the bibliog-raphy in annex C.


Increasing participation. When democraticgovernance moves closer to the people, citizenscan and do become more interested; more vote inelections22 and take a role in such civil activitiesas joining associations and gathering for demon-strations. Crook and Manor (1994) found suchchanges in Bangladesh, Ghana, and India; so didOrtuste and Custode (1997) in Bolivia, andBrillantes (1994), among others, in the Philip-pines. Participation has not increased in all coun-tries pursuing decentralization initiatives, but ithas in some.

Improved responsiveness. Bringing governmentcloser to the people can make it more responsiveto popular needs and desires, report Crook andManor (1994) about Bangladesh, Côte d�Ivoire,and India and Manor (1995, 82) about Nigeriaand the Philippines.

Generating local resources. Although the gen-eral track record is not one of great success, ef-forts to raise local revenue have at times workedout well. Rondinelli and others (1989) cite theHarambee era in Kenya (see also Smoke 1994)

and evidence from Nepal (see also Schroeder andWozny 1987). Manor (1995, 84) discusses newlocal taxes successfully imposed in Uganda. Onestudy in the 1980s (Blair and others 1989) foundsome local governments in Bangladesh did rea-sonably well collecting fees, mainly in connectionwith periodic markets.

Recruiting leadership. Democratic local govern-ments have indeed served as a training groundfor citizens and leaders, providing an educationin how democracy works. India, with elected gov-ernments from the national down to the villagelevel, offers a history of village panchayat mem-bers later running for district and state office. Forexample, more than 55,000 people served as lo-cal council members in Karnataka state in thelate 1980s (Crook and Manor 1994). If only 10percent of them have gone on to other electedpositions that means more than 5,000 peoplereceived on-the-job training for future work inelectoral politics.

Major problems in decentralizationThe many items on the negative side of the

21 Several analyses offer a good picture of experience with democratic decentralization in the late 1980s and early 1990s.The British Overseas Development Administration sponsored field studies on performance in two African (Côte d�Ivoire andGhana) and two Asian (Bangladesh and Karnataka state in India) systems (Crook and Manor 1994, Manor 1995). The FordFoundation supported studies on accountability in five anglophone African countries�Kenya, Nigeria, Tanzania, Uganda, andZimbabwe (Barkan 1996).

The World Bank assessed rural development, with field studies of social capital�the trust between people that facilitatestheir working together toward a common purpose�in Burkina Faso, Indonesia, and Tanzania, desk studies of 20 othercountries (see Binswanger and Shah 1994, Parker 1995, and Shah 1997), and at least one general paper on the democraticaspects of decentralization (Manor 1997, draft). (Publications are forthcoming from a conference on decentralizationsponsored by the World Bank and Food and Agricultural Organization of the United Nations in Rome, December 1997.)

The United Nations Development Program sponsored a study of administrative decentralization in Ethiopia, Kenya, andMexico (Cohen and Peterson 1995). It has also launched a series to analyze democratic decentralization and sustainablehuman development. The series is outlined in UNDP (1997a and 1997b).

The Expert Group on Aid Evaluation of the Development Advisory Committee of the Organization for Economic Coopera-tion and Development included a general assessment of decentralization in a series of evaluations on participatory develop-ment and good governance. The DAC sponsored a desk study of decentralization (Aasen and others 1997), held a three-dayworkshop on the topic in Paris, (fall 1996), and made recommendations in its final report on participatory development andgood governance (OECD 1997).

Michael Calavan of USAID recently completed a regional study of democratic decentralization at the microlevel, examiningvillages and subdistricts in Bangladesh, Nepal, and Thailand. (Calavan [1996] undertook the project while on a study leavefrom USAID, so it is not an official Agency report.) For two Latin American regional studies, see Bidus (1995) on CentralAmerica and Nickson (1995) on the structure of decentralization in Latin America.22 The contrast here with the United States is intriguing. In our own state and local elections, participation customarily dropsoff to a fraction of what occurs in national elections.


ledger amount to variations on two principalthemes:

■ Despite rhetoric to the contrary, all too often cen-tral governments do not really decentralize sig-nificant power and tend to deconcentrate ratherthan devolve authority

■ To the extent that central governments do ac-tually decentralize, the benefits are largely cap-tured by local elites, leaving other groups nobetter off than before

Reluctance to decentralize. In most decentrali-zation schemes little gets decentralized. Sometimesnational political leaders, such as Jomo Kenyattain Kenya, Julius Nyerere in Tanzania, and succes-sive generals in Bangladesh, are unwilling to letgo of personal power. Sometimes devolutionaryschemes that seem to be working get reversedbecause national leaders become jealous of theirlocal counterparts, as in the Indian states ofGujarat and Maharashtra, where state politiciansterminated the Panchayati Raj program (Webster1990, cited in Binswanger and Shah 1994, 59).Sometimes central bureaucracies undercut devo-lution, fearful of losing their field organizations tolocal control or becoming politically accountableto local officials. Not surprisingly, decentralizationinitiatives sometimes turn out to be centraliza-tion efforts in disguise, a cloak for more controlfrom the center (Ng�ethe 1994, Slater 1989,Walker 1991).

Local elite takeover. This is the other most com-mon reason for the failure of decentralizationschemes, so much so that one leading Americanpolitical scientist has simply dismissed decentrali-zation as �a carte blanche for vested interests.�(Lowi 1978, 259; see also Aasen and others 1997,15�16). In his 1968 analysis of South Asian de-velopment, Gunnar Myrdal cited elite dominance

as the principal factor behind what he saw as theutter failure of India�s Panchayati Raj experimentin democratic decentralization.23

There seem to be two patterns to the elite take-over of local government. In one, some combina-tion of policy innocence, inattention, or indiffer-ence at the national level opens the door for localelites to take over governance structures and per-vert them to their own uses. National leaders mayintend to empower the weak through decentrali-zation, but the traditional dynamics of the localpolitical economy too often skew things to thebenefit of the locally powerful. That was the cen-tral message in Myrdal�s analysis.

In the other pattern, elite dominance is not somuch the problem as the solution to another abid-ing national problem: how to secure the allegianceof the local power structure. In a tacit collusionbetween national and local elites, national lead-ers allocate development funds to the local levelknowing that local elites will siphon off the fundsfor themselves. In return, local elites keep thingsunder control and support the national leaders.

In this fashion, decentralization has offeredsome national leaders a chance to build a patron-age-based community network, giving them asupport base independent of the national partyand bureaucracy they must otherwise depend on(Blair 1982 and 1985, Westergaard and Alam1995). One can argue that if decentralization isdemocratic enough, eventually weaker segmentsof the community will learn to become seriousplayers at local politics. But in the short term, theability of democratic decentralization to materi-ally improve the lot of marginal groups appearsmodest at best, judging by examples from Africaand Asia (Manor 1995, 84).

Indeed, a strong case can be made that time, a

23 Myrdal (1968: 887�91, 1339�46). Many others have arrived at similar conclusions about Panchayati Raj (see Bendix 1969,338�56; Nicholson 1973).


strong central authority, and national determina-tion represent the road to equity and empower-ment of the weak. The history of civil rights inthe United States and rights for the ScheduledCastes (�Untouchables�) and Scheduled Tribes inIndia are examples of a central government chang-ing the rules so that marginal groups could enterthe political arena over time.

Subsidiary problemsSeveral other factors are problems as well.

Poor planning. Most local governments appearincapable of serious planning, beyond assemblingproject wish lists. Assigning priorities or buildinginformation bases for decision-making (project-ing future school enrollments, for example) seembeyond them (Manor 1995, Aasen and others1997). Worse perhaps, most local governmentsseem much more interested in building new fa-cilities (such as roads, buildings, and culverts) thanin maintaining them, so physical infrastructureexpands and decays simultaneously (Slater 1994,Gephart and others 1989).24 And while this isarguably no worse than bad planning by centralgovernments, it offers no reason to devolve powerto the local level.

Poor local revenue generation. Although somelocal governments can raise revenue, local effortsare generally disappointing. By contrast, there islikely to be more payoff in lobbying the centralgovernment for more resources than in trying toraise revenues locally (Garnier and others 1991).There is also the age-old problem that raising lo-cal taxes often means hiking them on local elites,who generally are adept at evading them.

Competition with other officials. Traditionallypoor relations between locally elected officials andbureaucrats are often compounded by the jeal-

ousy higher-level elected officeholders have forlocal authorities (Manor 1995, 87ff). National par-liamentarians, eager to be perceived as the pri-mary link between citizens and their government,are unlikely to look favorably on locally electedofficials who can distribute largesse to the sameconstituents.

Donor bias toward centralization. Donor pro-grams by their nature tend toward centralization,thus undercutting decentralization efforts. Condi-tionalities for donor assistance to decentralizationinitiatives must be met nationally, for example(Binswanger and Shah 1994, 33), and foreign as-sistance in the last several decades has generallystrengthened national over local governments(Aasen and others 1997). Moreover, unrealisticdonor schedules for project completion favor cen-tral control, even when the objective is to promotedecentralization.

Reasons forDecentralizing

Given decentralization�s unpromising trackrecord, why have governments tried it so often?25

What factors have motivated states to undertakedecentralization initiatives? Usually their reasonsfor doing so have been largely unrelated to de-mocracy.

Traditional central governmentreasons for decentralizing

Adapting policy. Even in as homogeneous acountry as Bangladesh or El Salvador�where thepopulation is strikingly uniform ethnically, linguis-tically, and religiously and where local geographydoes not vary greatly�a central government can-not adequately plan for important local variations.Local variations call for flexibility. To facilitate ef-

24 One explanation is that possibilities for graft are significantly higher in construction than in maintenance. Also, politiciansprefer giving the voters something new (rather than fixing something old) and engineers prefer the glamour of building (ratherthan repairing).


fective local planning and implementation, thecentral government must allow some decentrali-zation.

Supporting regimes. Many national leadershave used decentralization to build up a local sup-port base separate from national political partiesand bureaucracies. Generally this has meant buy-ing the support of local elites through local grants.In the short term such pseudo-decentralization efforts can strengthen politicalsupport for national leaders but they contributelittle to development or democracy.

Avoiding fragmentation. Central governmentshave also tried to avert regional disaffection andsecession by decentralizing power geographically.Sometimes this has happened in a federal system(as when tribal hill states were created in north-eastern India),26 and other times in a more uni-tary state (as when governments gave grants ofspecial autonomy to indigenous hill peoples andMuslims in the Philippines, and to Tauregs in Maliand Niger).

Transferring responsibility. Whether for ideo-logical reasons (the belief that local control is goodin itself) or fiscal ones, central governments cantransfer responsibility for service delivery throughdecentralization. In a number of African andformer Soviet countries, central governments havesimply become unable to exercise their traditionalfinancial and administrative responsibilities andhave passed them on to local governments by de-fault. Health care and education are prominentexamples of services that, if they are to be pro-vided at all, have to be provided locally in manycountries.

Democratic reasonsfor decentralizing

Each of the reasons cited above also has a demo-cratic counterpart putting it in a more positivelight that accords with much current donor think-ing. (USAID 1994, UNDP 1997a and 1997b, Nor-way 1994). Among the reasons donors supportdemocratic decentralization to strengthen local de-mocracy are

■ Local citizens should have control over howstate policy affecting their lives is formulatedand implemented

■ Power must be accompanied by accountabil-ity; governance structures should be account-able locally as well as at higher levels

■ Local governance can more satisfactorily ad-dress the particular needs of different culturalgroups and geographic conditions in heteroge-neous countries

■ Local public services can be more effectivelyprovided when they are locally managed, de-livered, and paid for

In other words, decentralization initiatives un-dertaken earlier for reasons of state have now beenundertaken for democracy�s sake as well. The dif-ference, of course, lies in citizen participation,public accountability, and genuine local autonomy.

Findings

DEMOCRATIC LOCAL GOVERNANCE SEEMS TO WORK

25 Others, such as Binswanger and Shah (1994, 3) ask the opposite question: if governments have such high expectations ofdecentralization, why have they tried it so rarely?26 The U.S. federal system is, itself, a compromise between national and state power.


in many different environments, which ap-pears to be one of its real strengths as a develop-ment strategy.

The Country ContextSome initial conditions were common in all

our case studies. Politically, five of the six had acentralized, top-down state structure, with somevariations. In Ukraine a rigid, authoritarian cen-tral government controlled all local political ac-tivity as well as much of the social activity. InMali, state control was limited to larger urban

areas, essentially leaving the countryside to its owndevices. Central and local governments in Bolivia,Honduras, and the Philippines exhibited variationson a patron�client relationship between center andperiphery. Only in India, where the PanchayatiRaj system had begun around 1960, was there alegacy of local democracy before the democraticdecentralization initiatives of 1983 and 1993.


3Not surprisingly, the economies of the case

study countries largely paralleled their politicalstructures. In Ukraine�s command economy, cen-tral planners tried with decreasing success to man-

age outputs and consumption patterns. The otherstates in varying degree had dual economies�asmall, relatively advanced formal sector coexist-ing with a much larger, subsistence-oriented in-formal sector. The informal sector provided someexports, benefiting the formal sector, but other-wise merely sustained itself on the margins ofthe economy. Such an arrangement tended to re-inforce the domination of local elites, who largelycontrolled the production of export commoditiesand owned the most productive local assets.27

Local governance tended to reflect nationalpolitical and economic patterns. In Ukraine, ad-ministrative deconcentration was the norm. Thecenter set up field offices to manage its vast socialwelfare programs but retained central control. TheMarcos regime in Manila ruled largely throughlocal bosses, who governed their domains throughthe pulitika system of patron�client relations. Thepattern was similar with Mali�s hereditary villagechiefs, though their power was less than in theother cases. In Honduras a somewhat more be-nign succession of presidents managed the mu-nicipalities through what amounted to appointedmayors (described by several observers to the CDIEteam as �decorative�). In most Karnataka villages,dominant castes used their combined ritual sta-tus and economic clout to wield authority overthe rest of the population.

27 In Karnataka, �exports� went from the rural sector to the urban sector, but otherwise the picture was similar. Much of Africamay be an exception to this pattern because local African elites have not had such control over commodity production.

Bolivia was somewhat different, because itsIberian culture exercised central hegemony, butindigenous organizations such as peasantsindicatos, or unions, performed many local gov-ernance functions. About half the population livein cities that enjoyed a measure of self-rule afterdemocracy was reintroduced nationally in themid-1980s, but the rest depended on what littlethe center dispensed from time to time.

Incentives for decentralizingThe motives for undertaking DLG were a mix

of the practical, the opportunistic, and the high-minded, as is usually the case with importantpolitical initiatives. In some ways, the histories ofthe six systems were similar, but what led theirnational leaders to embark on decentralizationvaried.

In Mali and Ukraine, the main incentive was aserious failure of the central government to de-liver the government services promised in an ear-lier era. After the collapse of the Soviet Union,the bankrupt Ukrainian state simply defaulted onits obligations to provide education, health care,water, and other services. In Mali the centralgovernment�s ability to provide social services fellso far short of demand after the mid-1970s thatit largely stopped supplying them outside majorcities.


In addition, countries undertook DLG for ide-alistic reasons. Ukraine�s 1996 constitution de-scribes the importance of local self-governanceas an end in itself. Mali�s DLG initiative came outof the National Conference, which took place soonafter the 1991 revolution that overthrew the pre-vious military dictatorship. In the Philippines, localself-governance came into being at the urging ofa small group of idealistic politicians headed bySenator Aquilino Pimentel. Leaders in Bolivia,Honduras, and Karnataka must also be creditedwith a genuine conviction that decentralizeddemocratic governance is a worthwhile end in it-self.

