World Bank

Public Disclosure Authorized
______FWes 1122
Polly Roerch
WORKING PAPERS
PublicEconomki
PolicyResearch
Department
TheWorldBank
March1993
WPS 1122
Fiscal Decentralization
and Intergovernmental
Relations
in Transition Economies
Towarda SystemicFrameworkof Analysis
RichardBird
and
ChristineWallich
Designinga well-functioningintergovernmentalfiscal systemis

essentialto the successof all the transitionaleconomies'major
reformgoals:privatization,macroeconomicstability,moreeffi-
cientperformanceandeconomicgrowth,andan adequatesocial
safety net.
PoCcyRaearWoingPpcsdhsninatchehfndingpofwoinpogeaanduagctheexchangeofideuamongBankstaff*nd
aflod inrcdinde copmatisauLbespapeis. dibutydbytbeRemhAdviy Stff,caythenamesoftheaugors,re&ct
enlytherviews,andshWdd and.
dhBoardcfocortngly.of h ngsia nors, any ofuha emtbeh cautr n' ow.eyshould
na be atiutcd to fthWodd Bank,a Bcud of Di,cw itsu umaect, or any of ihsb
membrcuin
PolicyR oroh|
PublicEconomics
WPS1122
This paper - a joint productof the PublicEconomicsDivision,PolicyResearch Department,and the

InfrastructureDivision,TechnicalDepartment,Europeand CentralAsia,andMiddleEastandNorthAfica
Regions- is part of a larger effortin bothdepartmentsto analyzeissuesrelatingto decentralization,and
in particular,the role of subnationalgovernmentsin the reformpmcess. Copiesof the paperare available
free from the WorldBank, 1818H StreetNW, Washington,DC 20433.PleasecontactBonniePacheco,
room HI 1-065,extension37033 (March 1993,85 pages).
The decentralizationof governmentin Eastern subnationalgovernmentsmay look to "coping
Europe representsa reactionboth from below (to mechanisms"such as holdingonto their enter-
tight centralpoliticalcontrol) and from above(to prises (whichprovidevital social services),
privatizethe economyand relievethe central developingextrabudgetaryrevenues,or borrow-
government'sfiscal stress). ing. Thesecopingmechanismsthreatenprivat-
ization,reducebudgetarytransparency,and
In all transitionaleconomies,the developing impedestabilizationpolicies.
structureof intergovernmentalrelationsis
intimatelyrelatedto suchcritical policy issuesas Bird and Wallichdescribethe risksto
privaization, stabilizaticl, and the socialsafety piivatization,to macroeconomicstability,and to
net. In the fiscal sphere,tax reform,deficit an adequatesocialsafetynet that presentpolicies
control,and intergovernmentalfinanceare a towardlocal governmentmay imply.Its themes
tripod.Unlesseach leg is set up properly,the arethat the subnationalsectorneeds to be more
wholestructurecouldcollapse. realisticallyfactoredinto nationalplans - and
that subnationalexpendituresbe more clearly
The presentstrategyof devolvingexpendi- assignedand revenueneeds more realistica9;y
turesdownwardwhile holdingbackon revenue assessed.S&chassessmentsare likely tv ac-
flowsand transfersto balancethe centralbudget knowledgea larger spherefor subrational
is unlikelyto succeedfor more than a year or governmentsand the need for accec!tm
two at bestL robust revenue sources. Giving local government
a share in the personalincometax is one possible
Net spendingreductionsat the subnational and perhapsdesirable approachto meetingthese
level may be difficultto achieve.From 10 to 40 revenueneeds.
percentof outlaysgo to the subnationalsector,
and in manycountrieslocal governmentsprovide Carefulattentionneeds to be paid to the
muchof the social safetythat makesthe pain of designand implementationof the
the economictransitionpoliticallytolerable. intergovernmentalfiscaltransferslikely to
And, most housingand many enterpriseshave remainprominentfeaturesof the
been shiftedto local ownership,with the mainte- intergovernmentallandscapefor years to come.
nance and subsidycost this implies.Sincethe Cautionis also neededon borrowingby
revenuesourcesassignedto local governments subnationalgovernment.Consolidatingand
cannot financeexpectedlevels of local activity, integradngextrabudgetaryfundsat the
the result of shiftingspendingdownwardis subnational(and national)levelsis crucialto
likely to be strongdemandsfor increased,rather enhancedbudgetarytransparencyand
than decreased,transfers.Altematively, macrostability.
The PolicyResearchWorkingPaperSeriesdissemirsates the fmdingsof workunderway in theBank. Anobjecii vof theseries

and
is to get these findingsout quickly, even if presentationsare less thanfully polished.The findings, interprAtations,
conclusionsin these papersdo not necessarilyrexesent officialBankpolicy.
Producedby the Policy ResearchDisseninationCenter

FISCAL DECENTRALIZATION AND INTERGOVERNMENTALRELATIONS IN TRANSITION
ECONOMIES:
TOWARD A SYSTEMIC FRAMEWORK OF ANALYSIS
by
Richard Bird
University of Toronto
and
Christine Wallich
Country Economics Deparhnent
Public Economics Division
The World Bank
Table of ConLonts
1. INTRODUCTIONAND SUMMARY . ......................................... 1
Why 8Local1Finance Matters 1...................

Traditionalvs Systmic Frameworks. ....................................... 2
Organizationof the Paper .............................................. S
Political/OverallStructural Reform Context ................................... 6
Borrowing .................................................. 8
Extrabudgetar Funds ................................................ 8
Infitutional Framework ............................................... 8
IN PROCESS:TRENDSAND THEMES ........................

II. DECENTRAULZATION 9
A Time of Transformation . ............................................ 9

The End of Centra Planning: Filling The Abhorted Vacuum ........................ 9
PoliticalD entatization. ............................................. 11'
The Size and Scope of Government: How Many Levels? .......................... 15
Size of Government Units: Efficiencyand Other Concens ......................... 17
Supporting Subnational Government: The Institutional Setting for Subnational Governmental
Reform ......................... i 8.............
The ChangingRole of Govermment:From Ownershipto Servic - Provision .............. 19
SimultaneousNational and IntegovernmentalFiscd Reforms. ....................... 23
Stablizadtio: The MacroeconomicContextfor D sentrizaton ...................... 24
II. WHO DOES WHAT? THE ASSIGNMENTOF EXPENDrrURES ....................... 29
The EfficiencyCase for DecentralizedLoca GovermenL .......................... 29

ExpenditureAsignment in Pratice ........................................ 29
Local Public Enterprise Ownehip ........................................ 35
Provisionof Social Services: Shiftingthe Safety Net Downstain .................... 37
ExpenditureAssignmentin Principle....................................... 37
Extabudgetary Funds ................................................ 39
IV. FINANCINGLOCAL PUBUC SERVICESTHROUGHUSER CHARGES ................. 40
V. THE CHOICEOF SUBNATIONALTAXES..................................... 42
Characteristicsof a Good Local Tax ....................................... 46

Local Propet Taxes ................................................ 47
Own SubnationalTaxes versus Shared NationalTaxes ............................ 48
TRANSFERS ..........................
VI. THE DESIGNOF INTERGOVERNMENTAL 50
Closing the Fiscal Gap ................................................ 50

Equalization ...................................................... 50
Fiscal Effort ...................................................... 52
MatchingGrants ................................................... 53
Other Objectives of Transfers ............................................ 54
A Case for Conditionality?............................................. 55
- ii -
Determiningthe Size of the Grant . . . . . . . . . . . . . . . . . . . ... .. ........ . ....... 56

Distributionof the Grant............................................... 56
MeasuringCapacity and ExpenditureNeed.................................... 57
VII. BORROWING ...................................................... 58
Past Borrowing Pr actic e s

............................................... 59
Sourcesof Borrowing................................................. 60
VIII. PRIVATIZATIONAND OWNERSHIP ....................................... 61
AssetTansfers ............................. 61
Privatizationvs Entrepreneuriaism .................................. 62
Timing of Asset Sales: Sales vs Give-aways .................................. 63
Revenuesfrom Privatization.................................. 64
IX. CONCLUSIONS.................................. 64
REFERENCES ................. 84
List of Tables
Table I Trends in Decentralization:GovernmentOrganization. 16