Finally, leaders had opportunistic reasons forundertaking democratic decentralization.Karnataka�s chief minister, Ramakrishna Hegde,wanted to build up a local base for his JanataParty, which was still relatively new in the early1980s, so it could compete with the older, well-established Indian National Congress. Similarly,President Gonzalo Sánchez de Lozada of Boliviahad a strong incentive to increase the appeal ofhis National Revolutionary Movement. Hondu-ran President Rafael Leonardo Callejas wantedto strengthen his political party but also had amore immediately practical incentive�to ensurecontinued access to foreign aid at a time whenCold War�inspired U.S. assistance to CentralAmerica was diminishing.

The legal enabling environmentExcept for Karnataka, in the countries studied

the legal foundations for DLG are quite recent. Inno case was significant enabling legislation en-acted before 1991. Essentially, these cases couldbe said to be still in the shakedown period of DLGimplementation.

The legal foundation for DLG is more funda-

mental in some case study countries than others.In four countries that legal foundation is embed-ded in the national constitution. India�s 73rd con-stitutional amendment, passed in 1993, suppliesguidelines for DLG, providing among other thingsfor representation of women and various minor-ity ethnic groups. The post-Marcos constitutionof 1987 in the Philippines affirms the importanceof DLG. Mali�s 1992 constitution mandates threelevels of local representative government.Ukraine�s 1996 constitution devotes a whole chap-ter to the subject.

These constitutional provisions required imple-menting legislation. In 1993, Karnataka passedan act spelling out its three-tier structure andbegan implementing it.28 The Philippines enactedits Local Government Code in 1991. As inKarnataka, the details take up an entire volume.In Mali, the Mission for Decentralization, theagency established to plan and implement thegovernment�s decentralization initiative, was re-sponsible for developing the enabling laws. Thedecentralization mission spent two years on thistask, and in 1995 and 1996 the national assem-bly voted the key bills into law. The new structureis expected to begin operations with electionssometime in 1998. As for Ukraine, an estimated20 detailed acts will be required to translate itsconstitutional requirements into law, a processexpected to take at least two years. In the mean-time, laws from the early 1990s, combined withexecutive decrees, provide the legal basis for itsDLG efforts.

In the two Latin American countries, the ba-sic justification for DLG lay in national legisla-tion. Honduras enacted its Municipal Reform Lawin 1990 and began implementing it the followingyear. In Bolivia, the Popular Participation Lawwent into full effect, with local elections at theend of 1995.

Basic provisionsThe structures created by this enabling legisla-


tion show striking variations.29 Ukraine has thusfar been the least innovative, leaving in place es-sentially the same system that existed in the So-viet period, with two very important differences:genuinely democratic elections are now held tofill the elective positions, and local governmentunits have some real autonomy in managing theiraffairs.

Bolivia�s new system is perhaps the most in-novative. It provides for elected mayors and coun-cils in the country�s 311 municipalities, as wellas for a parallel system of elected vigilance com-mittees�bodies charged with planning infrastruc-ture investment and monitoring the performanceof the mayor and council.

Karnataka�s system is groundbreaking in man-dating that women must hold one third of allcouncil presidencies and vice presidencies, as wellas one third of all seats on local councils. More-over, the law requires that Scheduled Castes (for-merly Untouchables) and other low castes be givenseats in proportion to their percentage of localpopulations.

In the Philippines the Local Government Codestipulates that special bodies called local develop-ment councils include nongovernmental represen-tatives (to be chosen by the NGO community) tofill one fourth of the positions. And in Mali, thecomposition of the 701 new communes was de-termined with the direct involvement of the con-cerned citizenry.

Many of the local self-governance systems aregenerous in allocating national resources to localbodies. The Philippines and Karnataka allocate 40percent of national funds to local units. Boliviaallocates 20 percent. Mali is considering a gener-ous allotment. Ukraine is unpredictable but seemsto average local allotments of more than a third

of national funds. Honduras, however, has beenrelatively tight-fisted, delivering no more than 1.5percent of national revenue.

Accompanying these resource allocations areincreased areas of responsibility for local govern-ment units, which also vary greatly. Two systemshave devolved virtually everything but police powerand such obvious central responsibilities as inter-city roads and transport. In Ukraine, where the cen-tral government has been incapable of meeting itsobligations, local bodies are responsible for virtuallyall spending on education, energy, health, heat, pen-sions, public transportation, and water and sanita-tion. In Karnataka, local governments provide allthe basic social services, such as health and educa-tion.

Bolivia and the Philippines have devolved spe-cific sectors or subsectors to local levels. In thePhilippines local units handle agriculture, health,sanitation, and welfare; in Bolivia local unitshandle investments in (but not the day-to-dayoperation of) education, health, and recreation.Specific sectors will also be devolved in Mali, butthe final decision on which is yet to be reached.Honduras took a different approach, giving mu-nicipalities general responsibility for planning andadministering specific services, such as water andsanitation.

Strategic implicationsCountries have launched DLG initiatives under

a wide range of conditions, provided a wide rangeof resources, and devolved different assignmentsto local governments. There is clearly no blue-print for taking up DLG in a national develop-ment strategy. But the breadth of this range im-plies that DLG can be an appropriate strategy in avariety of circumstances. It is not surprising, then,

28 Laws passed in 1959 and 1983 provided for earlier versions of local self-governance in Karnataka.29 Fuller details are given in the Impact Evaluation reports (see footnote 18).


that international donors have shown great in-terest in promoting democracy through DLG.

Donor andHost-country Roles

It is axiomatic in international developmentthat any donor-assisted development activity hav-ing even a remote chance of success must be�owned� by people in the host country. The com-mitment, enthusiasm, responsibility, and finan-cial support required to see the effort throughmust, in the end, be supplied domestically, whetherby government or private sector. In the currentera of flagging donor interest and stamina, theaxiom becomes even more apt. This is especiallythe case with DLG, which must be adapted andcrafted to meet each country�s circumstances ifit is to be launched at all. Still, donors will con-tinue to have an important and even critical rolein promoting DLG and helping its momentumbuild to the point where national and local com-mitment can sustain it.

Previous aidAll six countries in this assessment have a sig-

nificant track record in supporting decentraliza-tion, though seldom as democratic initiatives. Few,however, enjoyed earlier success at promotingdevelopment through their decentralization efforts.Most of these efforts have had some external aidin the past, though often in the distant past andindirectly. In India, assistance for decentralizationgoes back to the Ford Foundation�s support forIndia�s first pilot project in the late 1940s. Even-tually (with large-scale official American aid) that

project became the massive Community Devel-opment Program of the 1950s and PanchayatiRaj enterprise of the 1960s.

Since the early 1970s, though, the U.S. gov-ernment has provided no official support.30 In-stead, the national government in New Delhi hasbecome, in reality, the external agency support-ing DLG in India�s federal system. Each state hashad considerable autonomy designing its own ver-sion of Panchayati Raj. The central governmenthas assisted with various efforts, most notably thefive-year plans that began in the 1950s and havecontinued into the 1990s.

Official U.S. assistance to local governments inHonduras goes back to a 1968 program to fur-nish credit to municipalities. Other efforts overthe years supported urban infrastructure andmunicipal management in the major cities andtechnical assistance on a wider scale.USAID/Honduras assisted in the development ofthe 1990 Municipal Reform Law and then sup-ported the new law�s implementation with theMunicipal Development project. It came on lineshortly after the law was passed in 1990.

USAID�s contribution to decentralization in Malidates back to support for local health-care sys-tems in the mid-1980s. Since the 1991 revolu-tion, USAID has supported the government�s ef-forts to design and implement its present decen-tralization initiative.

In the Philippines, USAID has supported localgovernance since the 1970s, initially by support-ing public administration programs and sponsor-ing local officials for training in the United States.This assistance continued in various forms, in-cluding the Local Development Assistance projectof 1990�95, predecessor to the present Gover-nance and Local Democracy project, which is di-

30 The Ford Foundation has supported a number of local (generally experimental) development schemes in recent decades, ashave other donors.


rectly supporting implementation of the LocalGovernment Code of 1991. As in Honduras,USAID was in close touch with host country offi-cials as the code was being formulated; thus itwas able first to modify the Local DevelopmentAssistance project and then to design the Gover-nance and Local Democracy project to fit in withit.

In Bolivia there was some earlier USAID assis-tance in public administration, although only ona small scale. But Agency officials were in con-stant contact with their host-country counterpartsas local governance legislation was being writtenand implementation plans were being devised, soit could start up assistance efforts once the initia-tives came on line. The Democratic Developmentand Citizen Participation project was ready soonafter the government began carrying out its newPopular Participation Law in late 1994.

These five cases contrast markedly withUkraine, where there was, of course, no USAIDactivity before the Soviet Union�s breakup. In1993, through its Municipal Finance and Man-agement project, USAID began helping pilot mu-nicipalities shore up their badly sagging efforts toprovide government services in a crumblingeconomy.

Strategies adoptedIn the four countries where DLG initiatives have

been launched, USAID has taken an experimen-tal pilot approach in its efforts. Ukraine began with3 pilot cities, Bolivia with 6 municipalities, Hon-duras with 33 municipalities, and the Philippineswith 7 provinces and 1 independent city (encom-passing 168 local government units). In Bolivia,Honduras, and the Philippines, USAID selectedlocalities most likely to succeed. It could be ar-gued that selecting superior sites taints develop-

ment assistance because they are abnormal, andtactics successful there won�t necessarily work inmore ordinary settings. But given the bleak trackrecord of past decentralization efforts, it probablymakes good sense to �build on the best,�31 in thehope that workable approaches pioneered therewill also be effective in less favorable circum-stances.32

However, there is some danger that the build-ing-on-the-best strategy will put �the best� toofar ahead of �the rest,� since the main reasonsome places are the best is that they are moredeveloped than other places. To strengthen theiradvantages with extra assistance could exacerbateregional imbalances and create political problems.Honduras is a case in point. There, the pilot areasnestled along the more advanced north�south axisof the country were the primary target of USAID�sMunicipal Development project. The more back-ward eastern and western regions were not tar-geted and therefore were left without the kind ofDLG resources and support provided the partici-pating project municipalities.

While there may have been good argumentsfor host country governments to begin DLG ef-forts with experimental pilots, the decision in ev-ery case has been to launch the initiative simul-taneously nationwide. Bolivia implemented itsPopular Participation Law in all 311 new mu-nicipalities at one stroke, and the Philippines putits Local Government Code into effect in all 76provinces and constituent local units simulta-neously. Honduras�s 1990 Municipal Reform Lawwas applied throughout the country. Ukraine putits laws and decrees on local governance into ef-fect everywhere at the same time. Mali, by con-trast, initially implemented a somewhat restrictedlocal governance structure in 19 urban communesbeginning with the elections of 1992; when itlaunches the new setup, scheduled to come on


line sometime in 1998, it will cover the wholecountry.

The case for the full-scale approach is obvious:countries eager to push development as fast aspossible do not have time for laboratory experi-ments. In addition, politicians tend to find it diffi-cult to explain why some areas should be selectedfor pilot experiments while others are left out. Itis more expedient to simply include everyone andevery place.

The contrast between these two approachescould create some tension, if the host countryand USAID want to proceed at a different pace,each for its own valid reasons. Interestingly, how-ever, this did not occur in the countries CDIE vis-ited. Host country governments appeared to ap-preciate USAID�s need to concentrate on develop-ing effective approaches, while for their part,USAID Missions seemed to appreciate the hostcountry�s need to treat allareas similarly.

Balancing democracy andpublic administration

DLG brings a democracy component to a de-velopment activity that, for a long time, concen-trated on public administration. One result is thata strategy that had a single purpose now has two:public administration, or outputs; and democracy,or inputs. These two purposes can be comparedin a number of ways (see table 1).

The issue here is how donors balance thesetwo sides of local governance in their programs.The initiatives analyzed in this assessment covera range of possibilities, with Ukraine at one endand Karnataka at the other. In Ukraine, the im-mediate challenge was that local governmentswere unable to do much more than try, from dayto day, to furnish some fraction of the services

the Soviet state had once provided. The immedi-ate need was to keep things working, for unlessthe local administration actually delivered services,citizens would see no point in participating in lo-cal governance. Accordingly, at the outset theproject emphasized administration more thandemocratic participation. As local governmentcapacity improved, the emphasis shifted towarddemocratic participation.

In Karnataka, local administration has alwaysbeen in place; the task was to figure out how toensure citizen participation in directing it. DLGefforts in India, thus, tend to emphasize democ-racy. Different tiers of the system do have admin-istrative components, to be sure. At the highertwo levels of district and taluk (somewhat analo-gous to the U.S. county and township), there areoffices of development specialists in such areas aspublic health, animal husbandry, and road main-tenance. They have been in place more or lesscontinuously, albeit augmented or downsized asstate and central governments have launched,upgraded, or abandoned successive developmentprograms. Whether or not DLG is in place at anygiven time, administrative functionaries remain.The DLG initiatives have therefore concentratedon the democratic, political side of local develop-ment. They have worked on arranging the tiersof panchayats and determining specific powers,revenue allocations, and representation for women

31 USAID/Honduras has carried this notion one step further. Mutually agreed-on project compliance standards were set forthe 33 municipalities. After it was determined that 14 had not met the standards, they were dropped from the project.32 This approach is the one most commonly taken by USAID and other donors as well.

Table 1. The Two Sides ofDemocratic Local Governance

Democracy Public Administration(Inputs) (Outputs)

politics managementparticipation service deliverydemand side supply sidedecision-making policy implementationcivic education institution building


and minorities, rather than on providing services.

One could argue that in Ukraine, USAID shouldhave paid more attention to democracy, whereasin Karnataka the central and state governmentsshould have paid more attention to administra-tion. But the objective situation into which a DLGinitiative is placed will often be unbalanced oneway or the other, and part of DLG�s attraction isthat it has sufficient flexibility to move either wayto right such imbalances. Ukraine had to deal withthe reality of an imploded central government; inIndia the challenge has been to bring popular con-trol to local development.

Most of the other initiatives seem more evenlybalanced, with USAID�s Philippine and Hondu-ran projects emphasizing the administrative sidea bit more, while the Bolivia project was weightedslightly more toward democracy. The Mali case,however, is slightly different, reflecting the coun-try context�a government highly committed toparticipation, but local administrative capacitythat is extremely weak. There, USAID�s assistancehas intermingled democracy and public adminis-tration elements, rather than emphasizing one orthe other.

Strategic implications Building on experience. Previous USAID as-

sistance undoubtedly made it easier to build localself-governance projects in Honduras, Mali, andthe Philippines and may have helped in Bolivia aswell. This was so for both donors and host coun-tries. USAID learned something about how tosupport decentralization initiatives, and host coun-try professionals built expertise in the field throughtheir involvement. In Ukraine, where neither do-nors nor local officials had experience with localself-governance, both found it harder getting pro-ductive activities rolling than did their counter-parts in the other countries.

There are also some differences between do-nors and host country governments. USAID di-rect-hire officers rotate in and out, so institutionalmemory suffers, whereas host country profession-als often stay in place much longer and can recallexperience from decades past.33 Valuable as suchexperience has been in certain settings, such asthe Honduras and Philippines, it is apparentlynot essential. The Ukraine project may have hada tougher start, but it did get off the ground and,in the judgment of the CDIE team, had made genu-ine headway.