Table 2 Decntrlization: Scale and Size of Government ...............................
Table 3 Reform and the Size of the GovernmentSector .22
Table 4 MacroeconomicIndicatorsi.2
Table 5 Tax Administrationin TransitionSocialistEconomies ............................ 28
Tabls 6 ExpenditWeAssignmentto SubnationalAuthorities .............................. 30
Tanle 7 A PossibleExpenditureAssignment....................................... 38
Table 8 RevenueAssignmentsto SubnationalAuthorities ..............................44
Table 9 Structureof SubnationalGovermmentFinance . ................................. 1
Table 10 SubnationalGovernmentBorrowing ....................................... 59
Table 11 PrivadzatioaRevenues:Ownershipand Disposition .. 62
Acknow_edeMents
We wouldlike to thankthose who commentedon thispaper, or whootherwisesharedtheirinsightson

subnatlonalfinance in the variouscountries,includingFelixJacob;Zelko Bogetdc,RobertVan Pulley, Thn
Condon,GaborPeteri, ShankarAcharya,Anwar Shah, Heng-fiu Zou, Tony Pellegrin4Jim Hicks, Luca
Barbone, 7ThanosCatsambas and AmareshaBagchl. Researchassistanceof Rltu Nayyar is gratefully
acknowledged. SusheelaJonnakutyprovidedexpertwordprcessing. 7Thisis thefirstin a seriesofpapersunder
RPO BB67770M.
FISCAL DECENTRALIZATIONAND INTERGOVERNMENTAL
RELATIONSIN TRANSION
ECONOMIES:
TOWARD A SYSTEMICFRAMEWORKOF ANALYSIS
... A chang of governmentis not a chang of system, merely one of the politicdl conditionsfor
it. Tho change of system is a historical process that seer- likselyto require a long period of
time.&'
. INTRODUCTIONAND SUMMARY
Why 'Local" Fimnce Matters.
1. An importantand unduly neglected aspect of the process of transition in Eastern Europe is the extensive
decentralbiation,both politicaland fiscal, that is taiing place in many of the countries newlyemerging from behind
the socialist veil. Decentralizationrepresentsboth a reaction ftom below to the previously tight political control
from the center and an attemptfrom above to further the privatzation of the economy and to relieve the strained
fiscal situation of the central govermnment.The aspirations of, and the role of, local government are becommg
stronger in most of the transitionaleconomies.
2. The on-going reforms in subnationalfinance in these economiesare much more important than seems
generally to be recognied. In some countries in transition, the subnationalgovemments (broadly defined) are
quantitativelysignificant, accountingfur a large, and often gromwing,share of public sector activity. In aU the
transitionaleconomies,the developingstructure of intergovenmentalrelationsis intimatelyrelated to such critical
policy issues as privaizaon, the social safety net, and stabiliatioL Within the fiscal sphero, tax reform, deficit
control, and intergovenmentalfnce constitutea tripod: unless each leg is set up properly, the entire structure
may collapse.
3. Our assertion is that the designof a weUfumctioningintergovernmentalfiscal system is key to almost all
of the transition economies' major reform goals. Intergovernmentalfinances wil have a major impact on the
efficiencywith which the transitioneconomiesperform; on their macroeconomicstability;on the social safety net
and on the successof their privatzation policies.
4. Fiscal balance at the subnadonal level is key to macroeconomicstabiliqy. Given the paramountcy of
stabilizationgoals in almost all of the countries, the central governmentshave used all the tools in their meager
arsenal to bring stability to the macroeconomyY One such tool has been to shift the deficit down, ensuring an
adequaterevenueflow to the center - oither by shifting expenditureresponsibilitiesdown or by retainingadditional
revenues, even where budgets at the subnatonal level are severely iupaiued Hungary and Russia come to mind,
but there are others.
5. The resultsof this policy, ironically,could contributeto further instability: the budget constraint in these
countries is still so soft that subnaional governmentsmay respondby accumulatingarrears, borrowing(strong-
arming their enterpises to financepublic expenditures- sice enterprises have easier access to bank credit than
do the governmentsthemselves)or developingextrabudgetarysourcesof revenue. Pushing the deficit down may
not reduce it, but may merely repress it. Worse, some of these 'coping mechanisms' also reduce the transparency
of budgetary operations, and make the pursuit of appropriatenational budgetarypolicies even more difficult than
it is now.
-JJanosKomai, T'heSocialistSystem; (PrincetonUniversityPress, 1992), p.577.
Zlbe terms subnational and 'local' are used more or less interchangeably,unless speakingof a three tier, or
federal system, in whichcase subnationalrefers to all levels belowthe federal, and local refers to the lowest level.
6. A well-fimctiong - responsibiity for which has bee trafered 'downstairs' in a
number of countries -cannot be effetively caried out by an undefund subuationalsector. Where the safatynot
is cridcal for the success of Ihotransition to a market economy, nd it therfore a national policy objctive, -as
in Russa, and ,tguably, in Polundand other countrie - the burden of finaing the safety not should not be on
local governments,even though, being closer to tho people, subnationl gove ormnts migt administer it.
7. From an atlocativeprpctdve, the system of intove n fiscal rebltionswill also dtermine in a

major way the officiencywith which thb eoanomies perform, and their fu omic o. Locl
governmentownesip of enteprs - and local ir in their revenus - will inevitablyencourge domestic
protectionism (e.g. 'domestic purch rqimes and inter-jurisdictionaltrade bariers to protect local
monopolies)harmful to economicgrwth
8. Finaly, becaus some countries -notably Ruia- ar, so to speak, 'nation-building,' with significant
centrifual fores emanatingfrom the ubnatonal level, inteovermentl finane is also crueil to futurenatioal
cohesion. In Russia it may be crucial even to the contied existence of the Federadon. Even where national
cohesion is not at risk, demands for greatr 'participatory democracy' have heightened demands from the
subnationallevel. The strong desire for local politieal control and unfettered local autonomyis a natural outcome
of the end of central plmning. Combinedwith an inadequate fiscal base, however, loal aspirations and fiscal
capabilitiesare increasinglyat odd., which is giving rise to growing tensionsin a number of countries.
9. Even gvimhi2n is complicatedby the current environment.Followingthe transfer in many countries

of a significantnumber of anterpie to ownership, responsibilityfor privatizationnow also rests with
1matonal
subnationalgovenments. Their interess hero ar unclear at bese they still rely on ethirI enteprises for profits
and tax revenues, ad if privzd, subational govenments will hav to assume the substantial'non-production'
expenditues - for nurei_e, housing, schoob, etc.- tht will e spun off by the enterpris as they privatize. A
system of inegonmeta finacs tht doe not accommoda thes shifts in expenditureresponsibilitiesand
provide coreponding revenues, will impedothe proco of privatization.
10. In sum, intergovermentl fiscal rlations - far from being merely a 'local matter' - are key to simost
of a8lof the transitioneconomies' reform goals.
Traditionalvs SystomicFrumewors
11. The traditional anaysis of intergovenmenta finance examines the fiscal functionsof local and central
govermmentsin terms of their respectiveroles and responsibilidesfor stabilization,incomedistribution,expenditure
provision, the appropriate assignment of tax functions, and the design of a transfer system that provides
appropriateincentives. The 'benefit model' of service provision suggest that local governments- whose role in
this analysis is essentially that of service provider - should be financedto the extent possibleby charging for the
servicesthey provide, with local taxes makingup the remaining gap, supplementedas appropriateby trsasfers and,
in limited degree, borrwing.
12. This perspectiveis of courseimportantin the trasitional economiesalso. Butit missesseveralkey features
of the roles, responsibilitiesand economicfunctionsof subnationalgovemmentsin the former socialisteconomies.
First, the emphasisin traditionalanalysis on the local goverment as a serviceprovider ignores its importantroles
as produr and as wngr,as wel as the complicatedrelationshipsbetweenenterprisesand local governmentin most
transitionaleccnomios. The importuntrole of local governmentsas owners means thst they have a major role to
play, either as potential impedimentsto, or supporters of, Drivatization. Moreover, the asset stock recuntly
trnsferred to them in the decentralition process representsa potential source of revenue(or, in some instances,
of losses). The interactionof subuationalfinanceand privatizadonthus merits careful attention in the transitional
economies.
-3-
13. Second, th traditionalapproachignors th shrining role of govenment. in most t itional conomies,

govenment, both local avd cetral, playeda major producdonrole and wa the major invoeor in the economy,and
tae expendite side of the budget is clutered with expditur - not only sbidies, but direct investmnt,
inventory finance and wages - which in a market conomy, are not the responsibility of govemment.
Unfortunately,in all des countries govemmentreve is cclining mora npidly tha governmnts moabl, to
dives themsevas of thes expenditu responsibiities, thus ontributinqto u1bahinitproblems. One repon
in may countrie ha boen to try to shift the deficitdownwans by mang local govemnts responsiblefor morm
expenditure, while smtanously reducingcentrl tansfers to the onatonal sctor.
14. The obviousneed for flexibilityin today's apidly changingenvirment bas led many centralgovenmts
to attemptto preservosom dogreesof freedomby continuingwith the 'negotiated' tax sharng systomsof the pat.
But this approch is less and lea acceptablein countrieswhere domandsfor fair" treatmnt and ec2alizationare
strong and where local govemments are seeking gater autonomy. It is also incompatible with the efficient
provisionof local public services. It is theeforo especiallyimportat to replacethe presnt 'bargained' outcomes
with an intergovrnmetal fiscalframework- of transfers, of local revenues,and of local expenditureresponisibilities
- that is both firm enough to serve as a basis for actionand flexible enough to be compatiblewith the on-going
structual changes and reform.
15. The present sftrgy of devolvingexuendituresdownward while holding back on tansfers is unlily to
prove successfl for more than a year or two at best. Indeed, net expenditure eduction at the subnatonal level
may prove very difficult to achieve. The subnationalsector is significantin expendituroterms in many trasition
economies-in Rusia, and the CSFR, it accountsfor over 40% of total outlays. Elsewhr they account for 10-
30% of total outlays. At pesent, for example,state enterprisesprovide a wide range of social sector outlays: with
privatiznton, manyof these outlayswill have to be takenover by local governments. Moreover, in most countries,
the local governmentsector plays a critical role in tenn of providing a 'social sae net' which makes the pain
of the wrenching economic transition more politicaliy tolerable. Since the trvenue sourcs assigned to local
governmentssimplycannot financethe levelof expectedlocal activityin th i_mmeditofture, the result of shifting
expendituresdownwardis likelyto be strong demandsfor increasd, rather than reduced, transfers. In sum, as and
when expenditureassignmentsbecome cleaer, and decentraliation develops further, the likelihoodis that the
subnationalshare in other transitioneconomieswiUlalso grow. This means that tho intergovernmentalsystem wil
have to be sensitive to these developments:Notable across the transition economs is the fact that there has been
a general absenceof concrete empiricalestli -leg of 'correspondencewbetweentax and expenditureassignmentsand
need for transfers. Legislation on intergovenmental finance relations has typicaUy been developed with no
quantitativeassssment of its implications.
16. Despiteall the talk about *autonomy'and fiscal 'independenceo,few of the countriessurveyedoffer their
subnationalgovemmentsmuch in the way of fiscaldiscretionon the expenditureside. Indeed,most recentlegislation
has focussedon codifyingthe revenueside of the equation-before decidingwhat expenditureresponsibilitiesof
subnationalgovernmentsshould be, putting the cart before the horse, so to speak. Expenditure 'autonomy' is
enshrinedin legislation,but not really effective:there are still centrl 'norms' which govern allowableexpenditures
in virtually every sector, and centml spendingmandatesremain in place.
17. The new subnaional governmentswiUlalso needsubstantialhelp and guidancein developingadequatelqol

revnuo systems. The local *autonomy' that has been enshrined in recent legislationhas often meant separating
the spheres of responsibilityto the greatest extent possible, so that each level of governmentreigns in its own
sphere: in revenue terms, this translates into a pm-dispositiontowards *tax assignment models, and a goal of
reduced*transfordependance' on the center. The fact that in the pre-reform system, most subnationalgovernments
receivedrevenues from 'their' enterpises has led to a strong bias towards derivation-basedsharing models in
whichrevenues go to the jurisdiction in which they are collected.
18. It is critical that subnationalrevenueneeds and possibilitiesbe realisticallyfactoredinto nationalplans for

tax reformLFailure to look at intergovemmentaldimensionsof national tax policy changes can lead to problems.
-4-
In Russia and Hungary, a share of major natonal taxes has boen assigned to the subnationalIlvel, and recent
national tax policy changes have wrought havoc with the aggrpto subnationalrevenue base, with vecy eaic
effects across subnationaljursdictions. Secod, any thought of aasidg. ut', when the tax itself is undergoing
continuousdefinition,is problemical. In m-y transitioneconomie,however, local governmentshave comeout
stronglyin favorof revenue assigmet under which theygot 'all' the 'renuesfrom vtheirtaxes. Thisseeming
revenueindepndence leaves them with sgnifict residualvulnbility to tax policy changs, however,a fact that
seems genrally not to have ben recowzed. If this vunbity wore to take ito accout, as it dsould be,
national fiscal policy reforms would of course be constaned. Rcent tax reforms at the national level,
administrative weans and other factors havo meant significat rvenue vowlatiliy,with significantover and
undershootingof budgetay tagets. This situaon obviouslyhasimplicationsfor the desirabilityof 'tax assigment'
modelsper s, andthe designof sharing/transforsystms.
19. In most countries, for example,the pndcipal locaI" tax is some form of property tax: but in many cases,
nationalgovernmentshave hamstrungthe revenuepotential of this tax by grting exemptions to newly privatized
land, housing, and enterprises.In other instances,local governmentsare supposd to recive a share of certain
national taxes, often on a derivationbasis (e.g. origin, or residence):but such povisions raise both techcal and
allocativeproblems,in additionto biasing nationaltax policy decisionsin often undesirableways. Even whenthe
local tax base is feasibleandadequatein principle,in practiceit will gerally take some years before the system
can be expected to produce sufficient revenuesto meetperceivedlocw,;needs.
20. Despiteal the talkabout *autonomy-and fiscal independence, none of the countriessurveyedoffer their
subnationalgovemvents any fiscaldiscretion,even over the muior taxes they ane wssigped:rates and basescontinue
to be set by the center, in the context of the unitary tax sysem. lho paucity of locally-controlledtax resourees
in
mosttransitionalcountries,whencombiaedwith the universl reluctnce of politiciansto tax constituentstoodirectly
and openly, makesit inevitablethat hard-pressedlocal governmentswiUturn to other avenuesfor revenue:theywill
wmand increasedtransfers, theywill try to borrow, and they will try to exploittGthe full the new assets theynave
acquiredas part of the decentralization-privadzation process.
21. The level, designand effects of interaovemmentalfiscal transfer. thus constitute key elements in the
emergingintergovemmentaland local finance system. of the tnsitonal economies.There is a general absence
of 'formula'- basedtrasfers. These remain negotiatedin aggregateand by individuallocalities. Criticaldecisions
must be madewith respectto the overal size of suchtransfers(the 'distributable pool') as wel as with respectto
the distribution formulas to be employed, taldng into account both the severe fiscal pressures on the central
governmentandthe vital tasks to be performedby the local govemnts in the emergingnew strcture of the public
sector. Practicalmeasuresof *taxcapacity, *expenditureneed' and perhaps fiscal effort' need to be developed
for use in such formulas. Institutionsmust be developed at both the central and local levels to ensurethat transfers
(or subsidizedcredits) are put to the best possibleuse: such institutionsmay include, for example, at the center,
agenciesconcernedwith monitoringthe perfonrance of local govemmentsandwith providing them with needed
technicalassistanceand supportand, at the subnationallevel, various intermediarybodies, particularlywith respect
to the many very small municipalitiesthat have been createdin most countries.
22. As noted above, careful attention has to be paid to the design and implementationof the intergovemmental
fiscaltransfer. that seem likely to remain prominentfeaure of the fiscal landscapein most countriesfor years to
come. Cautionis even more necessary with respectto borrmin by subnationalgoverments. A number of
countries in transition have granted virtually unlimited bonowing powers to local governmentswithout taldng
adequately into account the macroeconomicnecessity of monitoring and controllingsuch bonrwing. As yet,
however, the relativelyundevelopedstate of financialinstitutionsin most countries makes abuse of the borrowing
power more a potential than a real problem.
23. The same is not true of the potentialmisuse of 'locallv owned" state assets, unfortunately,since in every
country such misuse is already visible. Such local assets may represent a windfall - at least in those cases where
they are not accompaniedby centmlpolicies(e.g. rent or price control)that turn them into a liability. Butthey may
- 5-
also often be misused,whetherthey are privadzedor not. A local heating plant that is privatized,for example,but
remains a (local) wraturalmonopoly obviously raises regulatoryproblems: who deals with these problems, and
how? Whatshould be done with the proceedsof asst sales? Under what (very limited) circumstancesshould such
proceeds be used to finance current local expeitur ? If an asset is not privatzed, but is used to geneote on-
going revenues for the local govemment(and perhp some forign partnr, with the municipality using its
ownershipof land or assets as its contributionto equity), the temptationto sustain and sttengthen any monopoly
position is obviouslyagain a problem, as is that of regulation. Even more damaging is the high pwbability that
more speculativeventuu - examplesincludetouristlodge andbakeries - are likelyto fail. Equally damagingmay
be the dissipationof scare local managerialskills on trying to squeewzmoneyout of such entrepreneurialgambits
rather than focusingon tho more essenal task of trying to provide local public services efficiently.
24. A specialproblem arises with reap' t to the prices cbrged by those local public enterprises to be retained
in tho public sector. In the past, central governmentsin the socialistcountries used low and fixed prices (including
rents and other urban user fees) and wage controls, as part of their distributionaltool kit. The now governments
in these countries ar now beingasked, in a reformenvironmet, to get the prices ight and to use tax policies and
targetedgubsidiesinstead. There are two importantdimensionsin the transitioneconomies:First, changesin public
sector prices L.rugc ) clearlyneed to be coordinatedwithwage reformon the one hand and withmore general
price reforms on the other hand, in order to avoid undueimpacton incomeand major distributionalshifts (as well
as on profits and hence governmentrevenue). Second, central governmentprice mandates -prices were typically
centrally administered- need to be devolved to the subnationallevel.
25. Finally, in somncountries,(Russia,China, andothers) price reformand other structural shiftswiUcontinue

to bave effectson the tax base and its distributionacrossregions.Coal and energyprice changes, for exampleimply
potentiallylarge shiftsin the vadous tax bases, both vtilbina tier of governmentin the tax bases of differet regions
within a country (e.g. a gain in value added in a primary-productproducing area after price libealization and a
correspondingloss in the consumingjurisdiction)and botwoe levels of governmentdependingon how taxes have
been assignec or shared. Such shifts must be taken into account when designing the intergovemmentalfiscal
system: they maywellconstrainthe reform of intergovernmentalrelations,eliminatingoptionsthat mightbe feasible
in other countries. At the very least, greater flexibilityis needed than in countries where economicstructuresare
more fixed. Intergovernmentalfiscalreformclearlymust be viewedas an ongoingRrocss in countriesin trnsition.
Oreanizationof the Paper
26. This paper addressesthese issues of intergovenmental finance in transitionaleconomies. It argues that
a broader frameworkthan is customaryis needed to analyzedecentralizationand intergovernmentalfiscalissues in
such economies. This frameworkmust incorporatosuch elementsas the likelihoodof continuingstructuralchanges
in the economy and continuingpolitical shifts, t4e need to undertake intergovernmentalreforms while coping
simultaneouslywith stabilizationpressures and the increasedimportanceof the social safety net, the likelihoodof
continued (local)public ownership on a significant scale, the financial implicationsof such ownership and its
possible conflictswith the overall privation objective, and continued vestiges of price and wage controls and
other rigidities. A comprehensiveand atcurate analysis of decentralizationmust take these elementsinto account.
27. The analysis in this paper is based on detailed recent studies of local and intergovernmentalfinance in
Hungary,Poland, Russia, Romania,and China, and on less detailedinformationfor the Czech and Slovak Federal
Republic(CSFR - especiallythe Czech Repub!ic),Bulgaria, and Vietnam. Salient characteristicsof the local and
intergovernmentalsystem in each of these countriesa'are presented in the main text for illustrative purposes.
21See also thefollowingworking papers: Bird and Wallich, 'Financing LocalGovernmentin Hungary, No. 869,
March 1992; Bahl and Wallich, 'Intergovenmental Finance in China," No. 863, February 1992; and Wallich,
Christine, 'Fiscal Decentralization:lnteraovernmentalFinance in Rus&iia' Additionalsource materialsare cited
in the bibliographyat the end of the paper.
-6-
Althoughtheprincipalfocusof Ae disussionis on the cental ad at Europon countria (especiallyHungary,