Experimenting in DLG. For donors, pilot effortsare valuable experiments in DLG. In theory, thisshould hold true for host governments, but in thereal world of political demands and service deliv-ery shortfalls, governments are often compelledto launch a program everywhere if they do itanywhere. Donor pilot approaches can still beuseful, in that they facilitate accumulation of use-ful experience that can be applied on a larger scalewhen the national program is up and runningand able to receive outside assistance.

Balancing democracy and public administra-tion. Some countries need to give more attentionto the administrative management side of DLG,while others need to concentrate more on democ-racy, and still others need a more even mix. Incountries such as Ukraine, where government atall levels has trouble delivering basic services, ad-ministration clearly demands first priority. Thiswill be true in many countries making the transi-tion to democracy. Where services have been inplace for some time, as in the Philippines, a stron-ger emphasis on democracy is in order.

Resistance andPolitical Will

Democratic decentralization is an exercise in


shifting political power, which inevitably createsapprehension, jealousy, and often outright oppo-sition. In most developing countries, most politi-cians and civil servants operating in the nationalpolitical arena perceive that power will be devolvedto the local level at their expense.34 This is espe-cially true in the unitary states characteristic ofmost of the developing world.

It is not surprising, then, that decentralizationaroused anxieties and resistance at the central levelin all of CDIE�s case studies. Opposition came fromboth political leaders and civil servants. It oftenhad two sources, related to both the constructiveand dysfunctional roles politicians and bureau-crats can play in development (see table 2).

Politicians want to determine national policyand launch programs to implement it, whetheras ministers in the executive branch or as mem-bers of the legislature. This is as it should be; it ispart of their calling as political leaders. Nationalpoliticians must have vision and ambition if acountry is to progress and develop, and they needto be able to turn those dreams into reality. Powerdecentralized too much�power fragmented�canthwart such ambitions, so it is understandablethat good politicians may object to decentraliza-tion.

But politicians can also object for less worthyreasons. All too often political leaders use the pre-rogatives of office to build up their own powerbases and personal fortunes, rather than to de-ploy national resources in the national interest.And for national politicians, it is often simpler tosiphon off public resources when they are cen-tralized than when they are dispersed through de-centralization.

Similarly, good bureaucrats rightly feel an ob-ligation to maintain integrity and high standardsin government activities and may well believe thatturning control over to less well trained local lead-ers could put those objectives at risk. Could a vil-lage council, for example, be entrusted to main-tain national reading standards as a teaching ob-jective in primary education? Conversely, the wishto keep central control often masks a desire tohang on to the perquisites of office and the op-portunities for illegal financial gain that aboundfor bureaucrats in the developing world. The ma-jor threat decentralization poses for many bureau-crats may simply be the drastic drop in their op-portunity for corruption.

In all our case studies, resistance came fromboth politicians and bureaucrats. Sometimes itcame directly from the top: in Ukraine, the presi-dent abruptly decreed that oblast (provincial) gov-ernors would be appointed rather than elected,making them accountable to him rather than thevoters. More often, political opposition came fromother sources. Provincial bosses in the Philippines,long the masters and primary beneficiaries of thepulitika system of patron�client relations, provedunderstandably reluctant to give up power to lo-cally elected councils. Many objected even morestrongly to new requirements for NGO represen-tation on local development councils. Largely be-cause of their objections, four years after DLGimplementation had begun, only slightly morethan half the local government units in USAID�sGovernance and Local Democracy project sites hadany NGO representation.

In Bolivia, presidentially appointed prefects(provincial heads) assert control over municipaldecision-making through a matching-fund ar-rangement in which provincial funds are added

33 This can be true also of USAID foreign service nationals, though generally they do not have the long experience thatcharacterizes many host country government development professionals.34 Many local political leaders in industrial countries have similar reactions. In the 1960s, for example, many mayors wereangry when the U.S. government sought to create community institutions in urban neighborhoods, bypassing them (see Blair1982).


to selected municipal projects. Similarly, nationallegislators in Honduras have refused to appropri-ate the mandated 5 percent of national revenuesto municipal budgets. In Karnataka, leaders inthe Legislative Assembly have tried to underminethe panchayat structure on several occasions overthe years. Most recently, for example, they refusedto hold local elections in the early 1990s andwhittled down the authority devolved in the 1993Panchayati Raj Act from what it had been in theearlier 1983 act.

Bureaucratic opposition takes different forms.In Mali the Mission for Decentralization was re-peatedly impeded by its parent ministry. Onlywhen it was transferred to the prime minister�soffice�an excellent example of commitment andpolitical will�was it able to function effectively.In the Philippines and Honduras, central line min-istries have been reluctant to give up control oversectoral field programs as completely as DLGimplementing legislation demanded. In Ukraine,the Ministry of Finance has jeopardized localgovernment�s fiscal autonomy by exercising itspower to approve local budgets and, significantly,to alter the proportion of taxes collected thatmunicipalities are permitted to keep. Anothercommon pattern is for the central ministry torelinquish control over policy and implementa-tion but maintain jurisdiction over personnel. InBolivia and Karnataka, salary levels and postings

are determined in thecapital city. The Philip-pines has a similarpolicy for postings, butonly for financial offi-cials.35

The importanceof political will

Historically, themain reason decentrali-zation initiatives havefailed is that central

governments have been unwilling to devolveenough power for local governments to functionproperly. The opposition of politicians and bureau-crats has been strong enough to weaken or killmost initiatives. In the cases CDIE studied, how-ever, decentralization efforts have succeeded de-spite serious opposition, largely because there hasbeen enough continuing political will at the high-est level to overcome or at least neutralize resis-tance. The government leaders in Bolivia, Hon-duras, and Karnataka came into office with localgovernance as a major objective. Each leaderpushed the initiative through to implementation.In the Philippines, the Local Government Code of1991 originated in the senate, but presidentsCorazon Aquino and Fidel Ramos both backed itstrongly. Similarly, Mali�s President Alpha OumarKonare has staunchly backed his country�s initia-tive. Even in Ukraine, where opposition was fierce,the support of presidents Leonid Kravchuk andLeonid Kuchma was critical to DLG�s success.

In all six cases, local governance represented adistinct break with the past. All six systems hadbeen highly centralized,36 so it took sustained po-litical courage to chart a new direction�in mostcases through more than one leadership cycle. Twosuccessive presidents backed centralization inHonduras, the Philippines, and Ukraine. InKarnataka, where there was a lapse after the1980s, it took a constitutional amendment in

Table 2. Sources of Opposition toDLG From Politicians and Bureaucrats

Role Politicians Bureaucrats

Constructive, Set policy, initiate Ensure standards,prodevelopment programs maintain program

integrity

Dysfunctional, Build patronage, Maximize perks,antidevelopment increase own wealth seize opportunity for

corruption


1993 to relaunch the initiative. In Bolivia muchdepends on whether President Gonzalo Sánchezde Lozada�s successor supports local governance.Without such backing, the Popular ParticipationLaw will probably wither. On the eve of full imple-mentation of its local self-governance initiative,Mali appears to enjoy ample political will, basedon support from the president, prime minister,and people.

While the political will needed to initiate andsustain DLG, at least in its early stages, is prima-rily found at the top, it is hard to find at the grass-roots level. There may have been supporters atthe district or provincial level, as Manor (1997,26) reports for a few cases, but genuine popularsupport has been scarce. Perhaps it will build overtime, but there is little sign of this so far. InKarnataka, when the bold local governance ini-tiative of the mid-1980s was scuttled by a subse-quent state government, there was little protestfrom the citizens despite a long history of DLGprograms. Among our sample cases, however, Malimay be an exception, since its initiative has evokedsuch widespread enthusiasm that any future na-tional government would probably face seriousproblems if it tried to stuff the DLG genie backinto the centralized governance bottle.

Strategic implicationsDLG will inevitably be perceived as injurious

to strong players in a political system, so there issure to be some resistance from politicians andbureaucrats. Donors need to understand whatunderlies such resistance. When legislators opposelocal governance for fear of losing their links toconstituents, for example, astute donors will helpfigure out how those connections can be main-tained. The challenge is to convince both politi-cians and bureaucrats that the devolution of power

will not be at their expense, and that policy canbe fruitfully initiated and high performance stan-dards maintained at many levels. Donors shouldprobably continue the political dialog with nationalleaders well past the first blush of a DLG initia-tive. Resistance will inevitably continue, and po-litical resolve at the top will need to be reinforcedfor local governance to take root and become sus-tainable.

Representation,Empowerment,And Benefits

Much of DLG�s attraction as a developmentstrategy lies in its promise to include people fromall walks of life in community decision-making.The hope is that as government comes closer tothe people, more people will participate in poli-tics. All sorts of constituencies�women, minori-ties, small businessmen, artisans, parents ofschoolchildren, marginal farmers, the urbanpoor�will then get elected to office (or havegreater access to those in office). That will givethem representation, a key element in empower-ment, that is, a significant voice in public policydecisions that affect a group�s future. Local policydecisions reflecting this empowerment will servethe wider public, providing more appropriate in-frastructure, better living conditions, and en-hanced economic growth. These improvementswill then reduce poverty and enhance equityamong all groups. As a formula, the argument37

could be written as follows (with the ▲▲▲▲▲ symboliz-ing an increase in):

▲▲▲▲▲ participation ÿ ▲▲▲▲▲ representation ÿ

35 In Mali matters of postings and salaries remain to be worked out.36 Karnataka was something of an exception in having had various forms of Panchayati Raj beginning in the early 1960s.But even there Chief Minister Ramakrishna Hegde�s 1983 initiative devolved authority more boldly than any previous effort.


▲▲▲▲▲ empowerment ÿ ▲▲▲▲▲ benefits for all ÿ▲▲▲▲▲ poverty reduction

RepresentationAssuming that a central government has a

genuine commitment to devolving power, the veryfact that democratic decentralization is takingplace means many new constituencies can gainrepresentation through public office. Businessmen,local notables, large farmers, professionals, andpossibly some labor union leaders will quickly finda place on local councils. Some will already haveenjoyed influence through links to the pre-DLGpolitical structure, since central governments, nomatter how authoritarian, always need a few tiesto the local level. Some local people will have hadprivileged access to higher levels of governance.But with DLG that circle of influence will openup local governing bodies to the electoral process.

Expanding representation will increase the like-lihood of pluralism and competition, for differentconstituencies will often disagree with one an-other and among themselves. Some localbusinesspeople may want to restrict outsiders fromentering the retail trade in dry goods, for example,whereas others will see opportunities for them-selves in opening local markets to external par-ticipants. But local elites probably are more likelyto collude in their own material interest than tocompete. For example, local business operatorsmay want to limit the number of licenses a mu-nicipality can issue for selling fertilizer or run-ning taxis, thereby keeping prices higher for thepublic and added profits for themselves. Thus ifonly the local elites gain representation on localgovernment councils, the public is unlikely to bewell served.

In our six cases, two other constituencies have

gained representation: ethnic groups and women.Ethnic groups that previously had no real politi-cal voice nationally have been able to assume con-trol of local governments in areas where they forma majority. In large areas of Bolivia, Quechua andAymara community representatives now sit onmunicipal councils that didn�t exist before imple-mentation of the Popular Participation Law. Inthe Philippines� Cordillera region, indigenousgroups such as the Kalingas and Gaddangs nowenjoy majorities on local government bodies. Simi-larly, councils in new rural communes in Maliare likely to show heavy representation from suchethnic groups as the Songhai and Dogon in areaswhere they are numerically strong. In Ukraine, itis not a question of ethnic minorities but of cul-tural and linguistic divisions. Ukrainian speakersdominate in the west and Russian speakers in theeast and south, so elected local governments aretypically dominated by members of the local lin-guistic majority.

Women are on local councils in countrieswhere their inclusion is mandated. Karnatakarequires that one third of elected members of alllocal bodies be women and that women hold onethird of council presidencies and vice presiden-cies. Karnataka has also mandated that the Sched-uled Castes (the former Untouchables) be allottedmemberships and executive positions equal to theirproportion of the population in the area.38 In Malithe major political parties are promising similarreservations for women, but no decisions havebeen made yet.

Where female representation on local councilsis not mandated, women have fared poorly in elec-tions. Honduras and Bolivia boast few women onlocal councils. In Bolivia the percentage of localelected offices held by women is actually lower (8percent) than it was under the earlier, more re-

37 The most committed proponent of this view is probably David Korten, who has made the case for it often (Korten 1990).USAID�s New Partnerships Initiative uses a somewhat similar logic. Its strategic approach emphasizes partnering between civilsociety, the business community, and democratic local governance (USAID 1997a).


strictive system (11 percent). And in the Philip-pines, women hold only about 10 percent of electedlocal offices. In Ukraine, where the overriding con-cern is to keep local governments functioning, gen-der representation has not surfaced as an issue.Women were relatively well represented in profes-sional and public life under the old regime, andsome women serve on elected councils. It appearsthat the proportion of women is fairly high on ur-ban neighborhood and rural village councils, butmuch lower in the higher tiers of government.

There are other ways to ensure representation.In the Philippines, the Local Government Coderequires that at least 25 percent of the votingmembers on local development councils be fromNGOs, chosen by local NGO constituencies. Bo-livia has taken a different approach setting upparallel vigilance committees to monitor electedbodies. Vigilance committee members are selectedfrom traditional local governance systems, suchas peasant syndicates and neighborhood councils.This composition establishes continuity betweenolder and newer local self-governance systems.

To ensure representational integrity in Mali�snew DLG system, the Mission for Decentraliza-tion set up special ad hoc regional and local groupsto help determine the territorial configuration ofthe new communes. These broadly constitutedgroups convened public meetings in which citi-zens determined which villages should be linkedto form the new communes. In the process, theexisting 270 local government units became 701communes. Presumably the new system will re-flect natural social divisions and allow self-identi-fied neighbors to manage their affairs together.Meanwhile, unlike many African countries, Malihas allowed the hereditary village chiefs to remainin place, although their role in the new system

remains unclear.

EmpowermentThe third element of the formula argues that a

group�s increased representation will lead to em-powerment, but local governance has deliveredonly partially on this promise. There is some goodnews: local governance has brought empower-ment to ethnic groups that are minorities nation-ally but geographically concentrated in certainareas. Andean councils can now steer investmentstoward primary schooling in Bolivia, for example,and Cordillera municipalities in the Philippinescan enforce legal restrictions on land purchasesby outsiders. In some urban areas, poor neigh-borhoods can take charge of their community�sdestiny by winning electoral control of local coun-cils, as in large sections of El Alto, the lower classbedroom municipality next door to wealthier LaPaz inBolivia.

On empowerment for women, however, thenews is not as good. There is no indication thatwomen acting consciously as a group have hadmuch effect on local public affairs in Bolivia orHonduras (possibly because they have not achievedsignificant representation). Worse, in Karnataka,women have played little part in council affairs,tending to remain silent or participate only as theirhusbands direct. Consider the municipal corpo-ration of Bangalore, Karnataka�s capital city,which has a population of well over a million.The mayor and deputy mayor are both women,as are 34 of the 100 elected corporation mem-bers, yet in all but a few instances their husbandscontinue to call the shots. At the very least, wemust conclude that even bold affirmative actionon representation does not easily or rapidly em-

38 In its latest (1993) version of DLG, Karnataka�like many other Indian states�requires the so-called Other Backward Castesof Shudra communities�traditionally next to the Scheduled Castes at the bottom of the Hindu hierarchy�also be givenmemberships in local councils proportionate to their share of the population (so far there is no similar quota system for theirbecoming council presidents). The mechanism for implementing these quotas is to declare specific seats reserved for women,Scheduled Castes or Other Backward Castes, allowing only people of that group to run.


power women.