Poland,andRomnuia),similartndencis andproblemsarevisiblein all th- countriesmtioned and maywellaWso
be manifestin theotherrepubicsof theformerSovietUnionand fonne Yugoslavia, as wellas in similarcountries
elsewhore.Komai(in Tho SocialistSystem)list 26 'soclist' countes in 1987. If we add to this lt the 15
republicsof theformr SovietUnionnd theS of fonnerYugodavia(anddeductthenow-vanihd EastGrmany).
thermwould appear to be at least43 countrieswhichonlyfiv yar ago had moreor les 'Socialist'syms. By
1992,howevor,is probablyfairto say all but two of themecounto (Cub nd NorthKorea)ar nowin momo
'trnsitional stage,motot momor all of the daslymiof this paper ould be applicable.This paperthaefore
representseeially an intetimreporton what is suroto pmve an on-goingand importantproblemin all the
trnsitonal ecoomies.
28. Thepaperis organizedas follows:Sectin II reviewsr numberof geea ius withrespectto theroform
of governmental structurethathavearisenin thetransitional
economic.This ectionelabomates manyof thepoints
madeaboveaboutthe importanceof thd changig contextwithinwhichquestionsof local governmnt roles,
responsibilities,
and financesare beingdeided. Section77 thenouldinesbrieflywhatappeauto be theapprpriate
roleof subnationalgovernment, xuchcountries,with specialattentionto its expediture responsibilities.The
remaiLderof thepaperthendiscussestheissuesrelatingto eachof the posible methodsof financingsubnational
government in transitiotaleconomies:userchages (SectionIV), subationaltaxes(SectionV), intrgovermental
transfers(Section VI), subnationalborrowing(SecdonVII), and the use of subuationalgovemmentasset (Section
29. The matix below summarizessome of the salientfeaour found acros the tramition socialisteonaomies.
| ern~kagmeotal Fiwa Reltion in Trmtansn Ecnm2os

r SalientCbarteristics
Political/OverallStructura Reform Context
* Intergovemnmental
financesreforms are taking place in a context of golitical decentralization.
* Govemnment reorganizationis also taking place throughoutthe trnsition and socialist economies,
and striving for 'local autonomy', many countries appear to be eliminatingthe 'middle tier' of
governmentin their constitutionalreforms.
* Intergovermnentalfinancesreforms are taking place in contextof overalUsbinking role of
* Intergovernmentalfinancesrefonns are taking place in context of a changingrole of government.
* Intergovernmentalfinancesreforms are being undertaken simultaneouslywith overU national In

reforms.
Stabilization
* Intergovernmentalfinances reformsare taling place in context of macroeconomicinstabilityand
inflation, a weak fiscal NOsitiOn
of federal/centrd govenment, and a shrinkingrevenue base at
national level.
-7 -
Box ... (contnud)
|* AMakimi ond overall mco concernsdomiate the agenda so dt intergovernmentali finacial

reform needs to take place in tdis strne contoxt,but cae needs to be takn to prevent the
stabiliduaionobjoctivefrom fully domnting the redign of th inegovermumtsl fian regim.
| Stabilizuon concr lead th center to deal with budgey stin by pushing the deficit down',
andhoping the subiational level can do the cutting.
E In some cas, the center has shifted (potentialy large and open-endad
responsibiliti-sfor 'social welfare down to local level in an attemptto balamcoits budget.
* Expenditure autonomy' is enshuinedin legislation,but not really effective: there are still centra
norms and mandate in placo, and transfer dependancealso impedes autonomy.
* Subsidiesreprent a major item in the subnationalbudgets due in part to central prioe setting on
public sector prices.
* Use ot 'norms' to define how a serviceshould be provided ;imitsexpenditureautonomy.
* Cental 'm4a' on wages, and other expenditureslimits expenditureautonomyand continueto

be importanL
Tax Assi e and Iafm
* Attemptsto make the subnationalsector 'independent' are desiredby the subnationallevel md the
center. center wants each tub to sit or. its own bottom", to assig taxes and be done with it, and to
reduce 'transfor dependance'; This is generallynot possible, because the tax instrumentsare not
sufficient,and user charges not yet at a level to contribute significantlyto cost recovery.
* Subnationalgoverments want to shake off contral control, and be masters in their own house, and
see appeal in the apparentautonomythat 'transfer independence"provides.
* Tax assignmentstypicallyunder-fimdlocal sector: low-yieldingrevenue sources are assignedor

those which will bear fruit only in the very long term (property tax).
f:
|* There is a genea absenceof concrete empiricalestimat of 'correspondence'of tax and
expenditureassignmentsand need for trasfers and legislationon intergoveimmentalfinances
relationsis typicallydevelopedwith no quantitativeassessmentof its implications.
* Generd absence of 'formula"- basedtransfers; they remain negotiatedin aggregatevolumeand by

individuallocalities.
Generalabsence of transparencyin allocationof investmentgrants/capitalfinancing.

-8-
Box .... (continued)
Privatizationand OwnaR
|* Major continuedrole of subnationalgovenments in ownershipventures at a timo when

centr/national priorities sQupoxtprivatztin This relates to the inadequat revenue bau given to
localitiesand their continued financialdependenceon nterprisesand tranfers for rovenue sources.
|* Potetlly significantnegaive consequec for the budget of pnvatizaon (in some cowutres)as
enterpnsosocial expendituresgo 'on budget,' and newly pnvatzed enterprises become hader to
tax. Subnationalgovernmentsmay not want to lot go of this rliable fine=.ing source.
* How to use the saitlry revenuesfrom pnvatzation (debt repaymet, nvestment finance, current
expenditures). How to take advantageof the asset stock held by local governments.
| orurinD
* In most tansition and socialisteconomies,there is generous legal access of subnational

governmentsto unlimitedborrowing, someimes even foreignborrowing.
Extrabudetav I?unds|
* Extabudgetay funds continueto play a major role, complicatingbudgetarytansparncy and

macro-stabilization.
Institud2 Famework
* In may transition economies,the instit>tionalarrngements needed for diaogue on

intergovenmental issues is missing, even though the need for such dialogue is, as noted above,
greatr than in makdeteconomies.
Tax Administration
* 'Inverted. administative systemsin many countres where the local administion offices may
have greater loyaltiesto local governmentthan to central administratve systems.
-9-
H. DECENTRALIZATIONIN PROCESS:TRENDSAND THEMES
A T;me of Trmnsforation
30. Perhapsth most iniportantthing to remmber about the procemsof politicaland fiscaldecentralizationnow
engagingthe countriesunder reviewis how very r it is. Apat from China, whichtook the first tentativesteps
in this directionin 1979, the subationl sytems now emeging in the transitionaleconomiesar of very recent
origin, as indcated by the dates of some of the relevant basic laws:
Poland - L Self-Governmet Act, 1990