Similarly, the Untouchables (Dalits) ofKarnataka�s councils either don�t participate ordo so only at the direction of their local patrons.This is discouraging, because the Scheduled Casteshave had mandated representation in elected lo-cal government systems since Panchayati Raj wasintroduced in the early 1960s. After almost fourdecades, they still do not exercise a significant voicein local affairs. They are typically unable to de-mand more equitable siting of water supplies andelectricity lines, two areas of public infrastruc-ture investment that have been notoriously ab-sent in areas where they live. If this has been thestory for the Untouchables after almost 40 years,is it reasonable to expect much more for women,whose representation has been mandated for onlya decade?

However, there may be greater scope for indi-rect empowerment at higher levels of government.At the state level in India, for example, Dalits andwomen�s groups have brought enough pressureto bear on political leaders to wring concessionsfrom them on such matters as university facili-ties (of particular interest to Dalit organizations)and the regulation of alcoholic beverages (of spe-cial concern to women�s organizations). In thePhilippines, USAID�s Governance and Local De-mocracy project is experimenting with similarways to give disadvantaged groups a voice nation-ally.

And there is also some evidence that ordinarycitizens feel more empowered through local gov-ernance. In the Philippines, for example, recentsurveys by the Social Weather Stations� pollingorganization found that citizens believe they candeal more effectively with local government thannational government and are also more satisfiedwith local government. Perhaps more significant,there are smaller class differences in this feelingof citizen effectiveness regarding local government.At the lowest (barangay) level, lower class people

tend to see themselves operating at about the samelevel of effectiveness as do the higher classes. Onnational issues, class differences are much wider(Rood 1997). These survey results do not allowus to connect the feelings of empowerment withspecific donor efforts, such as the USAID GOLDproject, but they do indicate that USAID is work-ing in fertile areas in supporting coalitions ofwomen, the urban poor, and fisherfolk.

Distribution of benefitsAs the formula set out at the beginning of this

section suggests, the degree of empowerment af-fects the distribution of DLG benefits. InKarnataka, local elites still control elected coun-cils and steer most benefits to themselves, reflect-ing the lack of empowerment of marginal groups.For example, they upgrade existing high schools(which their children attend) rather than spendpublic funds to expand primary and secondaryeducation (to serve all children). In Ukraine, busi-nessmen (or �biznizmen� as they are referred topejoratively) and entrepreneurs who have sprungup with the collapse of the Soviet system havesteered many benefits their own way by captur-ing privatized state enterprises through bribes andinfluence peddling and evading local taxes throughintimidation and payoffs.

In areas of Bolivia and the Philippines whereethnic minorities have taken over local councils,it is too early to tell whether benefits will be broadbased or go primarily to a few community lead-ers. In one Bolivian municipality CDIE visited, thecouncil decided to invest in school constructionin remote hamlets, an initiative that should spreadbenefits widely. But the same council also decidedto build an automobile racetrack and sports coli-seum. Income generated by these ventures couldgo toward improving primary health care (for ex-ample, to promote treatment of scabies, a chronicdisease in the area), or it could go to a few coun-cil members as kickbacks for construction con-tracts or skimming from gate receipts. As for theCordillera, it is unclear whether community lead-


ers will regulate local mineral extraction to mini-mize environmental damage and maximize roy-alties for local investment, or sell off long-termcontracts for mining rights and pocket the pro-ceeds themselves.

Outside the regions where specific ethnicgroups dominate, the evidence is also mixed. Insome areas of the Philippines, for example, localbosses have simply spread their tentacles to en-velop the new councils, rigging elections and ap-pointing their own retainers to the seats reservedfor NGOs (or ignoring the provisions for NGO rep-resentation altogether). The flow of benefits re-mains unchanged.

In other places the story may well be different.The data gathered by the Social Weather Stationssurvey, which show an increased personal senseof political effectiveness, may imply that benefitsare being distributed less unevenly. In Bolivia andHonduras, there is stronger evidence that localcouncils can undertake activities benefiting con-stituents across thesocioeconomic spectrum. At the insistence of itsvigilance committee, Cochabamba municipalityis building primary schools in its outlying (andpoorer) neighborhoods, moving away from its oldpattern of building them mostly in the wealthiercentral area. Other initiatives in Cochabamba toimprove the city�s central hospital and reconstructa municipal park will benefit everyone. In manyHonduran municipalities, considerable effort isgoing into providing sanitary drinking water foreveryone.39

Poverty reductionAlthough USAID has not specified poverty re-

duction as a goal of local governance initiatives,other donors have. The World Bank, for example,lists poverty alleviation as one of two overarchingthemes (along with environmental protection) tobe pursued in its own assessment of decentraliza-tion (Binswanger and Shah 1994, 4). But there islittle evidence so far that DLG initiatives can domuch directly to reduce poverty, at least in the

short run (Manor 1997, 90�91).

The main reason for this short-term pessimismis that when governance is decentralized, localelites get most of the power and steer benefits tothemselves, or at least maintain the existing dis-tribution patterns (which largely benefit them,anyway). Local elites may be even less likely thannational elites to target government resources tothe poor. In any event, the CDIE country studiesdid not find them any more willing to do so, afinding Manor also reported (1997, 90). Manornotes that he has yet to discover evidence of anycase where local elites were more benevolent thanthose at higher levels (1997, 90).

Still, many local government activities benefitthe poor, although they are not targeted specifi-cally to them. Providing electricity or water forall citizens, for example, can help raise productiv-ity for rich and poor alike. Human capital invest-ments in education and public health now beingundertaken in many areas of Bolivia promise toincrease productivity of both poor and nonpoor.But such undertakings, especially in education,are uncertain and will take a number of years tobear fruit.

Strategic implicationsIt appears that as we proceed analytically

through the representation�empowerment�ben-efits formula, the results become successively lesssubstantial. DLG initiatives to date have increasedrepresentation, but provided less in the way ofempowerment, and even less in making the dis-tribution of benefits more equitable or reducingpoverty. However, this does not mean that pro-moting DLG is a futile endeavor. Increased repre-sentation offers significant benefits in itself. Whenwomen or Dalits hold public office in Karnatakaand sit on local councils, it demonstrates thatmales of hereditary privilege are not the only oneswho can enjoy such positions. The presence ofNGO representatives on local development coun-cils in the Philippines shows people there are al-


ternatives ways to participate in local decision-making where traditional political bosses still havetoo much power.

A second benefit of increasing representationamong marginal groups is its attendant leader-ship experience. Karnataka�s new structure pro-vides 85,000 elective positions, which meansmore than 28,000 women are panchayat mem-bers at any given time. Among them, some 1,900women are presidents and an equal number arevice-presidents of their panchayats. While manywill fail at these tasks or enjoy only modest suc-cess, some will acquire skills that will equip themto be leaders in other endeavors, for example, incivil society or in state or national politics. Simi-larly, representatives of the Scheduled Castes andOther Backward Castes are gaining leadership skillsin their reserved slots. Indeed, a substantial num-ber of present-day state and national leadersamong the Scheduled Castes�who often have sig-nificant political clout�came from this group.

Increased representation on local councils alsomakes it clear to children of both sexes and allethnic communities that they can aspire to pub-lic service. For example, a Dalit girl who has anaunt or neighbor on the village council is morelikely to have higher ambitions than one who per-ceives inevitable exclusion as herfuture.

These promising benefits of increased repre-sentation, however, will take some time, perhapsdecades, to unfold. Scheduled Caste reservationshave been in place since the beginning ofPanchayati Raj almost 40 years ago. This is alonger term prospect for change than donorsmight want, but it is surely worth the effort. InBolivia and the Philippines, it will be some time

before previously marginal minorities can trans-late the empowerment they have gained in theirgeographic areas of strength into improved lives.People new to making public decisions have a fairlylong learning curve before they can actually shapelocal public policy in ways that benefit their con-stituencies. And even when they do, benefits maybe long in coming. Building new schools, for ex-ample, doesn�t mean that effective teachers willimmediately be there to staff them. And it takestime to educate successive groups of children, andlonger still for them to use that education in mak-ing their lives more productive. In sum, while weare most probably talking in terms of decades here,not five-year donor projects, these kinds of im-provements in public life can start with donor-supported DLG initiatives.

One last question to consider is how narrowlyfocused empowerment initiatives must be in thoseareas where marginal groups are in the minority,which is the most common situation. Some do-nors think empowering these marginal minoritygroups should be a fundamental goal of DLG, whileothers view empowerment in more general terms.In its policy document on governance, the UnitedNations Development Program refers to helping�empower people previously excluded from deci-sion-making.� It lists poverty eradication and gen-der equity among the goals of decentralization(UNDP 1997a, 19; and 1997b, 8�9). The guide-lines of the Development Assistance Committeeof the Organization of Economic Cooperation andDevelopment speak of empowerment as �increas-ing access to and influence over resources andinstitutions� by marginal groups, such as low-income people and women (DAC 1995, 8).

Within USAID there are similarly contrastingviews. The Philippines Mission, for example,

39 Of course, it is possible for these efforts to produce less desirable results. The schools in poor neighborhoods may be sounderfunded as to provide no real education, the central hospital may use its resources mainly on high-tech procedures servingfew patients, the park may charge an admission fee that will restrict its use to the wealthy, and the pipes for drinking water maysomehow never stretch beyond the more well-to-do sections.


40 See USAID (1997a, vol. II, 77�86); also USAID/Manila�s web page (http://www.usaid-ph.gov).41 See the NPI Resource Guide (USAID 1997a). In its introductory chapter, the guide uses various forms of the word�empower� 22 times, �local government� or �local governance� 53 times, and �civil society� 89 times.

stresses the connections between its efforts to sup-port DLG, civil society, and empowerment of thedisadvantaged.40 However, USAID�s principal stra-tegic statement on the topic is more general, stat-ing that the Agency�s role is to �stimulate andreinforce democratic elements at the city and com-munity level�(USAID 1994, 18). And the New PartnershipsInitiative�s Resource Guide deals at length withlocal empowerment, but does not specify whatgroups or elements will gain empowerment. Itspeaks instead of empowering communities andcivil society through DLG.41

The issue is whether empowerment can en-able marginal groups to use local politics to pressfor policies of direct benefit to them, or whetherthe whole local community must be empowered.Is it feasible to expect marginal groups, such aswomen, minorities, and the poor, to pressure lo-cal governments into delivering services more eq-uitably by targeted efforts or redistributing ongo-ing activities? Or is it more likely that such groupswill garner benefits only when they participate incoalitions that push for efforts that benefit theentire community? For example, are marginalgroups more likely to be able to persuade localgovernments to set up programs to educate girlsand build sanitary wells in poor areas, or push forinitiatives that offer universal primary educationand provide sanitary drinking water for everyonein the community?

Perhaps in some places traditionally marginalgroups can be mobilized to act in their own be-half locally, even where they are minorities. Thisappears to be true in some parts of the Philip-pines, but not in other parts where local bossesstill maintain control. Elsewhere, includingKarnataka and a number of Indian states, mar-ginal groups may be unable to act locally but canpress successfully for their own interests at a

higher level. In still other places, such as Boliviaand Honduras, marginal groups can direct ben-efits to their members only through universal ef-forts that benefit everyone locally.42 For the DLGstrategist, this means that great care and sensi-tivity will be required to settle on an appropriateapproach to empowerment. Strategies that willwork in settings such as upland Philippines ap-pear quite unsuited to areas like rural Karnataka.

Fiscal AutonomyAnd Regional Equity

In all the case study countries, central govern-ments granted significant authority to local gov-ernments. In Ukraine, because of the centralgovernment�s inability to provide services, localgovernments are basically free to undertake what-ever they can find the money for. In Karnataka,almost all traditional government services exceptpolice power have been transferred to local au-thorities.43 But national and state governmentsthere earmark about 85 percent of funds for lo-cal use, specifying the general programs (such ashealth or agriculture) for which they must be used.Every district in the state, for example, must spenda given amount per capita on malaria eradica-tion, whether malaria is endemic in the area ornot. In theory, Honduras also gives localpolicymakers latitude, but the actual allocations(1.5 percent of national revenue) severely limitwhat local governments can do.

Bolivia and the Philippines have transferredfewer sectoral responsibilities to local government:in the Philippines, agriculture, health, sanitation,and welfare, but not education; in Bolivia, educa-tion, health, and recreation, but not agricultureor sanitation. The major difference between thetwo systems is that Bolivia has confined activityto investment in the transferred areas (and the


42 This applies in industrial countries, too. William Julius Wilson, in his well-known works The Truly Disadvantaged (1987)and When Work Disappears: The World of the New Urban Poor (1996), argues that U.S. social programs to help minoritiesand the poor succeed in the long run only if they benefit all groups, because only then is it possible to build the constituencyneeded to launch and maintain such efforts.43 In no case was police authority transferred, and except for informal courts with limited jurisdiction in Karnataka, none ofthe sample countries transferred any judicial authority.

central government retains control of civil ser-vant salaries and careers), whereas the Philippinegovernment has devolved all activity in the speci-fied sectors, including employees and payroll. Bothcountries have loopholes in restrictions. Becauselocal governments in the Philippines are allowedto take on activities that support the general wel-fare, some have become involved in environmen-tal projects. In Bolivia some local governmentshave launched sanitation projects under the healthrubric.

Fiscal autonomyMost literature on decentralization holds that

fiscal autonomy, or financial independence fromthe central government, is essential for effectivelocal government. The rationale is the lower thelevel of government, the more nearly it can ap-proximate a political marketplace in which citizensdetermine what they want and are willing to payfor in taxes and fees. And when government ser-vices are tied to payment for those services, citi-zens can voice their preferences to elected office-holders and not reelect those who do not act onthese preferences. Thus if citizens want moreschools or better health care, their political leadersshould be able to levy the taxes and fees to financethose improvements. If people think too much isbeing spent on municipal construction or councilmembers� perquisites, they should be able to presstheir leaders to limit that spending.

Even with fiscal autonomy, however, some cen-tral government grants or subsidies are necessary,for few if any local governments are capable ofraising sufficient resources to meet all local needs,even in advanced countries. Indeed, decentraliza-tion initiatives have often failed because centralgovernments devolved too little to the local level.Sometimes central grants have been too feeble,

and sometimes local governments were not givenenough authority to raise money. Sometimes thecentral government allocated enough money butearmarked so much of it that local authoritiescouldn�t spend it where it was most needed. In allthese instances, local governments did not haveand could not accumulate the resources to do whatwas needed.

Judging from the evidence in our sample, atleast two fiscal ingredients are necessary for localgovernments to succeed:

■ Central allocations must be predictable andcommensurate with responsibilities devolved.

■ Local governments must have and use author-ity to raise local resources. Central governmentallocations may cover tasks transferred to thelocal level, but generating local revenue is es-sential to improving local governance andmeeting citizens� needs.

Allocations fromcentral governments

Comparing how countries have managed re-source allocation to help local governments meettheir new obligations is like comparing apples andoranges, since the relationships between grants andresponsibilities differed so greatly. In the health sec-tor, for example, Philippine local governments mustimplement programs, meet the civil service pay-roll, and make needed investments. In Bolivia, lo-cal governments attend only to investments in thehealth sector; the central ministry handles salariesand programs. However, in Bolivia, Karnataka, andthe Philippines, grants to local governments arefar more generous than they were and seem to belarger generally than in most other countries�an


exception to the pattern of central tightfistednessin most decentralization initiatives.