Hungy - Local Self-Goveamet Act, 1990
RomWnia- Local Govenment Act, 1991
Bulgar - Lcal Self-Govemmnt and Local Administraion Act, 1991
Rusan Fderation - Law on the BudgetaryRights of Local Self-Govamnet, 1991.
In most cases, additionalimportantlegislationcame even later - e.g. to set out subnationl revenue sources (such
as Poland's 1991Act on Taxes and lA Fees) - and in manycountries is stiU to com (e.g. with rspect to the
precise ownershipand control of some importa local assets). Key elementsof the legal frameworkneeded for
efficient local goverent are stiU missingin most countries.
31. Moreover, oncethe basi ucue of local govemmentwas established,local electionsstill had to be held,
new local governmentsestablished,and a myriad of importan administraivedetails determined. In Romania, for
example, the first local electionswoer not held until February 1992 and the new mayoa did not take office until
April 1M. In the Russian Federation,the first local electionswore held in November 1990, and in Hungary in
April 1991. Even after a year or two's experience, as in Hungary, Bulgariaand the Czech Republic, some mayors
and other local officids, partiularly in the smaller communities,still bave littlo idea of what their real powers,
re sibilities,and limitationsart. Nor are matr much beuterat the centralgovernmentlevel where, in most
countries, no one really seems to be in charge of dcenrliatiadon, with the result that some key issus appear to
have beon decided (or left undecided)more by accident than design, parcularly issues concerningthe intercton
of the new local governmentswith such other lements of the transitionalreforms as privatition, tax reform,
financialreform, preo reform, subsidy reform, and budgetaryreform. Tbis neglect is now causing trouble in an
number of countriesnotjust for the local governient sector but also, to varying degrees, for the successof all these
ambitiousreform plans.
The End of Central Plannina: Filling The Abhorred Vacuum
32. Unlessmore systematicattentionis paid to these problemsin the near future, muas are Ilkelyto get worse
rather than better. Up to now, manyof the potentialdifficultieshavebeen kept in check largely by the continuance,
de facto if not always de jure, of the refonn m under which subnationalgovenments received most of
their funds from central govemmentbudgetarytransfen which wore basicallyallocatedon the basis of negotiation
and bargaining. In this process, the centnl governmentalwayshas the upper hand, controllingas it does not only
the total amountof such transfers, but also who gets them, and when. When combinedwith the continuedcentral
control in most countriesof such critical determinantsof local outlaysas wage structures, enterprise prices, and so
-4In Bulgaria for example, laws passed eady in the reform process only spocified that local governmentsare
responsiblefor 'local matters. This vaguenessin the law allowedlocal authoritiesto claim all or no responsibility
on convenience. This deficiencyis now being corrected, and Bulgara is now in the process of preparing a series
of new Acts covering territorialdivision, local self-governmentand administraton, and finance to replacethe 1991
Act.
- 10-
on, and the ingrainedhabit of local depndence on centmalguidance, the persistenceof this system bas played an
importantrole in preventinglocal expendituresfrom exacerbatingthe genlly tenuousfiscalsituationof the central
government.
33. But this situationis unstable and will not long persit in the absenceof strong central political control.Y
Already, many countries are findingthat the combinationof th manifesdystrong desire for local politicalcontrol
and the unbalaced mdate - virully unfettered 'local autonomy', but a totally inadeque fiscal base - logally
bestowedon the new local govenments is causingtrouble. In the Czech Republic, for example, centralautodities
are properly concened about the national implicationsof both local povemment busines activities and local
borrowingas weUlas the fragmentationof the nationaleducationsystem. In Hungary, whorelocal gover_nN can
spend their budgets, most of which come from national transfers, any way they want, simiar problems are
emerging. And in Russia, which is with the CSFR the only fonmlly "federa transitional country, separtist
tendenciesfueled in part by fiscal pressuresand imbalancesare raising funamntal politicalproblems.
34. The CSFR, of course, has already decidedto split into two sparate republics, effective in 1993. Even
more dramatic, although of course largely motivatedby forces stronger than those discssed here, is the cse of
former Yugoslavia. These two are by no means the only countrie in central or eastern Europe to be etmically
heterogeneous. It may weUbe tat in the end the only way such countries will be able to hold together in a
democraticfashion will be by reconstitutingthemsdves as vey "loon" fedrations - like Switzerlandor Canada,
in which the subnationalunits (cantonsor provinces, respectively)are also 'sovereign states in some important
respects. In these circumstances,central-ational negotiationsbecome as much or more exercises in diplomacy
as in the designand implementationof efficientand equitablegovermental structuresAYIn contrast, the discussion
in the presentpaper assumes, as does the legal frameworkin every country consideredhere except Russia (and the
soon-to-vanish CSFR), that the transitioncountries are organized essentily as unitary states, and that the essential
role of local governmentsis to ensure the efficientprovisionof local public sevices and to act, where it is efficient
to do so, as agents of the centrl government.
35. At present, in most of the countnes under review, those local officials who are not simply waiting more
or less passively,either for orders to come down from above as they have for decadesor, more hopefuUy,for their
future to be unveiled for them, haveunderstandablybeen using their eaergies primarly to attemptto wheedlemore
moneyout of the central government. In Romania,for example,where the entire 1992 trnsfer budget was spent
by the middle of the year - largely on centrally-mandatedsubsidiesto enterprises- it is not suprising that Bucharest
city officials were to be found two or three times a week at the Ministryof Finance trying to secure more fumds.
Other local officials, forhtnateenough to possess clear tide to significantassets, have been selling them off and
using the proceeds for current expenditures(as in some towns in the Czech Republic)or trying to use their assets
as equity in joint ventures with foreigninvestors(as in other towns in both Hungary and Romania).
36. On the whole, however, there is as yet little eidence at the local level of wide deviationsfrom desired
national standards in the provision of essentialservices, other than that resulting from the general fiscal penury
afflictingall these countries. Similarly,perhapspartly reflectingthe relativelyundevelopedfinancial sector, there
is also no evidenceof either significantbudgetarydeficits at the local level or of excessive (or often, any) local
borrowing. As it gets harder to squeeze muoney out of the center, however, and as it gets more difficult to get easy
financefrom seUingoff assets, subnationalgovernmentsin the transition economiesare going to find it increasingly
temptingto move in both these ditions.
althougheven
-China, where such control largely remainsin place, is of course an exceptionto this genemalization,
there many of the problems discussed in this paper are manifest.
-I'Foranalysis relevant to such federai states, see Shah, *Perspectiveson the Design of IntergovernmentalFiscal
Relatiers," Worling Paper No. 726, 1991,and Bird, FederalFinancein Comprative Petpective (Toronto, 1986).
- 11 -
37. With respect to education, for example, the finacial incentives in the preent Hungariansystom may in
the future, unless changesare made, lead to larger townswith seondary schoolsbeing unwillingto acceptstdents
from smallertownswithout suchschools. whe completelyunconditionalnatureof intergovenmemtalfiscaltansfers
in Hungary make it very difficultto penaliz such actions. Simily, national educationstandardshave vanished
in the Czech Republicbut have not as yet been replacedby either guidelinesor nrles: totalitarian educationwas
undoubtedlyan vil, but it is not obvious that anachy is better. For the fist time in decades theimis clearly a
possibilitythat wide variations in the qualityof educationwill eerge dependingon the ficl resources available
to the municipalityin which a studkt lives. Similarlywide vartions may equally anse in health, in so-al welfare
expenditure,and in other aue of local govenint activityas the power of the center to commandvaniss. This
outcomeis unlikelyto be caosiderd desirableby most citizens.
38. It is one thing for local expaoditurechoices to reflect the wishe of local officials who have been
democraticallyelected and who wa spendingfunds secured from the citim who elected them. But it is quite
anotherwhen sovices with substantialjuisdictional spilloversare underprovidedin poor areas owing solely to lack
of fiscal resources, or when trasfer moniesprovidedby taxpayers as a whole for specificnational piposes are
misspentby local officials responsiveonly to their own local electors. Yet both of these outcomesare all too
probable over time, given the virtusallyunrestricted 'local autonomy' conferred by some of the new local
governmentlaws (e.g. in Hungay). For much the same easo, the almost unrestrainedborrowing authority
bestowedon new and untried local authoritiesby some of these laws carries with it the potentialof eventud fiscal
disaster.
39. As yet, the subnationalsystem is too now in most countries for most of the bad things that mighthappen
to have occurred. But ther aro enough signs to be worrying, and it is critical to use this time of transformation
and reform to continue to develop the local govemmontstuct in a more positive direction in a number of
importantrespects. It is, for example, importantto carify exactlywhat is, and is not, meant by local autonomy
and to flesh out more clearly the role envisged for local govenunents. It is also important to developadequate
institutionsat the centrl level to support the manynew, and often small, local govenunentsthat have beencreated.
Moreover,the role and nature of institutionsinrmediate betwee the central governmentand the local govemments
has also tuned out to be both a critical issuoin many countries,and an impoant differentiatingfactor fromcountry
to country. Thes questions are discussedfurdtr in the balance of this section.
PoliticalDecentralization
40. The reformof local and intergovernmentalfinanceis taing place in the contextof politicaldecentralization
and in particular a strong commitmentto decentrliation of certain responsibilitiesand revenueto self-governing
local administrations 1 In reactionto decadesof over-talized
z. public administration(and related party control),
most Eastern and Central European countries are gngagingin a process of extensive decentralization,gradually
reducingthe role of governmentin genera and the powersof central govemmentinstitutions,and especiallytheir
planningand control apparatus, in particular.
41. In line with this tendency,autonomyand controlover 'local matters, and accountabilityfor them,is being
increasinglydevolvedto the newly-electedlocal governments.The centrl governmentwants 'each tub to sit on its
own bottom', to assign taxes and be done with it, and to reduce transfer dependance'. Local governmentswant
to shake off central control and be masters in their own house, and see appeal in the apparent autonomythat
'transfer independnce' provides. For thoseand other reasons,the principleof localautonomy is nowenshrined
-'In Bulgaria however, the pace of the former exceeds(and precedes) the latter. There is a risk at present of
increasing fragmentationin the number of self-governingunits in Bulgaria because of the perception of local
governmentsthat 'self-governing status will lead to greater budgetary transfen from the center. In order to
counteractthis trend, it is importantthat aspirationsfor local autonomybe temperedby specifyinga thresholdfor
self-financing/ownsource icvenue mobilization.
- 12 -
in the legislationof most transitionalcountries. As a rle, this legislationaffims d liation, local financial
autonomy and the administraive 'independence' of the uunadonal level from cental control. The language
differs, but the geonel intent is clearly to free subnational govenments from the 4ead hand' of centralized control
and to let local democracyflounish.
42. Thesepoliticalreforms are paralleledby dcentaliton of fiscalresponsibilities.Someformof subnational