Most of the case study countries were gener-ous about assigning revenue to local governments.The Philippines guaranteed 40 percent of inter-nal revenue allotments, Karnataka allotted 40 per-cent of the state budget, and in Mali there hasbeen discussion that urban communes might re-ceive 40 percent of all national revenue collectedlocally.44 But each government uses a different basefor calculating the share. The �internal revenueallotments� in the Philippines include both cus-toms charges and national income tax, whichKarnataka�s �state budget� does not. Althoughland and sales taxes are included in both, the termsmean different things in the two countries. In-deed, it may well be that the 20 percent of na-tional tax revenues allotted to local governmentsin Bolivia is more in relative (or even absolute)terms than the 40 percent of the state budgetallocated in Karnataka. In Honduras, the com-mitment of 5 percent of national revenue is clearlylow, no matter how national revenue is calculated,and in the first four years of local governance pro-grams allocations never went above 2 percent.

Central governments in Bolivia, Karnataka, andthe Philippines made good on their commitmentsfrom the start. For these cases, whatever the levelof allocation made, local governments could counton a specific share. Even in Honduras, local gov-ernments knew they could count on a certainyearly fraction of the original pledge.

In Ukraine, however, allocations were so un-predictable local governments could never be surehow much revenue they might get. Most taxesare collected locally in Ukraine. Governments atsuccessive levels are allowed to keep a share ofthe various taxes, with the remainder moving upto the next level. Such a system could work ifshares were constant over time, but they vary from

year to year and even from quarter to quarter.Moreover, in any given year, the shares vary fromone local government unit to another. The mostegregious example is the value-added tax, thesource of almost a third of Ukraine�s internal rev-enue in 1996. The proportion of VAT revenuesretained locally ranged from 20 percent to 100percent in the places CDIE visited�a reflectionof the government�s effort to reduce regional dis-parity. Nonetheless, because of the variance inrevenue from the VAT, local governments can nei-ther consistently provide needed services normatch peoples� ability to pay for them with thecost of their delivery. Moreover, mayors and re-gional governors spend much of their time lobby-ing with the central government, haggling aboutretaining or enlarging their shares.

Local revenue mobilizationThe other fundamental issue in fiscal autonomy

is allowing local governments enough authorityto raise funds on their own. Central funds are neverenough to do all the things citizens expect of lo-cal governments, particularly given the height-ened popular expectations that surround DLG ini-tiatives. People expect the local government to domore than the central government, and to someextent it can, though its accomplishments tendto be modest. Localities can do more in a givensector by reallocating funds from one area to an-other (to the extent they are permitted such flex-ibility). Or they can launch efficiency drives tocut local government payrolls, tighten procure-ment procedures, or root out fraud.45

Local governments in Ukraine, desperate tostem the catastrophic decline in public services,grasp at every opportunity to increase taxes andfees, often running up against central governmentrestrictions. At times they are able to exerciseconsiderable ingenuity in getting around suchcontrols. For instance in one city, when the Min-istry of Finance forbade charging street vendors


44 The amount discussed for rural communes is 80 percent.45 In a democratic system, such reform efforts, if overzealous, can backfire. In one Bolivian municipality CDIE visited, thenew mayor fired 93 of 123 public employees and shifted a large part of local governance funds to the water and sanitationsector. When a legal change in early 1997 permitted municipal councils to oust mayors under certain conditions, his councilwasted no time removing him from office. Both dismissed employees and disgruntled consumers of neglected services hadreason to press for the mayor�s removal.

for trash removal, the council responded by re-quiring vendors to buy trash bags from the city ata fixed price.

Some efforts to produce local revenue have beencounterproductive. Ukrainian entrepreneurs com-plained to CDIE that the official business tax rateamounted to 90 percent of their income. Perhapsthey were exaggerating somewhat, but whateverthe actual rates,46 they were high enough to pro-vide about 40 percent of all local revenue in 1996.They were also high enough to drive much localbusiness into the parallel economy, where taxescould be evaded partly (by keeping several sets ofbooks) or even altogether (by operating completelyoutside the official system). The results have beenserious. Not only do local governments lose muchof the money they might collect with lower taxrates and fuller compliance, but also much of thebusiness effort driven underground becomes partof the illicit commercial activity that has becomeso pervasive in Ukraine since independence.

Municipalities in Bolivia have been grantedwide powers to assess taxes and fees and havetaken advantage of those rights to raise more thantwice as much money per capita ($44 in 1995)as they received from the central government($21). Honduras and the Philippines have similarpowers but appear to have used them less. In thePhilippines, for example, funds from the centralgovernment still provide 70 percent of most localgovernment budgets (it varies significantly fromone local government to another). In Karnataka,local governments have barely used their rightsto raise revenues locally, expressing a reluctanceoften found in developing countries to assess lo-cal taxes and fees. This is perhaps because suchlevies would fall most heavily on the wealthy (es-pecially in poorer countries) who are often un-

willing to pay taxes.

When they get up and running, local govern-ments in Mali will also find it hard to raise moneybecause of the poverty there. Mali�s per capitaincome is less than $300, and there is just notmuch money available to be raised. In addition,local governments will have to overcome the in-ertia from a general tax strike that has lasted sev-eral years.

Regional disparitiesSuccess in local revenue generation often cre-

ates a new problem: disparity between regions.Generally some regions are wealthier than oth-ers, and these can generate local public funds moreeasily than the poorer areas. The result, whenfunds are invested wisely, can be growing dispari-ties between regions�rich areas become richerand poor areas continue to struggle. Central allo-cation policies can attenuate these differencessomewhat,47 but can also widen them.

Regional differences were perhaps the most ex-treme in Bolivia. Central funds were allocatedstrictly per capita, but some areas generated fargreater sums locally than others. With the mostdynamic regional economy in the country, SantaCruz department (province), was able to raise $58per capita in 1995. That same year, Potosí depart-ment, which has been going downhill since its sil-ver mines began to play out more than a centuryago, raised only $14. Add to that the $21 per capitagrant from the central government, and Santa Cruzhad $79 to spend for public purposes, whereasPotosí had only $35. Such disparities were evenworse at the municipal level. Among places CDIEvisited, prosperous Cochabamba municipality raised$65 per capita on its own, but economicallystrapped Pucarani managed only $4. Clearly


46 CDIE�s translators were doubtless confused about average tax rate�the rate on all income�and marginal rate. Themarginal rate is charged on the last unit of income, often progressively. The first dollar of income might be taxed at 10percent, with rates steadily increasing so that all income over a certain level would be taxed at a very high rate. This issimilar to the income tax brackets most advanced countries use. Interestingly, the 90 percent figure was cited by all thebusinesspeople CDIE met, so whatever it actually means, it has become something of a symbol on Ukraine�s fiscal landscape.47 American public school financing is a useful analogy. In most (perhaps all) states, some effort is made to reduce differencesacross school districts by increasing per capita state grants to poorer districts, but not by so much as to discourage localrevenue mobilization in wealthier districts.

Cochabamba will increasingly outstrip Pucarani onmost measures of development.

The system in Honduras is similar to that inBolivia. Municipalities can raise local revenues ina number of ways, so similar disparities can beexpected to emerge, especially considering thecentral government�s stinginess in allocations tomunicipalities. The wealthy port city of PuertoCortés, for example, retains a certain percentageof the customs collected there, a resource obvi-ously unavailable to the many poorer interiormunicipalities.

The situation is more complex in the Philip-pines, where some features reduce disparities andothers increase them. Under the formula for cen-tral allocations to local governments, half the to-tal is allocated on a per capita basis, one quarteron the basis of land area, and one quarter in equalshares to all local governments. This means thatthe most sparsely populated areas get a bit more.Those are generally the poorest, most isolated andbackward areas, so the formula tends to reducedisparities. Under the formula for resources man-agement, however, local governments retain 40percent of all taxes and fees collected in their ter-ritory for natural resource extraction, as well as40 percent of any income from joint venture pro-ductive activity based on natural resources.48 Suchactivities are mainly mining and forestry, muchof which occurs in poorer regions, such as theCordillera. That increases equity, but in other partsof the country such provisions are more likely tobenefit areas that already enjoy prosperity.

Ukraine has made the most effort to use cen-

tral government allocations to reduce regionaldisparities. It has constructed an elaborate pro-cess of shifting state and local tax shares to differ-ent areas to ensure that in the end each local gov-ernment has about the same money to spend oneach citizen. Relatively wealthy Kyiv keeps only20 percent of the VAT it collects and has to handover 30 percent of its total revenues to the cen-tral government. Poorer Ternopol keeps all of itsVAT collections and receives about a quarter ofits annual budget as a subsidy from the centralgovernment. As a result, both jurisdictions haveabout the same public money to spend per per-son. However, this creates uncertainty, for localofficials have a hard time figuring out how muchrevenue they will have, given that the VAT is aconstantly changing target.

In the two least economically developed set-tings�Karnataka and Mali�considerations aboutregional disparity are perhaps less relevant. InKarnataka, there has been little local resourcegeneration so it can scarcely create such dispari-ties. In Mali there are significant regional differ-ences�between the wealthier south and poorernorth, for example. Over time, local revenue gen-eration could exacerbate those differences, but ina country of such extreme poverty, even thewealthier areas will not be able to generate muchrevenue, so imbalances could take a long time todevelop.

The policy dilemma here that donors and hostcountry governments must address is clear. If thecentral government tries too hard to equalize re-sources among areas by reallocating public mon-eys, local incentive to raise revenues is dampened.


48 As prescribed in the Local Government Code of 1991, section 290 (Philippines 1992, 116). See Manor (1997, 88�90) fora discussion of the possibilities of equity-oriented formulas.

(This could even encourage laxity in raising rev-enues by rewarding incompetence).49 But if thegovernment gives full play to local incentive byallowing local units to keep all the revenues theyraise, it risks increasing disparities betweenwealthy and poor areas. Clearly a balance betweenregional equity and local incentive is called for,but how the �right� balance can be achieved re-mains difficult to answer.50

Strategic implicationsFiscal autonomy is critical to local self gover-

nance; without it, DLG can be no more thandeconcentration�setting up local field offices toexecute central government policies. But fiscal au-tonomy alone will not ensure the success of localgovernance, even if allocations from the centralgovernment are generous, as they have been insome countries. Those subsidies must be supple-mented by resources raised locally for local gov-ernments to do all the things their constituentsdemand. This is feasible in Bolivia and Ukraine,less feasible in Honduras and the Philippines, andleast feasible in Karnataka and Mali, the poorestof the case study settings.

Where local revenue generation does work, itwill inevitably work better in some areas thanothers, which will increase regional disparities inservices provided as well as quality of life. Effortsto implement equity-oriented formulas,51 such asthose used in the Philippines, would be good tar-gets for donor technical assistance.

PublicAccountability

Democratic governance at the local or nationallevel can succeed only if public servants are held

accountable�government employees must beaccountable to elected representatives, and repre-sentatives must be accountable to the public.52 Awide range of mechanisms can be agents of ac-countability. Each has its strengths, but none ap-pears able to carry the whole load itself.

Bureaucratic accountabilityto elected officials

In all but Mali, significant control over localcivil servants was transferred to local elected bod-ies. When the health sector was devolved to thelocal level in Bolivia, for example, so was supervi-sion of health officers. But in every case this trans-fer of authority has been incomplete. For example,the central government of Bolivia still determinessalary, posting, and tenure, so a mayor could or-der his health officers to keep posted clinic hours,but could not discipline them if they failed to doso. The same is true, more or less, in Honduras,Karnataka, and the Philippines. Ukraine trans-ferred far more authority than the others, largelybecause the central government was unable to domore on its own.53

Ukraine aside, the pattern in the other fourcountries is that devolution, and hence account-ability, has been incomplete. One reason for thishas been the reluctance of central governmentsto truly decentralize. But another significant rea-son has been the unwillingness of governmentemployees to be decentralized. This is especiallytrue of those in the professional fields, such asengineers, health workers, and teachers, who al-most invariably want to enjoy urban amenitiesand send their children to good schools�ambi-tions likely to be frustrated if they are posted in-definitely to rural areas or small towns. Smallwonder, then, that when central governmentsdecree decentralization, bureaucrats in the lowerechelons are quick to declare that such initiatives


49 It could be argued that this does not appear to have occurred in Ukraine, for Kyiv still produces enough local revenue togive 30 percent of it to the central government for redistribution elsewhere. But as time goes by, citizens and businesses inKyiv may come to resent such enforced largesse and begin to demand tax reductions, just as residents and political leaders inupstate New York rail against the �handouts� they perceive themselves to be giving New York City through taxes.50 Surprisingly little attention is paid to this issue in available literature, both academic and donor-produced. Prud�homme isan exception (1994). In most analyses of local governance, especially those concentrating on its democratic aspects, fiscalanalysis tends to get short shrift or be neglected altogether, as Cohen and Peterson (1996, 14�15) observe.51 See Manor�s discussion of the possibilities here (Manor 1997, 88-90).52 It could be argued that local bureaucrats are also accountable to the citizenry, but it is more practical to view the linesgoing from bureaucrat to elected representative to citizenry, because it is the elected official who must in the end sanction thebureaucrat (through his or her supervisory role) and it is the populace who must ultimately exert control over the represen-tative at elections. To be sure, bureaucrats should be responsive to citizens, (this responsiveness is one of the measures ofgood DLG) but they should be accountable to elected officials.

should not apply to them. And central govern-ments almost always back down. At a minimum,they arrange programs so that field officers main-tain links with their original line ministry andthereby enjoy some insulation against local con-trol.

Such an arrangement sounds like a recipe forimpasse between local bureaucrats and electedofficials, but this need not be the case. InKarnataka over the several decades of PanchayatiRaj, for the most part the two have achieved anaccommodation�elected officials direct govern-ment civil servants in their jurisdictions, whilethe line ministries write annual evaluation reports(and determine promotions and postings). Thearrangement is not perfect, but it does work.

Elected officials� accountabilityto the public

Free, fair, regularly scheduled elections anduniversal suffrage are the most direct mechanismsfor ensuring that those who govern are account-able to the citizens. Without elections, local gov-ernment is not democratic. But elections are crudeinstruments of popular control, since they occurat widely spaced intervals (two to five years in thecountries studied here), and address only thebroadest issues. In a well-established system theyallow voters to register general approval or disap-proval for an incumbent�s performance or to se-lect new public managers when incumbents va-cate their offices.

But DLG needs more than this kind of blunt

direction from the citizenry. People must be ableto indicate their likes and dislikes between elec-tions, as well as their views on specific proposals.There must also be ways to publicize citizens�views and uncover wrongdoing in local govern-ment. For this more continuous and fine-tunedaccountability,54 varied arrangements are avail-able.

Elections. Five of the case study countries havemanaged at least one free and fair local election.55

Some have managed more. India has conductedelections regularly since Panchayati Raj was in-troduced around 1960, and the Philippines hashad elections since 1988, but in both countriesthere have been flaws. People in the Philippinesstill speak of guns, goons, and gold as perennialfactors in elections�and in some areas, the chargerings true. Intimidation and vote buying also hap-pen in India. In both places there are public offi-cials who have won and held onto their officesthrough fraudulent means. But on the whole, elec-tions in these two countries as well as the othersin the CDIE sample have reflected the public will.India in particular has shown over many decadesnow that elections can be a powerful tool for ac-countability.

A number of old-line local Filipino bosses, aswell as civil servants, journalists, and others in-volved in civil society in Ukraine and Mali, main-tain that voters are too politically immature, arenot �ready� for democracy, and make uninformedchoices. But such arguments can be and are made


53 The sole exception was the Finance Ministry, which largely retained control over budget officials at all levels, to ensure theintegrity of the governance structure. The result was a somewhat uneasy arrangement in which local officers of the FinanceMinistry in effect worked for both the mayor (or provincial governor) and the ministry in Kyiv.54 This kind of between-elections accountability can also be labeled citizen participation in government, in that suchparticipation is a kind of mirror image of accountability. Citizen participation concentrates on input and involvement ingovernment activities and operations, whereas accountability aims at public oversight and control of government. Thisdiscussion concentrates on the accountability aspects of this duality.

in all democracies. Still, most democratic systemshave endured bad leaders and survived�some re-peatedly�so long as other mechanisms of account-ability were in place.

Political parties. In many systems, oppositionpolitical parties are a powerful engine for enforc-ing accountability. The party in power all too of-ten has strong incentives to evade accountability,but opposition parties have their own incentivesto uncover wrongdoing at city hall and publiclyhound incumbents for their misdeeds. Theypresent a constant vision of a viable alternativefor doing public business�a different and perhapsbetter way. Such a visible alternative helps keepthe party in power on the straight and narrowpath.

This scenario requires a strong party systemat the local level, though, which is often not thecase.56 In Ukraine, after the political implosion ofthe Communist Party, the party system becamefeeble and fragmented. There is one regional partywith some strength in the western part of thecountry, but for the most part there are only piecesof political parties. And although it has been morethan a decade since democracy was restored inthe Philippines, the party system there has yet tostabilize. Instead, it is weakened by personal loy-alties, opportunism, and a tendency for parties tobreak up and re-form. At this early stage in itsdemocratic history, Mali also has a highly frag-mented political system, with nine relatively stableparties and dozens of splinter groups operating

almost entirely at the national level.

The other three countries have reasonablystrong party systems. In Honduras the two-partysystem is vigorous and stable at all levels. Boliviahas been less stable; five parties currently enjoyserious national support, and the party systemhas become the major organizing principle in mu-nicipal government.57 Karnataka had a strong two-party system at state and district levels throughthe 1980s. In the 1990s, that expanded to a three-party structure. Interestingly, however, and incontrast to what seems to be happening in Bo-livia, the parties have not established much pres-ence at the lowest level.

Social capital and civil society. In its New Part-nerships Initiative, USAID emphasizes support forthree sectors�business, civil society, and DLG.Civil society is usually defined as organizationalactivity between the individual (or the family) andthe state. Its democratic role is to advocate forconstituents, to act as watchdog over the state,and to support political competition generally(Hansen 1996). Civil society is essential to DLGin USAID�s strategic vision, but in the six politicalsystems studied it has not played a large role atthe local level. Even in the Philippines, which isat the high end of the spectrum of civil societydevelopment among the six cases, the evidence isnot allpositive.

At the low end of the spectrum, Ukraine hasessentially no civil society at the local level, nor isthere much sign of social capital�the trust that


55 Mali, the sixth country, had local elections in 1992 for 19 urban areas. Local elections for the entire country are supposedto take place sometime during 1998.56 As a corollary, all parties must adhere to some consensus on the rules of the political game. For example, party competi-tion is to be framed in peaceful terms, and elections are the means to transfer political power (as opposed, for example, tomilitary coups d�état).57 In the initial phases of the Popular Participation Law, municipal councils were elected on party tickets, but the parties didnot appear to dominate the election of vigilance committees in most places. This is likely to change as the vigilancecommittees become more integrated into the political system, with its strong party orientation.

facilitates people working together toward a com-mon purpose�which would help build civil soci-ety. The main reason, of course, is the country�srecent Soviet history, in which the state controlledalmost all organized social activity. The whole webof organized life collapsed with the communistsystem, leaving families and individuals autono-mous and without social moorings. CDIE foundpeople in the major cities just beginning to orga-nize in their apartment buildings for such purposesas maintaining common areas, providing security,and dealing with outsiders encroaching on publicspace. A number of donors, including USAID, theEuropean Union, and the Soros Foundation, aresupporting efforts to build social capital and civilsociety, but the road ahead is clearly a long one.

In a way, India is the most startling case, be-cause DLG has been in place there the longest.One would assume that civil society had at leastgotten off to a good start. And so it has at thestate and national levels, where organizations ofwomen, environmentalists, professionals, farm-ers, and others have been part of the political land-scape for some time. But at the village and mandal(township) level, such activity is hard to find inKarnataka. Even civil society�s precursor, socialcapital, appears not to have developed outside ofcaste groups.

Social capital seems reasonably strong in Bo-livia, probably reflecting its tradition of peasantsyndicates and urban neighborhood organiza-tions.58 But it appears not to have carried overinto civil society, for these bodies have yet to be-come participants in municipal political life. Mali,if anything, has a stronger base of social capital,with its rich tradition of associational life and

strong interpersonal networks at the village level.There are signs of a civil society emerging there,including the rapid growth of NGOs (there arenow more than 600) and of registered village as-sociations (more than 2,000) since 1992. Butwhether such groups can form a viable civil soci-ety will not be evident until Mali fully launchesits DLG initiative after the impending communeelections.

Among the six case study countries, only thePhilippines shows signs of a civil society that isan effective instrument of public accountability.And even there the story is mixed at best. In a1996 survey of USAID�s Governance and LocalDemocracy project, only 57 percent of the par-ticipating municipal governments demonstratedclear evidence of genuine NGO participation.59

That participation is likely higher than in the coun-try as a whole, since project sites were chosen inpart because they were likely to succeed at DLG.And the 57 percent figure can be interpreted indifferent ways. For instance, the Local Govern-ment Code of 1991 specifies that one fourth or1,300 positions on local development councilsbe filled by NGO representatives. Only 293 ofthose slots were occupied by representatives freelyselected from the NGO community (rather thanchosen by the mayor) and actively involved incouncil work (Rood 1997, 16�18).

Even with these qualifications, though, it isclear that genuine civil society participation ex-ists in the Philippines in some places, even aslocal bosses continue to dominate in others. Fora country long plagued with elite control at alllevels (Timberman 1990), this is a promisingstart.


58 Many of these bodies became the building blocks for the structure of the community organizations and vigilance commit-tees that have taken shape under the country�s Popular Participation Law.59 Rood (1997, 18). �Clear evidence� means some combination of representatives chosen by the NGO community activelysitting on local development councils, local government investment plans developed with real citizen input, and environmen-tal plans developed in similar fashion.

In Karnataka and Bolivia, civil society has madeless progress, but the reasons for this are not im-mediately obvious. Perhaps the answer lies partlyin the functional nature of traditional social struc-tures, especially in rural areas. Caste councils andAndean farmer groups appear in many ways tohave been mainly governing bodies, whose call-ing was to set rules for members and settle dis-putes�not to act as advocates with local govern-ment agencies or compete with other organiza-tions in getting the attention and support of localgovernment. And there seems to be no obviousway to transform them quickly into players innew DLG systems, because their experience sim-ply doesn�t prepare them for it.

Building social capital and civil society willprobably take a long time, perhaps a very longtime, as Robert Putnam (1993) argues in hisanalysis of their growth over the centuries in Italy.Places such as Bolivia and Karnataka surely donot have as far to go as Ukraine, but there willhave to be much social change before civil societycan become a major engine of social account-ability at the local level in all these settings.

The media. At the macro level, the media havetwo basic roles. The most important is to makepolitical news public. Only if people know whatis going on, good and bad, can they hold theirgovernment accountable. Without vigorous me-dia to spread it, political news remains the prop-erty of the inside few.

The media�s second role is to help uncover gov-ernment misdeeds. This investigative function issubordinate to the main public information func-tion, though, for there are other institutions�political parties, civil society, and the legal sys-tem�to uncover malfeasance. At the local level,this role becomes even more secondary, or per-

haps unrealistic, for small newspapers and radiostations generally cannot afford to sponsor thework necessary to conduct good investigative jour-nalism.

Thus, it is the public information function thatthe local media must perform: to make politicalnews from all sources (including government,opposition parties, and civil society organizations)available to the widest possible audience. Whatthis means in most developing countries is radio.Newspapers tend to be published only in largerpopulation centers, and are accessible only to thosewho can read them (although in many societiesit is traditional for literate people to read aloud tononliterate audiences). Television reception is con-fined to those who can afford a TV set and livewithin the effective radius of a transmitter. Butradio, especially the AM band, is cheap to oper-ate, does not require line-of-sight transmission,and has great audience potential.60 Local news,talk shows, and question-and-answer programsare all excellent ways to spread political newswidely.

This use of radio is probably most advanced inthe Philippines, which has hundreds of local AMstations, many of which broadcast political fea-ture programs. There is also considerable regionalpress in both English and vernacular languages,although largely in the provincial capitals. Boliviaalso has many local radio stations, and at leastone very active network that provides politicalfeatures. Bolivia�s print media, however, originateonly in the largest cities. Honduras has some lo-cal radio, although not as vigorous a system as inBolivia. Its print media have similarly limited cov-erage. Mali has 77 private rural radio stations likelyto become pillars of accountability as the newDLG system takes shape (CDIE found evidence thatthis was beginning to happen). But the low distri-


60 There are at least twice as many radios per 100 households in the six countries as there are TV sets or daily newspapers received.The range in the early 1990s was from 81 radios per 100 households in Ukraine to 4 per 100 in Mali (UNDP 1996, 1997).

bution of radio receivers (4 per 100 people) is aconstraint. In India, as in so many former Britishcolonies, the government retains a monopoly overradio. There is some local press and newspaperspublished in district towns feature some politicalnews, but districts are large (a district in Karnatakaaverages about two million people), so local cov-erage is bound to be thin.

In Ukraine, both print and broadcast media arestill emerging from the heavy control imposedduring the Soviet period. Newspapers, television,and radio are all finding their way in the uncer-tain new world of press freedom. There is someregional news coverage, and there are local talkshows. At least a few mayors have gone on televi-sion to publicize their budgets and invite publicdebate about them. Thus, there is some promisethat the media can become an effective instru-ment of accountability, but for the most part thatpromise has yet to be realized.

Public meetings. Several countries have insti-tuted public meetings to insert civic opinion intolocal governance. A number of Ukrainian may-ors have launched public budget hearings, andpublic hearings have become common for manylocal government bodies in the Philippines. In bothcountries, official council meetings are open tothe public. In two countries, open meetings atthe lowest level of local governance are mandatedby law: five cabildos abiertos a year in Honduras;and two gram sabhas in Karnataka. In Mali, theMission for Decentralization organized an ongo-ing series ofregional and local meetings to involve people di-rectly in the government�s decentralization initia-tive.

The results of these efforts have been mixed.Some mayors in Ukraine and the Philippines usecouncil meetings and hearings to solicit citizen

views and mobilize support for their programs,but others ignore them. In Karnataka, elected of-ficials were evidently so burdened with embar-rassing questions at village meetings that theyquickly abandoned their gram sabhas, or heldthem at odd times or in obscure locations. SomeHonduran officials have made their mandatedmeetings largely ceremonial functions, but oth-ers have used them to involve citizens in localgovernment decision-making. In Mali the meet-ings have been an unqualified success, informingpeople about decentralization and helping thembecome stakeholders in its implementation, forexample, by having them participate in the redis-tricting process that established the new com-munes.

Formal grievance procedures. Two of the casestudy systems have instituted formal proceduresfor citizens to redress grievances against electedofficials.61 The procedures in Bolivia are especiallyambitious. The vigilance committees, whose mainduties are to make plans for local infrastructureinvestments and to monitor municipal budgets,are also empowered to wield a legal instrumentcalled a denuncia against local councils. If con-vinced that its municipal council has actedwrongly, a vigilance committee can lodge a com-plaint with the national executive branch, which,if it finds the denuncia is in order, passes it on toa special committee of the Senate. If the Senatein turn finds merit in the complaint, it can sus-pend central funds to the erring council. A recentlaw also allows municipal councils in certain cir-cumstances to formally dismiss a mayor. Bothmechanisms have been used sparingly thus far.In the Philippines, a recall procedure can be usedagainst governors, mayors, and municipal coun-cil members, somewhat similar to those intro-duced in many American systems in the earlypart of this century.

Opinion surveys. The USAID Governance andLocal Democracy project in the Philippines has


pioneered the public opinion survey as an instru-ment of accountability by building local capacityto design and conduct them. These polls measurecitizen satisfaction with government services.Results to date are promising (Van Sant and oth-ers, 1998, 3�14 to 3�16), but the polls are stillexperiments. Thus, it is not clear whether theycan be replicated on a larger scale or be self-sus-taining once project support is withdrawn. Andquestions of autonomy for local polling organiza-tions have yet to be resolved. Nonetheless, thiswhole area offers considerable promise for futuredonor support in countries with a fairly sophisti-cated survey capacity at the national level, suchas Bolivia, India, and Ukraine.

Patterns of accountabilityThe mechanisms used to promote account-

ability in the six countries CDIE studied rangefrom somewhat effective to virtually useless.Table 3 gives an overall picture of how they havebeen working in these countries. To be sure, suchan exercise has to be impressionistic, since thesamples were small, not necessarily representa-tive, and not equally weighted. Public meetings,for example, are clearly less vital than electionsas instruments for popular control. Moreover,some instruments (civil society, the media, andparties) exist independently of local governanceinitiatives, while others (local elections and pub-lic meetings) exist only as part of them. Finally,there are obvious measurement problems.

The table distinguishes between �viablesources of accountability,� mechanisms proveneffective in ensuring accountability, and �poten-tial sources,� mechanisms with some promiseof making elected officials accountable but notenough of a track record to be considered vi-able. Mechanisms that seem to be in place andviable (that is, effectively ensuring accountabil-

ity) are indicated with two checks; those show-ing promise of being a source of accountabilityhave one check. While these estimations are ap-proximate, they give a rough idea of the account-ability the seven mechanisms provide in the sixsettings.

Four patterns emerge from the table. First, thecase study DLG systems have used varying com-binations of accountability mechanisms, thoughnone has instituted them all. Nor have any twocountries tried the same combination of mecha-nisms. Second, so far, no one mechanism hasproven viable in all settings, or even most of them.Making bureaucrats accountable to elected offi-cials shows the most potential for working atpresent, but it has only shown much real successto date in Karnataka. In the other cases, seriousconstraints remain.

Third, the systems have achieved quite differ-ent rates of progress with the mechanisms ofaccountability they have used. Bolivia has threethat appear viable, Honduras and Karnataka twoeach, the Philippines one, and Ukraine none. Po-tential sources of accountability show another pat-tern, with the Philippines far out in front (sixmechanisms), followed by Honduras, Mali, andUkraine (three each) Bolivia (two), and Karnataka(one). And finally, success does not appear to be afunction of time. The country with the most re-cent DLG program in place (Bolivia) has the mostaccountability instruments up and running, whilethe system with the longest experience(Karnataka) has done less well.