govemmentstucture had existedin most of the transitioneconomies mder the socialtstregime. However, the fiscal
systemof this regime was essenidaly 'unitary,' with the sub-national level being little more than an edmin
unit or 'departmet' of the center, with no indepe;dnt fiscal or legisltive responsibility. Kornai (I992, op cit)
sets out the former socialisSsystem in considerable detail. It is revealing that this lengthystudy hardly refers to the
existenceof subnationalstate administrations,noting only that they are tightly cntrolled in al respects by the
centralbureaucrcy. In effect, local govenents were basically nits (or branch offices) of the
udeconcentrated'
central govermmentand had little or no financial autonomy. This was true even of those countries such as the
USSR, and Czechoslovaliawhich were formallycalled 'federations'. It was also true in the RussianFederation,
despite the fact that Russia's 31 autonomousrepublicsand national regions wero said to be 'autonomous' and to
haveindependentbudgetaryrights. Policy-makingwashence very controlledand centrHzed and local government
hadvirtuallyno independenttax or expenditurepowes - part of a larger picture in whichthe budget itselfwas seen
only as the handmaidenof the Plan.
43. ReorganizingGovernuent. Now, however, virtually every transitional country is to varying degrees
decentralizing,deconcentrating,and delegadngfunctionsand responsibilities.YDecentalization has manymerits,
both in political and economic terms. It has, for example, the potential of improving the efficiency of local
governmentby subjectingits actions to the scrutiny of the local electorate. Autonomyover local decision-making
and expendituresalso frees local governmentsfrom the heavy hand of centrd control. As yet, however, these
benefitshave not been attained for two quite distinct reasons: the persistenceof the old ways and certain flaws in
the new dispensation.
44. In Romsani, for example,there are at least three importantreasonswhy local autonomyhas little meaning
in reality as yet. First, although the Public FinanceLaw clearly says that every local government is supposedto
have budgetaryautonomy,it also gives power to the central govment to change local budgetary allocations.9'
Moreover, as noted above, the habit of deferenceto centrl commands, even if no longer legally binding (or
enforceable)still persists in many areas. Secondly,at the present tibm local govermnments in Romaniaactually
deliver very few services themselves. Instead, their basic role is to contract with reaies autonomies(local public
enterprises)for the deliveryof most local public services. These 2ngj, like al enterprises in pre-reformRomania,
were of course just another part of the public sector under the previous regime. Under a 1990 Law on the
Restucturing of State-ownedEconomic Units, these enterprisesare now financiallyindependent. Althoughthey
are legallyresponsibleto a board comprisedof me4bers of the relevant local government, they appear to be acting
both more independentlyand at the same time more in response to the many cental controls and egulationsto
WDeconcentration refers to dispersionof responsibilitieswithinthe central governmentstructure from the center to

regional*branchoffices', and differs from 'delegation' in whichlocal government may execute certain functions
on behalf of the central government(and be accountableto it for their performance), and 'decentralization', in
whichfull decision making and implementationauthorityis transferred to local govenmment,which is accountable
only to its own constituents.
2'Art. 51 of the Public Finance Law states that 'the distributionof the revenuesand expendituresor categoriesof
local budgets ... is establishedby the decisionbodiesof each county [district]and of the city Bucharest.' ButArt.
22 of the same law states explicitly that the central governmenthas decisive power in the formation of local
budgets. Each year, local budgets are submitted to the Ministry of fmance no later than July 1. The Ministryof
Financeexaminesthe draft budgetsand has ten days in whichto "ask for the necessary modifications' in order to
establishbudget balance. Most importantly, 'in case of divergencethe (central] Govemment is to decide'.
- 13 -
which they are stil subject thn at the behest of local govemments. Finally, as in most of the transitional
economies,Romaniahas createda large numberof small, poorlocal goverments with almost no resourc0sof thoir
own. Autonomy' does not mean the same tbing to a small, poor rural commune as it does to a large, rich city.
45. Lcal institutionsare more developedin countnes such as Hungry, but these three problems- conflicting
legal interoptations, tangled rmlationswith enterprises, and considerable diversity in the capability of local
govemments - exist almost everywhere, and cast some doubt on the real significance of local autonomy.'
Moreover, even where such implementationproblemsdo not exist, the fiudamental problem with th way local
autonomy is being approache in the transition oconomiesis that such autonomy is not moeningfil if it is not
accompaniedby capacity, and it is not desirableunless it is accompaniedby responsibility.
46. To makelocal autonomymenjrtef, local govements need adequatolocally-controiledrevenues. Not

only must decisionsabout the provisionof governmentservices with local benefitsbe made by local governments,
but such goverments must have the resources to carry out such decisions. As developed further below, local
expenditurediscretionis stil limitedin imponantways: some such limitationsmay be good; others clearly are not.
Moreover, the fimdamntal inadequacyin most couttries of the local revenuebase (see later discussion),and the
resultingcontinueddependanceof local financeon ad hoc transfersmeans that revenue capacityis also inadequate.
Greater expenditurediscretion in those expendituresfiuly assigned to the subnationallevel and greater flexibility
for sub-nationalgovemrmentsto raise their own revenuesseem needed in most countries. Such revenue flexibility
and autonomyis a crucial step to maling local govemmentmore accountableto local residents, and to improving
the efficiencywith which local public services are delivered.
47. One meaningof local autonomyis thus that local governts shouldbe able to raise and spend revenues
from their own revenuesourcesin any way they wish - subjectto cenain limitaions on the revenuebases to which
theyhave access (see below). Since local governmentsare responsibleto local taxpayersfor what they do with the
funds they raise, in principle it is neither necessary nor desirable for the central government to tell local
govemmentswhat they can do with their own money. If those who pay for local services financed from local
revenues, the local taxpayers (and voters), are unhappy with what their locally-electedauthorities do with their
money,they may dismiss them.
48. On the other hand, if the centml govemmentis unhappywith uwht some local governmentdoes with the
money raised from local taxpayers, its unhappiess may be unfortunate from some perspectives, but it is
unavoidable.Whatlocal autonomymeansis preciselythat there is nothing that the centralgovernmentcan or should
do to alleviateany such distress it may feel - barring extremecases such as corruptionor when other nationallaws
are breached.
49. Local autonomyemphaticallydoes not mean, however, that local authoritiesshouldbe able to do whatever
theywant with other people's money- that is, with taxes paid by other than local residents. Truu local democracy-
presumablythe ultimate reasonwhy 'local autonomy' is considereddesirable - means *responsible' democracy,
which in turn requires full and explicitpoliticalaccountabilitvto the sourceof local governmentresources for what
is done with those resources. Local autonomyin this sense is thus not only consistentwith, but also requires, that
two importantconditionsbe satisfied:
1. The access of local authorities to taxes that may be *exported- to non-residents should be
severely restricted.
2. Subnationalgovernmentsshould as a rule be accountableto central government (that is, to

taxpayersin general) for the use made of central governmenttransfers.
Acceptanceof these conditions,as is necessaryif efficient resourceuse is to be achieved, has strong implications
both for the design of the local tax system and for the design of transfer systems, as developedbelow.
- 14 -
50. Th'eeenal economicrole of subntional governient is to provide to local residentsthose public services
for which they armwilling to pay. As already noted, local govemnmetsmust be accountableto their citizens for
the actions they undertsketo the extentthose citizens finance those actions. Similarly, local governmentsmst be
accountable to the central govemmentto the extent they are in effect acting as 'agents' and are financed by
transfers. Accountabilitvin this sense is the public sector equivalentof the 'bottom line' in the private sector.
51. Such accountabilityclearly requir thst subeationalgovernmentsshould, whenoverpossible, choa for