Strategic implicationsThis analysis presents several implications for

donor strategy. First, a number of approaches canwork�there is no �one best way� to promote ac-countability. DLG programs, accordingly, can be

61 In Honduras, jealous political leaders have used the legal system to harass sitting mayors by having criminal charges broughtagainst them, but so far the courts have judged the cases to be without substance.


crafted to any setting. A second implication is thatfor accountability to be effective, several instru-ments must take root and succeed. One, or maybeeven two mechanisms, even if they work, willprobably not suffice to establish accountability ona self-sustaining basis.

A third and more sobering implication is thatmany of these approaches will take a long timebefore they function properly, and indeed maynever do so fully. As long as local bosses retaintheir power in various parts of the Philippines,for example, they will likely be able to bend elec-tions in their favor. And if civil society has notemerged at the local level in Karnataka over thepast several decades, it would seem unrealistic toexpect it to do so in the near future. In some cases,though, it may be possible to devise a mix of in-struments to make up for those that are unlikelyto work. For instance, to hold local governmentsaccountable in the Philippines it might be worth-while to strengthen the party system so that itcould function in tandem with civil society andthe media.

Mali offers a case in point. Its DLG initiative isscheduled for full implementation after the localelections, slated for sometime in 1998. As reflectedin table 3, there are three potential sources ofaccountability to build on: elections have beeninitiated successfully in 19 urban areas; civil so-ciety is manifested in hundreds of NGOs and thou-sands of officially registered village associations;and 77 private rural radio stations have begunoperating. To tap into these sources, one approachwould be to begin with Mali�s self-defined com-mune populations. In effect, the citizens chosetheir own political boundaries for the new sys-tem of local government. In doing so, they in-cluded villages with long-standing affinities andtheir thousands of old and new village associa-tions. This nucleus of civil society should be nur-tured.

A second approach might be to encourageMali�s many splinter parties to form local coali-tions (as USAID-sponsored efforts are doing inUkraine through the National Democratic Insti-tute and the International Republican Institute).Other approaches include encouraging NGOs touse radio stations for public affairs programming,as is done in the Philippines, or helping the mediapromote election debates between local candidatesfrom different parties.

Performance andAccountability

Most of this assessment addresses the democ-racy or �D� component of DLG, but what localgovernments achieve in governance�the �G� sideof DLG�is equally important. For governmentcannot just be something; it must also do some-thing. And if citizens are to think it worth theirwhile to get involved in DLG, local governmenthas to be engaged in meaningful activities. Thiswas most obvious in Ukraine in the early 1990s,when local governments were in danger of foun-dering because the end of the welfare state andrunaway inflation had left them with responsi-bilities let go by the center and without resourcesto meet them.

But Ukrainian local governments have begunto right themselves and to deliver�if still incom-pletely�such basic services as education, energy,social welfare, and transportation. In fact, a gooddeal of USAID�s assistance in the three pilot citiesof its Municipal Finance and Management projectwent first to rehabilitating urban services such astrolleybus lines, public housing management, andland registry. And in these pilot cities, service de-livery has climbed back from the low point ofseveral years ago, although not yet to the levelsmaintained in the Soviet period. Some services(such as housing ownership) are being transferredto the private sector, but most remain in the pub-lic sector, which still lacks the administrative ca-


Table 3. Mechanisms for Accountability

Bolivia Honduras Karnataka Mali Philippines Ukraine

Bureaucratsaccountableto elected officials ✔ ✔ ✔✔ ✔ ✔

Elections ✔ ✔ ✔✔ ✔ ✔ ✔

Political parties ✔✔ ✔✔ ✔

Civil society/social capital ✔ ✔ ✔✔

Media ✔✔ ✔ ✔ ✔

Publicmeetings ✔✔ ✔

Formalgrievanceprocedures ✔✔ ✔

Opinion surveys ✔

✔✔ = viable source of accountability✔ = potential source of accountability

Ele

cted

offi

cial

s a

ccou

ntab

le to

citi

zenr

y

pacity and the resources to do its job well.

The remaining question is whether citizens willget involved if instruments of public accountabil-ity, such as elections and media coverage, can beput into place. Will they become part of civil so-ciety, support political parties, use the media tostay informed, and attend public meetings? Willdonors such as the Eurasia and Soros foundationsand the National Democratic and InternationalRepublican institutes be effective in promotingthese activities?

CDIE found indications that such efforts werebeginning to show results, but it is still too earlyto tell what will happen down the road. For ex-ample, opinion surveys in the Philippines indi-cate that people perceive that services delivered

by local governments have improved and showmore satisfaction with local government servicesthan with those provided by the national govern-ment. And, as noted earlier, people think they aremore effective in dealing with local governmentthan with the national government. In Karnataka,an in-depth survey by the Institute of Develop-ment Studies in the United Kingdom found boththat service delivery had improved with the rein-troduction of DLG in the 1980s (following a hia-tus when local government was administered froma higher level) and that citizens were more satis-fied with it. The authors of the study attributethe improvement to two things: more direct ac-countability of elected officials to the public andbetter flow of information from citizen to govern-ment (Crook and Manor 1994).


Strategic implicationsThe degree to which accountability improves

performance in local government appears to de-pend largely on how far a system has progressedalong a democratic path. Only when mechanismsof accountability have been in place long enoughto flourish can they begin to influence perfor-mance. In places such as Ukraine and Mali, wheredemocracy is just beginning, more effort will prob-ably have to be devoted to getting local govern-ment institutions up and running.

National AdvocacyDemocratic local governance does not operate

in a political vacuum. Like any other new institu-tion, it requires sustenance from outside, both forfunding and for the legal environment that givesit authority and sets its parameters. Early suste-nance can come both from the host country thatsets the initiative in motion and from donors (bothfinancial and technical). But eventually nationalpolitical leaders will move on to other agendas,assigning DLG a lower priority; donors will alsoend their support sooner or later. Indeed, donorsare increasingly shifting their support among sec-tors within countries and closing down their op-erations in more and more countries. If DLG isto endure in this less nurturing milieu, it mustdevelop its own support bases quickly, especiallyat the national level. Most notably, it needs advo-cates at the center to ensure it receives the neces-sary policy attention and fiscal support.

During its fieldwork, CDIE quickly becameaware of the rapid growth of associations carry-ing out advocacy at the national level. Sophisti-cated associations of municipal leaders haveemerged in Honduras, the Philippines, and Ukraine

over the last several years. The Ukrainian Asso-ciation of Cities lobbied to establish a legal basisfor municipal government in the June 1996 Con-stitution. Since then it has been working intenselyto influence the implementing legislation. In thePhilippines, leagues of cities, municipalities, andprovinces have assumed an active role in advo-cating the cause of DLG in Manila. The Hondu-ran Association of Municipalities, after a long andmoribund history (it was founded in the early1960s), became an effective advocate for imple-menting the Municipal Reform Law in the1990s.62 Mali has a mayors� association for the19 urban areas already enjoying elected self-gov-ernance. Curiously, considering the increasinglyactive civil society that has developed in recentdecades at the state and national levels, Karnatakahas no meaningful advocacy associations for lo-cal governance that affect political life at the statelevel.

Strategic implicationsAdvocacy groups at the national level can help

protect funding sources from rival claimants, pre-serve local government autonomy against politi-cal and predatory bureaucratic behavior, promotepublic consciousness through civic education, andhelp maintain a supportive legal environment.Such groups seem to be impressively effective inboth political systems that are pluralistic at thenational level, such as the Philippines, and in sys-tems that exhibit little pluralism at any level, suchas Ukraine.

USAID has supported DLG advocacy organiza-tions in Honduras, the Philippines, and Ukraine,and will support a similar effort in Bolivia as partof its new democracy project there. This supporthas been partly financial, but the technical assis-tance has probably been more important, draw-ing on the long experience of U.S. local govern-


ments advocating their cause in state capitals andin Washington, D.C. Promoting national advocacyorganizations for democracy should be part of alldonor efforts to support DLG.

Conclusions,LessonsLearned, AndRecommendations

Conclusions

D EMOCRATIC LOCAL GOVERNANCE CAN BE A

powerful tool for promoting democratic de-velopment. The potential gains from a well-con-ceived and well-implemented DLG initiative areconsiderable and more than justify the effort. Butlocal governance initiatives cannot do everythingnor fulfill all the objectives USAID and other do-nors wish to realize in their democracy initiatives.This assessment has undertaken to analyze sev-eral USAID programs, the host-country govern-ment initiatives supported by these Agency efforts,and to the extent possible, other donor experiencein the DLG area and relevant literature. From thisoverall exercise, it is possible to draw a number ofgeneral conclusions about the strengths and limi-tations of DLG.

■ Participation and empowerment. DLG can sig-nificantly increase political participation ofmarginal groups, and it can empower geo-graphically concentrated minorities; it appearsunable to empower marginal groups that arenot geographically concentrated, at least in theshort run.

■ Poverty reduction. DLG can help alleviate pov-erty by strengthening the capacity of local gov-ernment to deliver services that benefit thewhole population, in sectors such as educa-tion, health, and water supply; it shows lesspromise in reducing poverty through effortsdirected specifically at marginal groups.

■ Fiscal sustainability. DLG can partly sustain it-self through local revenue generation, but thiswill tend to exacerbate regional imbalancesunless supplemented by central subsidies topoorer areas.

■ Accountability. When a variety of mechanisms,such as civil society, media, and political par-ties are used together, DLG can improve ac-countability of local government bodies to thecitizenry; used in isolation, these instrumentsappear much less effective.

Lessons Learned

1. Representation does not necessarily lead toempowerment, but is valuable in and of itself. Itis possible to increase representation for marginalgroups, such as ethnic minorities and women, bymandating that a certain proportion of represen-tatives belong to a certain group, but this will notlead automatically to empowerment. Still, increas-ing representation for such groups helps improvetheir status, showing the community at large thatthey also can hold leadership positions. Being partof an elected body provides leadership experiencethat can be useful when people run for higheroffice. And when members of previously unrep-resented groups hold office, they serve as poten-tial role models for younger members of theirgroup.

62 In Bolivia a similar association is planned, but had not begun operation at the time of CDIE�s visit.


2. Some groups may be able to attain empow-erment directly through DLG, but for others itwill have to be indirect. DLG can empower geo-graphically concentrated minorities (mainly in-digenous communities and smaller ethnic groups)by giving them the opportunity to command theirlocal governments. But this kind of empowermentis possible only when enough members of a groupare concentrated in a particular area. Marginalgroups that are more dispersed geographically canattain representation in local governance bodies,especially if it is mandated by the central govern-ment, but that is not the same as empowerment.For these groups, empowerment may have tocome more through alliances with the larger com-munity than through their individual efforts.

3. DLG holds promise for reducing povertythrough broad-gauge efforts encompassing the en-tire local population, but much less so for localefforts aimed specifically at the poor. Benefits forthe poor are more likely to come through generalimprovements (such as community electricity,sewage, and water) and a growing economy, whichwill benefit everyone, rather than efforts target-ing one group (such as education for girls, or elec-tricity for poor neighborhoods). This is so princi-pally because it is more feasible to assemble apolitical constituency to support local governmentactivities including the whole population thanthose benefiting specific marginal elements.

4. Fiscal autonomy is critical to DLG. If localgovernments are to serve their constituents andbe accountable to them, they must have fiscal in-


4

government subsidies to poorer areas can allevi-ate the problem but tend to be politically difficult,because the subsidies mean richer areas are sup-porting poorer areas�and they may resent doingso. Moreover, such subsidies can be counterpro-ductive if they reduce incentives to raise taxes inthe richer areas.

6. Political will is needed at the highest level tocounter political and bureaucratic resistance.While political will is important to any develop-ment goal, it is critical to democratic local gover-nance. The primary reason for failure of decen-tralization initiatives historically has been centralgovernments� unwillingness to devolve enoughpower to local governments for them to be able tofunction. Since the shift in power and authorityembodied in democratic local governance makespolitical and bureaucratic opposition almost in-evitable, much political will is almost always nec-essary to see initiatives through to fruition.

7. Bureaucrats must be held accountable toelected officials. Considerable energy must be de-voted to ensuring that civil servants are account-able to the elected officials in charge of local gov-ernments. At the same time, central and localgovernments need to give due regard to civil ser-vants� employment rights and job security.

8. Elected officials must be held accountable tolocal citizens. Public accountability can be real-ized through such mechanisms as elections, civil

society, formal grievance procedures, the media,opinion surveys, political parties, and publicmeetings. Each has proven useful in some settings,though none works everywhere nor has any alonebeen sufficient to ensure accountability. Instead,accountability calls for a combination of mecha-nisms tailored to the setting. If a particular mecha-nism seems unlikely to work�and civil societyand political parties in particular seem deficientin many settings�alternative mechanisms maybe used to support the weaker instruments or tosubstitute in part for them.

9. Local governments must deliver services. De-mocracy can help them do a better job deliveringservices but it cannot do this quickly. Performancein local government appears to correlate withprogress along a democratic path�systems withlonger, deeper experience in local democracy de-liver services more successfully and more in con-sonance with popular wishes. In countries withlittle or no experience in democracy, it is not rea-sonable to expect such mechanisms as civil soci-ety and a free press to immediately and smoothlyensure good performance from local government.It takes time to get the process right.

10. Associations of local governments can bepowerful advocates at the national level. Donorswill inevitably move from local governance toother development concerns, and sooner or laterleave the host country altogether. Under such cir-cumstances, advocacy organizations are likely to

dependence as well as political authority. It is not enough todecree that local governments be responsible for various activi-ties and to provide subsidies for them to discharge those duties(though such funding is usually critical to their success). Theymust also be able to mobilize resources locally and decide howto spend them.

5. Local government success at raising revenue can magnifyregional disparities. Richer areas will find it easier to levy taxesand fees than poorer ones, enabling them to make larger socialinvestments. This leads to better living standards for them butleaves poorer communities further and further behind. Central


be the only agencies arguing for continued atten-tion to the needs of democratic local governance.Accordingly, developing such groups as early aspossible should be a high priority.

11. DLG is a flexible strategy. There is no �onebest way� to develop local self governance; manyapproaches seem to work. Decentralization hasthe potential to succeed almost anywhere, if thecentral government gives it the necessary supportand if public accountability prevents local elitesfrom seizing most of the benefits. Those two con-ditions are tough to meet�hence the many fail-ures�but progress is not impossible, as some ofthe CDIE case studies show.

12. Previous assistance in other sectors can bea valuable base on which to build. Though notessential for success, previous external aid to lo-cal governments may provide a base of experi-ence, both individual and institutional, that cangive democratic local governance a head start.

13. �Building on the best� is both promisingand fraught with potential difficulties. Givendecentralization�s many past failures, donors� de-sire to select promising sites for pilot efforts isunderstandable. This maximizes the likelihood offinding workable approaches that can be repli-cated on a larger scale. But this strategy height-ens the chances that the seeds of democratic lo-cal governance will grow only in fertile spots, of-fering little to less promising and marginal areas.

14. Finding the appropriate balance of democ-racy (input) and public administration (output) isimportant. Some countries need more of one thanthe other, while some need to develop each in se-quence�for instance, building institutional capac-ity before improving participation and account-ability.