the services theyprovide, and, wherechargingis impracticable,they should financesuch services from taxes bomne
by local residents,except to the extentthat the contral governmentis, for reasonsdiscussedlater, willingto pay for
them. Public soctor activities are unlikely to be provided officiently unless the lines of responsbiity and
accountabilityare cleady establishedalong these lines.
52. On the one hand, then, local govemnmentsneed to be given accessto adequateresources to do the job with
which they are entrsted. On the other, theymust be held responsibleto those who provide these resoure - local
residentsor central goverments, as the case may be - for what they do with them.
53. In principle, local govemmentsshould therefore not only have access to those revenue souces that they
are best equipped to exploit - such as residentialproprty taxes and user charges for local services - but they
should also be both encouragedand permitted to exploit these sources as fuUy as possible, subject to certain
limitations.Unless local governmentsare given some degree of freedom with respect to local revenues, including
the freedom to make mistakesfor whichthey are accountableto their constituents,the developmentof responsible
and responsivelocal govemmentin the trnsition economiesis likely to remainan unaainable mirage. On the other
han;, if local governmcatsare given cate blauche to spend o-ter people's money,the alocationa, distuibutional,
and maccc consequenceswill almost certainly be undesirable.
54. There are of course dangersin permittinglocal governmentsevenlimitedfreedom. One danger in the eyes
of some is thst they will not utilize fully all the revenue sources open to them, thus allowing the level and quality
of public services in some areas to deterioratebelow the standard considered desirable. But this is not a real
problem. If the savice in questionis reaUyone of national importance (e.g. reseach) or one in which there is a
strong national interest in maintainingstandards(e.g. poverty alleviation),it should be nationally funded at least
in part and its achievementmonitoredcorrespondingly. If it is not a matter of national interest, why should the
national governmentbe concemed? If the local electors do not like what their local govenment does, or does not
do, they can 'throw the rscals out' at the next election. As noted above, the freedomto make mistakes, and to
bear the consequencesof one's mistakes, is an importantcomponentof local autonomy.
55. Another danger, more salient from an economic perspective, is that local governmentsmay attempt to
extract revenues from sources for which theyare not accountable,thus obviating the basic efficiencyargumentfor
their existence.To counter this inevitabletndency, central govenments should in principle deny or limit access
to taxes that fall mainly on nonresidentssuch as most naual resoure revenues,pr-retail stage sales taxes and,
to some extent, nonresidentialreal property taxes.
56. Anotherway to counterthis problemto someextent may be to establisha uniform set of tax bases for local
governments(perhapsdifferent for differentcategories such as big cities, small towns, and rual areas - although
this may not be politicallyacceptablein all countries), with a limited amountof rate flexibilitybeing permitted in
order to provide room for local effort while restrainingunproductivecompetitionand unwarrantedexploitation.
It is especiallyimportantto provideadequateflexibilityto exploit good local tax bases to avoid creating a situation
in which the only flexibilityavailable veal governmentsin their struggle to cope with budgetaryprssu is by
exploiting such economicallyundesiable sourcesof revenueas local businesstaxes or, evenworse, local business
enterprises. As emphasizedlater, the businessof governmentis not business,and enterprisingmunicipalitiesshould
not be encouragedto developlocal enterprisesin order to secure the revenuethey need to function.
- 1S-
Tho Size and Scopeof Government: How Many Levels?
57. In the pmroform period, virtully all the countries under review had a multi-tier organizion, with the
central level implemeaing its control over the local level throuh an of government.The central
level,variouslycalled'Oblast
govermentwa responsiblefor theoverallplanand budget,and the intenrmdiate
(in the Russian Fedation); *Judet (Romania);County Council (Hunguy), and *Voivoduin Poland, was the
administrative,control d surveillanceagencytough which o cer effectedconol and implemented
policy.w This middle tier ovemw tho expenditur of local governts and wa the chnol for centil fical
flowsto thelocalities.Substatvoly,however,noe of the countriesdiscsed her wasorganized,froma fiscal
perspective,u a federaton. Althoughthe RussianFederotin gv nomial autnomy to its 31 'auftomous (or
national)regions, okmup and republics, these sb-ntioad gove too wero nither fiscally idependent nor
able to determin their own expenditur and tax policies. At the other extrme in some ropects is China, a unitary
country with a vey decentralized inison.
58. At present, only Russia and Czechodova have a thre tier fml systm of govemment,comprising
the federal govemment, republics, and regions - and t Czech and Sloak Fedeal Republic, as it is formally
called, is scheduledto septo shortly into its two componet pua. All the other countries discussedhere have
a uitarv system of government. A first look at the lol stucture suggeststhat Poland, Hungary, and Romania
have a two-tier level of government(central and local), Rumds,China and Bulgaia a three-tierstruture (central,
provinces/county/regionaland local), and Vietnam a four-tier level of goverment (czatml, provincial,
municipal/districtand comneighborhood) a shownin Table 1. In fact, however, matters not so simple
ae
anywhere. Romania, for instance, has two tier of local govemmt (the old judets and municipalities):all of the
area of a judet is divided into unicipalities,but in principle each type of local governmentis independentof the
other. In Hungary, on the other hand, whilothe old countiesstill exist, they really have no functions,so that there
are m effect only two lovels of government,cental and locaL On the oter nd, in Budapest(as in tho citie of
the Czech Republic)there are *subcity (distict, or ward) councils and goverments with electedcouncilsand their
own budgetsand assets, so that thoe local level of govenmnt itsf ha two tiers - as indeedit does in the largely
unrefonned stucture of Vietnam. In Bulgai, the municipalitiesar the basic unit of self-government,but they
have contiguousbordes and therefore both urban and ural. The municipalityitslf has an elected Counciland
Mayor, but urban settlents within the municipaity also hav an elected Mayor (no Council) with delegated
executivepower, from the municipality(ie. county). This lower level is thereforea deconcentratd level of local
govemment.
59. As a rule the middle tier of goveiunt cotinue to aist in vestigial form in most tansitional
countries, but it is struggling both for a role and for legidmacy. Most coundriesseem to want to shed this
intermediat tier, associatedwith the formercontroleconomy,as if to put their past behindthem. In some countries
the role of the interdiat tier is now limited largely to overseeing the constitudonalityand legality of local
govemmentoprations. lI others, its role has become one of coornanting cental governmentpolicy at the
regic *allevel. In most, however, there are few quesdonsmore in flux than the presentand futre role, if any, of'
intermediatelvels of government.
60. In HAomgm,for example,ther are now 8 regions, 19counties, and 3070 localities. The regions,headed
by a central official called the Commissionerof the Republic, have two functions: (1) information and
dissemination,which involves collecting data on local govenmet activities and providing informationto local
govenments to facilitat the implementationof national policies; and (2) to ensure that local governmentactions
(e.g. on taxes)are constitutional. The functionsof the 19 counties, formerly the main teritorial control instrument
of the state, have also been severly restricted. counties retain some minor fee revenuesand are supposedto be
L' Theseare, strictly spealdng,decentralizedbranchesof the centralgovenmmentadministration,not an indpendent

tier. There are 88 oblasts in Russia; 9 oblasts in Bulgaria; 19 County Councils in Hungary; 49 Voivodshipsin
Polp.nd;and 41 Judets in Romania.
- 16-
responsible for functons of an intuAunsdicional natur that servo swvral localies. In contrast, the local
govemnments(almost twice the number of the former 1563 local councils) have considerable expenditure
responsibilities,the right to spend national transfes and shared ta a they so fit, and amost unlimitedpower
to borrow, own and disposeof property, and to manag, establish, or sel public eaterpises.
Table 1
Trends in cr t Governmnt Orgnization
country UnitayFdal # of Tiern Em N of Tiers (preet)

___________ _______________ (pro-rform
) Structm
Hungary Unitary 3: central/county/ Unitary 2: cental/local
local
Poland Unitar 2: cenral/local Unitary 2: voivodship8/
Ronania Unitary 3: central/regional// Unitary 2: central/local

local _ _ _
Bulgaria Unitary 3: central/regional/ Unitary 3: central/oblass

local (regions)obahina
(municipalities)
CSFR Federl 3: central/state/ 2 countries 2: republics/local
local
Russia Federal 4: central/oblast/ Fderad 3: federal/oblast/
myons/vi-tge rayons
_ __ _ _ _ _ _ _ _ _ Soviet_ _ _ _ _ _ _ _ _ _ _ _ _
China Unitary 3: centrl/province/ Unitary 3: central/

local province/
local
Vietnam Unitary 3: centDal/province/ Unitary 4: central/
local province/
municipal/
commune.
61. T,hestuation in Romaniais quite different. Althoughthe 41 judets continue to exist, there are now two
completelydifferentadministrativestructuresin eachjudet The first, headedby an appointedprefect, is essentally
the deconcentated territorial admnistaton of the state. The seoond, haded by the elected president of the judet
council, is the locl govenment at the judet level of local gov ere, which is responsiblefor works of regional
interest but whichhas no hieruchial authorityover the 2,948 other local governmentsU.
L'Me situationwhere a local govemmentjurisdictioncoincides territoriallywith a deconcentratedechelonof the

central governmentadministaon can be found in other countries. T.heRomanianshave basicallyborrowedthe
French local administtion system as it emaergedfrom the 1982decentralizationreform, at least as far as the roles
of the departmt (judrte) and the communes(comunesi orase) are concerned. (They did not copy the truly new
feature which that reform brought about, nmely the creation of the region as an additionalintermediatelayer of
- 17 -
62. In EgJg, the 49 vovoidahipsaneesetaly deoncetred publicservice inmdiaes for cti central
expenditures(mostly in hlth and socal welfa), although other central expenditures are camed out locally by
line ministries. All local governmentfunctions ar caried out directly by the 2383 gjina governmons. Prest
plans, however, ate for a new iermediat tier of local govenmet (probablyabout 300 units called powiats)
to be put in place in 1993. Similarly, in the CzechRepublic, whoe there are at preseat over 5000 local
govebmts, there ar also 7 regions and 75 districtse_nally left over ftom the preious regime, and plans are
being consideredfor creating anything from 3 to 20 to 7S "second-fierm local govamenats.
63. Finally, in tIe Russn EfSop, a sligbtlydifferentpaten is emerging, with the obast level tddng on
a laorg role and emerging as an iwpota unit in its own right. ndeed, if the fiSecalreforms at the subnational
level proposed in macmst logislationare implemted, the Russin Federtion will have decentlid expenditure
responsibilitiesand to a lesser degre, rvenues, to the oblst level and will take on the character of a tme federal
state.
Size of Goveme Units: Efficiey-and Other CMgM.
64. Althoughthe initialreaction against the previous intemediate govemmentsas udesrable remindersof
the old regime is dendable, so is the obviousconcernin many countriesto put somethingmore acceptablein
their place. As shown in Table 2, most of the east Europeancountries have created a large number of small, and
probablynot viable, local governmentunits. In Hungary, for example,the averagopopulationof local goverments
(excludingBudapest)is only less than 3,000. In Romani*,half the populationlives in local governnowtunits with
a populationof less than 4.000, such units accout for 90% of all local governments, and the average population
of all local govenments is only 7,500. Poand's locadgovenms ar smilarly fragmented, with an average
populationof only about 16,000.The Czech Republic,a compactcountry with a little more than 10 millionpeople,
has over 5000 local govermments.In Bulgariathere are 3,984 elected local govemet units (settlemnts) with
the averagepopulationof 1500. At the more aggregatelevel, there are 273 municipalitieswith the averge size of
around 30,000.
65. Such extensive iixm on may be good for getting govemment close to the people, but it also
substantiaUycomplicates intergovernm tal relations - it is, for example,imWposiblefor the central govemmentto
maintaindirect bilateal relationswith so many units -, reduces the political~roice of local governmentas a whole,
and in many cases yields localitiesthat are sunply too small to provide efficientlyall the public services demanded
from them. At the very least, even if it is decided not to reate a formal intermediatetier of government, some
mechanismfor coordinationand consultationis likely to prove necessary. Possiblemodels abound in the world,
e.g. the 'municipal federations that exist in many of the Nordic countries, under which localities orgaize on a
voluntarybasis to providecerin local services,or the Nordic secondarycommun-e, county-likebodieswith their
own elected officialsresponsiblefor providingservices of an inteijurisdictionalnature.
local (regional)government). Similar situationsexist in other countries, for instance,in Germany(Bavaria)where

the landkroiseare both local govemments(gebietskorperwhaften)and the lowestechelonof the Stateadministration.
- 18-
Table 2
Deloizationu Scaleand Sie of Govenmenw
Country of units of local governumet Average Size