Recommendations

Strategically, donors should

■ Use DLG to improve representation, responsive-ness, and accountability

■ Rely on other strategies to empower marginalgroups, reduce poverty, and rectify regionalimbalances

Tactically, donors should

■ Encourage DLG as a means of empowering geo-graphically concentrated minorities

■ Support local fiscal autonomy and revenuemobilization, but also encourage central gov-ernment subsidies to poorer regions

■ Support several mechanisms of accountability( civil society, elections, formal grievance pro-cedures, media, opinion surveys, political par-ties, and public meetings)

■ Support building social capital and civil soci-ety as a long-term proposition

■ Help develop municipal lobbying associationsto advocate for local governments in the capi-tal city

■ Continuously shore up political will for DLGnationally and locally

■ Build on previous donor support for decentrali-zation

■ Strive to attain a balance between the democ-racy side and the public administration side ofDLG

References


Aasen, Berit; Odd E. Arnesen; Stein S. Eriksen;and Arne Tesli. 1997. Evaluation ofDecentralisation and Development:Decentralisation in Developing Countries�Ex-periences and Lessons Learned. Report to theDAC Ad Hoc Working Group on Evaluation ofPrograms Promoting Participatory Develop-ment and Good Governance. Oslo: Royal Min-istry of Foreign Affairs, April.

Associates in Rural Development. 1994. Decen-tralization: Finance and Management Project.Final report. Burlington, Vt.

Barkan, Joel D. 1996. �Genuine Devolution toPopular Control vs. Mere Deconcentration inAnglophone Africa.� Paper for annual meet-ing of American Political Science Association.San Francisco, August�September.

Bendix, Reinhard. 1969. Nation-Building and Citi-zenship: Studies of Our Changing Social Or-der. Garden City, N.Y.: Anchor Books.

Bidus, Mark H. 1995. �Municipal Developmentand Democracy in Central America.� Preparedby USAID/Guatemala, Central American Pro-grams. International City/County ManagementAssociation.

Binswanger, Hans; and Anwar Shah. 1994. �De-centralization, Fiscal Systems, and Rural De-velopment.� Revised proposal. Washington:World Bank, 24 June.

Blair, Harry. 1982. The Political Economy of Par-ticipation in Local Development Programs:Short-Term Impasse and Long-Term Change inSouth Asia and the United States from the 1950sto the 1970s. Monograph Series No. 4. Ithaca,N.Y.: Cornell University, Center for InternationalStudies, Rural Development Committee.

�� . 1985. �Participation, Public Policy, Po-litical Economy and Development in Bang-ladesh, 1958�1985.� World Development 13,12 (December): 1231�47.

��, ed. 1989. Can Rural Development Be Fi-nanced from Below? Local Resource Mobiliza-tion in Bangladesh. Dhaka: University PressLimited.

��. 1995. �Assessing Democratic Decentrali-zation: A CDIE Concept Paper.� Washington:USAID.

��. 1997. Democratic Local Governance inBolivia. CDIE Impact Evaluation No. 3. Wash-ington: USAID.

Blair, Harry; and Gary Hansen. 1994. WeighingIn on the Scales of Justice: Strategic Ap-proaches for Donor-Supported Rule of Law Pro-grams. CDIE Program and Operations Assess-ment Report No. 7. Washington: USAID.

Bowles, Chester A. 1954. Ambassador�s Report.New York: Harper and Brothers.

Brillantes, Alex B. Jr. 1994. �Redemocratizationand Decentralization in the Philippines: TheIncreasing Leadership Roles of NGOs.�International Review of Administrative Sci-ences 60, 4 (December): 575�86.

Calavan, Michael. 1996. �Building DemocraticLocal Governance: A Development Challengefor the New Millennium,� draft manuscript.Washington.


Cheema, G. Shabbir; and Dennis A. Rondinelli,eds. 1983. Decentralization and Development:Policy Implementation in Developing Coun-tries. Beverly Hills: Sage Publications.

Cohen, John M.; and Stephen B. Peterson. 1995.Administrative Decentralization Strategies forthe 1990s and Beyond. Governance and Pub-lic Administration Branch, Division for PublicAdministration and Management Develop-ment, Department for Development Supportand Management Services. New York: UnitedNations Secretariat.

�� . 1996. �Methodological Issues in theAnalysis of Decentralization,� DevelopmentDiscussion Paper No. 555. Cambridge, Mass.:Harvard Institute for International Develop-ment.

�� . 1997. �Administrative Decentralization:A New Framework for Improved Governance,Accountability and Performance.� Develop-ment Discussion Paper No. 582. Cambridge,Mass.: Harvard Institute for International De-velopment.

Conyers, Diana. 1983. �Decentralization: The Lat-est Fashion in Development Administration?�Public Administration and Development 3, 2(April�June): 97�109.

�� . 1984. �Decentralization and Develop-ment: A Review of the Literature,� Public Ad-ministration and Development 4, 2 (April�June): 187�197.

�� . 1986. �Future Directions in Develop-ment Studies: The Case of Decentralization.�World Development 14, 5 (May): 593�603.

Crook, Richard, and James Manor. 1994. �Enhanc-ing Participation and Institutional Performance:Democratic Decentralization in South Asia andWest Africa.� Report to ESCOR. Overseas Devel-


opment Administration.

DAC (Development Assistance Committee of theOrganization for Economic Cooperation and De-velopment). 1993. DAC Orientations on Par-ticipatory Development and Good Governance.Paper OECD/GD (93) 191. Paris: OECD.

��. 1995. Participatory Government andGood Governance. Development Cooperationand Guidelines Series. Paris: OECD.

Department of Interior and Local Government(Philippines). 1992. Local Government Codeof 1991: Republic Act No. 7160. Manila.

Esman, Milton J.; and Norman T. Uphoff. 1984.Local Organizations: Intermediaries in RuralDevelopment. Ithaca, N.Y.: Cornell UniversityPress.

Freedom House. 1997. Freedom in the World: TheAnnual Survey of Political Rights and Civil Lib-erties. New York.

Garnier, Maurice; David Green; Elinor Ostrom;James T. Thompson; and Tina West. 1991. �De-centralization: Improving Governance in Sub-Saharan Africa,� Draft synthesis report; DFMProject.

Gephart, David; Louis Siegel; Ed Connerly; andLarry Schroeder. 1989. Bangladesh: RoadsMaintenance and Improvement Project Evalu-ation. Burlington, Vt.: Associates in Rural De-velopment.

Greenberg, Edward S.; and Benjamin I. Page.1993. The Struggle for Democracy. New York:HarperCollins College Publishers.

Hansen, Gary. 1996. Constituencies for Reform:Strategic Approaches for Donor-SupportedCivic Advocacy Programs. USAID Program andOperations Assessment Report. Washington:USAID/CDIE.

Inbanathan, Anand. 1992. The New PanchayatiRaj in Karnataka: An Evaluation. New Delhi:Institute of Social Sciences.

Inbanathan, Anand; T. A. Bhagyalakshmi; and R.Doraiswamy. 1997. �Democratic Local Gover-nance in Karnataka State, India.� Paper com-missioned by CDIE.

Japan International Cooperation Agency. 1997.The Issuewise Study Committee for Japan�s De-velopment Assistance. �Local Development andthe Role of Government.� Tokyo.

Korten, David C. 1990. Getting to the 21st Cen-tury: Voluntary Action and the Global Agenda.West Hartford, Conn.: Kumarian Press.

Lippman, Hal; and Harry Blair. 1997. DemocraticLocal Governance in Ukraine. CDIE ImpactEvaluation No. 2. Washington: USAID.

Lippman, Hal; and Jan Emmert. 1997. AssistingLegislatures in Developing Countries: A Frame-work for Program Planning and Implementa-tion. CDIE Program and Operations AssessmentReport No. 20. Washington: USAID.

Lippman, Hal; and Joel Jutkowitz. 1997. Demo-cratic Local Governance in the Philippines.CDIE Impact Evaluation No. 1. Washington:USAID.

Lippman, Hal; and Barbara Lewis. 1998. Demo-cratic Decentralization in Mali: A Work inPro-gress. CDIE Impact Evaluation Report No.5. Washington: USAID.

Lippman, Hal; and Patrick Pranke. 1998. Demo-cratic Local Governance in Honduras. CDIEImpact Evaluation No. 4. Washington:USAID.


Lowi, Theodore. 1978. The End of Liberalism, 2nded. New York: W. W. Norton.

Manor, James. 1995. �Democratic Decentraliza-tion in Africa and Asia.� IDS Bulletin 26, 2(April): 81�88.

�� . 1996. �Political Economy of Decentrali-zation.� Draft paper for Workshop on Decen-tralization, Fiscal Systems, and Rural Develop-ment, sponsored by the World Bank, HarpersFerry, W. Va., 13�16 May. Revised and ex-panded, 1997.

Myrdal, Gunnar. 1968. Asian Drama: An InquiryInto the Poverty of Nations. New York: Pantheon.

Ng�ethe, Njuguna. 1994. The Politics of Democ-ratization Through Decentralization in Kenya:Policy and Practice With Emphasis on the Dis-trict Focus for Rural Development. Paper forConference on Democratic Decentralization inAfrica and Asia, Institute of CommonwealthStudies, University of London,21�23 September.

Nicholson, Norman K. 1973. Panchayati Raj,Rural Development and the Political Economyof Village India. Occasional Papers No. 1.Ithaca: Cornell University, Center for Interna-tional Studies, Rural DevelopmentCommittee.

Nickson, R. Andrew. 1995. Local Government inLatin America. Boulder, Colo.: Lynne Reinner.

Olowu, Dele. 1994. Decentralization, Democrati-zation and Development in Africa. Paper forConference on Democratic Decentralization inAfrica and Asia, Institute of CommonwealthStudies, University of London,

21�23 September.

Organization for Economic Cooperation and De-velopment. 1997. Final Report of the DAC AdHoc Working Group on Participatory Devel-opment and Good Governance, Parts I and II.Paris: OECD.

Parker, Andrew N. 1995. Decentralization: TheWay Forward for Rural Development. PolicyResearch Working Paper #1475. Washington:World Bank, Agricultural and Natural Re-sources Department.

Prud�homme, Rémy. 1994. On the Dangers ofDecentralization. Policy Research Working Pa-per 1252, Washington: World Bank.

Putnam, Robert D. 1993. Making DemocracyWork: Civic Traditions in Modern Italy.Princeton, N.J.: Princeton University Press.

Rajas Ortuste, Gonzalo; and Luis VerdesotoCustode. 1997. La Participación ComoReforma de la Política: Evidencias de unaCultura Democrática Boliviana. La Paz:Ministerio de Desarrollo Humano, SecretaríaNacional de Participación Popular.

Rondinelli, Dennis. 1981. �Government Decen-tralization in Comparative Perspective: Theoryand Practice in Developing Countries.� Inter-national Review of Administrative Sciences 47,(2): 133�145.

Rondinelli, Dennis; John R. Nellis; and G. ShabbirCambia. 1984. �Decentralization in Develop-ing Countries: A Review of Recent Experience.�World Bank Staff Working Papers No. 581.Washington: World Bank.

Rondinelli, Dennis; James S. McCullough; andRonald W. Johnson. 1989. �Analysing Decen-tralization Policies in Developing Countries: APolitical�Economy Framework.� Developmentand Change 20, 1 (January): 57�87.


Rood, Steven. 1996. �Researching DecentralizedGovernance.� Paper for National Conference,Philippine Political Science Association,Lingayen, Pangasinan, 9�11 May.

�� . 1997. �Democratic Decentralization inthe Philippines.� Paper for Annual Meeting ofthe Association for Asian Studies. Chicago, 13�16 March.

Royal Norwegian Ministry of Foreign Affairs.1994. Support for Democratic Development.2nd edition. Oslo.

Schroeder, Larry; and James Wozny. 1987.Financing Local Panchayats in Nepal. Syracuse,N.Y.: Syracuse University, Maxwell School, Met-ropolitan Studies Program.

Shah, Anwar. 1997. �Fostering Responsive andAccountable Governance: Lessons from Decen-tralization Experience.� Draft paper. Washing-ton: World Bank.

Slater, Richard. 1989. �Democratic Decentraliza-tion or Political Consolidation: The Case ofLocal Government Reform in Karnataka.� Pub-lic Administration and Development 9, 2(April�May): 147�157.

�� . 1994. �Strengthening Local Governmentin Sri Lanka.� Paper for Conference on Demo-cratic Decentralization in Africa and Asia, In-stitute of Commonwealth Studies, Universityof London, 21�23 September.

Smith, Brian. 1985. Decentralization: The Terri-torial Dimension of the State. London: GeorgeAllen and Unwin.

Smoke, Paul J. 1994. Local Government Financein Developing Countries: The Case of Kenya.Nairobi: Oxford University Press.

Timberman, David G. 1990. A Changeless Land:Continuity and Change in Philippine Politics.

Singapore: Institute of Southeast AsianStudies.

United Nations. 1961. Technical Assistance Pro-gram. A Handbook of Public Administration:Current Concepts and Practice with Special Ref-erence to Developing Countries. New York:United Nations, Department of Economic andSocial Affairs.

��. 1989. Department of International Eco-nomic and Social Affairs. Prospects of WorldUrbanization, 1988. Population Studies No.112. New York.

United Nations Capital Development Fund. 1995.Poverty Reduction, Participation and LocalGovernance: The Role for UNCDF. New York.

United Nations Development Program. 1996,1997. Human Development Report. New York:Oxford University Press for UNDP.

��. 1997a. Governance for Sustainable Hu-man Development: A UNDP Policy Document.New York: UNDP.

��. 1997b. Decentralized GovernanceProgramme: Strengthening Capacity forP e o p l e -Centered Development. Draft Management De-velopment and Governance Division, Bureaufor Policy and Program Support. New York.

��. 1997c. The Global Research Frameworkof the Decentralized Governance Programme.New York.

USAID. 1994. Strategies for Sustainable Develop-ment. Washington.

��. 1995. �Core Report of the New Partner-ships Initiative.� Program and Policy Coordi-nation/SA, 21 July.

��. 1997a. NPI [New Partnerships Initiative]


Resource Guide: A Strategic Approach to De-velopment Partnering. Report of the NPILearning Team. Washington.

�� . 1997b. Center for Democracy and Gov-ernance, Strategic Plan, 1997�2002. Wash-ington.

Van Sant, Jerry; Harry Blair; Henedina Razon-Abad; and Todd Amani. 1998. Impact Assess-ment: Governance and Local DemocracyProject. Report prepared for USAID/Manila. Re-search Triangle Park, N.C. Research TriangleInstitute, February.

Walker, David B. 1991. �Decentralization: RecentTrends and Prospects from a Comparative Gov-ernmental Perspective.� International Reviewof Administrative Sciences 57, 1 (March): 113�129.

Webster, N. 1990. Panchayati Raj and the Decen-tralization of Development Planning in WestBengal: A Case Study. CDR Project Paper 90.7.Copenhagen: Center for Development Research.

Westergaard, Kirsten; and Mustafa Alam. 1995.�Local Government in Bangladesh: Past Expe-riences and Yet Another Try.� World Develop-ment 23, 4 (April): 679�90.

Wilson, William Julius. 1987. The Truly Disad-vantaged: The Inner City, the Underclass, andPublic Policy. Chicago: University of ChicagoPress.

��. 1996. When Work Disappears: The Worldof the New Urban Poor. New York: AlfredKnopf.

World Bank. 1994. Governance: The World Bank�sExperience. Washington.

��. 1997. World Development Report 1997:The State in a Changing World. Washington.

Wunsch, James S., and Dele Olowu. 1996�97.�Regime Transformation from Below: Decen-tralization, Local Governance and DemocraticReform in Africa.� Studies in Comparative In-ternational Development 31, 4 (winter): 66�82.



Date post:	03-Feb-2016
Category:	Documents
Upload:	dejan-seslija
View:	4 times
Download:	0 times