Hungary 3070 municipalities 2,834
Poland 2383 gnas 16,000
Ronuia 2940 urban and rural mniucipa4itie 7,S00
41 judeta S37,804
Bulgaia 273 municipalties 30,000
CSFR 2843 cides in the SlovakRepublic 1,867*
5500cities in the Czech Republic
Russia 2000 rayons 125,000
88 oblast 2,840,909
China 30 provincs 33 millon
Vietnam 44 provincs 1.47 mon
Note:* For the CSFR as a whole.
Sum,ortingSubnaional Government: The Ihstutional Setting for Subntioinsl GovemmentislReform
66. The lack of an approie contralgovnnt strucr to moitor and supporttho new local govanmts
is a common problem in central and eastern Europe. In Poland, Hungary, and Vietnamfor example, thDreexists
no institutionor designatedperson in cbare of the deentlion of the localitiesand the Ministty of Finance
still plays a lead role in the budgetaryfunctionand other fiscal issues at the local level. In Russia the formation
and implementationof fiscal polcy is done by the centrl governmentand three Parliamntay commisons. The
budget is submittedto the Parliamentfor approval, the collectionof revenueis done by the State Tax Service, and
expenditureauthorizationis dr the control of the MOF. In Bulgaria, the Ministryof Regional Development,
Housing Polcy and Construction is responsible for formulating the policy on terrtoral division and local
administrativefunctionsand guidingthe reformprogam whilothe MiniAryof Financeretainsprimy controlover
the linked issuesof local govrment financeand budgeting. In Romaniathe formulationof decentalization policy
and implementationof local governmentreformis tife resnsibility of tho Departmentof LocalGovernmet Affairs
of the Office of the Prime Minisr and the Deparmt of Tax Administationand Local Budget Managementof
the Ministry of Finance and Economy. Simildy, in the Czech Republic, both the Ministryof Finance and the
Ministryof Economic Policy and Developmentare involved in shaping the intergovernmentalfinance system.
67. Despite(or perhapsbecue of) the involvementof a number of centralgovenment institutionsin a number
of countries, quite a lot that shouldbe done sem to be going undone. Amongthe tasks for which responsibility
must be clearly establied are MomitOlinland ssg subnationdfnms both in total and individually. This
task might be located in the Ministryof Finance, in a special local govemmentdepartment, or it might even be
given semi-independentstatus with some local goverment involvement(like the US Advisory Commissionon
IntergovernmentalRelations or the similar body in Australia). There ae pros and cons of each possible location
that need to be considered carefully. But what is beyond dispute is the need to ensure that the central authonties
havea better underanding of both the existingsituationand the likely effects of any proposedchanges thanis true
at present.
- 19 -
68. In addition,reular and dogiledfingaial da shouldbe maintainedon subutionalgovenments.Ideally,

this requires;
(1) the establishme of uniformfinancialreporting(and budgetng)systms - perhapswith

and
differentdegre of complexityfor differentcategoriesof local governments;
(2) the desigaion of an agencyresponiblefor coUectingand procesng then datain a timely
fahion (whichwiUprobably,in a larg and complexcountry,requie use of some sampling
thniques as wellas occasionalcensusexercisesto providetheneeded frtame).
Thefirst of these quiem ts in tun requirmthedevelopment of suchsysteumand, muchmoredifficult,their

implementation, which will geally require subsatial trining and folow-up efforts. Developingsuch
'infrastricture' is neitherquicknor cheap,butit is esential if ad importantpublicsectoractivitiesare
substantial
to b decentlized - for whateverrsrn - withoutlosing sight of what is goingon in these importat pua of the
public sector. Implementationof this recomendation may rqui additionalcentnl govem ntot
institutions(1) to developand maintainthereportingsystems,(2) to tran and supportlocalgovenmentofficials,
(3) to developindependent auditingproceduresfor localadministrations and theirrovenueoeaningentexprises,and
(4) to run the stical side of the opation - withthelatter beinga maininputintothe workof the instituion
referred to in the first recommendationabove.
69. Additionaltrainingefforts are clearlyneeded to uagade the technicalcanacity of subnational overnments

to carry out efficientlyand effectivelythe expenditurofunctionsfor which they are now responsibleand otherswith
whichthey may be entusted in the future, whetherin exerciseof 'local autonomy or as delegatedagenciesof the
central goverment In some insance. (e.g. education)the neary support should probably como from the
relevant central minisy, while in other instane (e.g. road and waterworks) it may come fom the ministry,
regionalagencies, universities, private sources, or some mixture of the preceding. In view of the heterogeneous
nature of both the services that are (or may be) decenalized and of most countries, it is unlikelythe same solution
will be appropria in aU coass, but it is clealy a high-pority item to work out in detil exactly who is going to
supportthe new lood role in each functionl area and how that support is going to be delivered. Each functional
area - including the budgeting and financi rporting discussed earier - has different needs, problems, and
possibilities,and will likely have to be treateddiffetly.
70. Finally, an especiallyhigh-prioritytask for the neededcentral unit monitoringlocal finance is to analy
and evaluateintermovernmental fiscaltransfers. As indicatedlater in this paper, thereare many questionsthat must
be raised about the objectives, the design, and the effects of such transfers. There is also much to be said for
encouraging much more informed and open discussion of these matters than is common in many countries,
especiallywhen the whole process is as novelas it is in the transitionaleconomies. Regularpublicationof data will
of course help, but one cannot rely solely on an interestedparty - the central govermment- to carry out, and
certainlynot to publish, all the analysis that is really neededin any country in which intergovemmentalfinancial
issues are important, whether in political or economicterms, or both. For this reason and others, there is much
to be said for establishing as soon as possible in most countries one or more smal non-govermnentalresearch
institutesfocusingon local governmentproblems.
The Changin Role of Government:From Ownership

to Service - Provision
71. Decentralizationand intergovernmentalfiscalreforms are taking place in the contextof a major change in
the role of goverment in the trasitional economies.Ia the past, governmentsat both the nationaland subnational
levels have been involved in almost every aspect of economicactivity. The state owned all enterprises, from the
steel plant on the 'commanding heights' of the economyto the local laundry. The state organized production,
allocatedlabor, planned exports, and so on. The first stage of decentralizationin some countries began in the
1970's, when the ownership of some smaller industrial units and the retail sector were transferred to subnational
-20 -
govemments. By the late 1980's, local goverments in countries such as China owned a substantialpatt of the
economy- light industry, and some manufactuning.
72. More rect reforms aim to decentralie owneip to tho privat sector. The trnusitionis still far from
completo. Althoughall the east Europeancountriesare committedto privatzing stlte assets, to liberaliang prices,
and to withdrwing from the role of the state as direct pducer and prvider of ecoomic goods, many produetive
assets - indudrial assets, agriculturl and urban lad, the housing stock, commrcial property - remain in
govenmment hands, somedmes local, sometimesnational.
73. Whileeverything is in flux in most countries,as a nle at the local level the satus quo contasb starkly
with the role local govemmentsgenerallyplay in maurktec4onmies.In the lIser, local govenmments provide basic
urban inftucur and local public services (sometime through local public enterprs) and often act as local
administratorsof nationa policies in such area a education,health, and social sevices. In the former socialist
pattern, on the other hand, enterprises owned by local govenments had their investments financedfrom local
budget', and local budgets in tum derived revenuesfrom Otheir eantepriss. C
were close, and the distinctionbetween enterpris functionsand governmentfunctionsoften murky. Lcaities often
asked enterprisesto construct and financesuch facilitiesas roads and schools, clinics and even sports stadia and
clubs; 'donations' to the budget for one-off outlayswere also com rm Jt made little difference whether such
outlayswere directly undertakenby the enterpriseor via the more circuitous route of profit transfer, and funding
via the governmentbudget: the burden on the enterpriseand the not budgetaryposition was the same. All this has
now changed, or is in the process of change.
74. Until recently, for example, most local government revenue from enterprises came from 'profi:
remittances'. As a result of reforms to the national tax system, and especialy of the enterprise tax system,
however, these remittanceshave largely been replaced by taxes on *thei enterpriseeW. Somo countries still
provide explicitly for such 'ownership' rights to enterprise taxes: in Romana, for example, local governments
(iudetsand municipalities)have the right to both profits and dividendstaxes levied (by the central government)on
their public enterprises. In Poland, prior to the racet tax reform, both profits and dividendtaxes similaly went
to the 'founding' sbnatidonalgovernment,and in additionthesegovemnments (vovoidshipsandgmina) got specified
shares of the similar taxes levied on state enterprises which had significant activities in their territories. This
processwas so complicatedthat the Polish tax administrationspent almost as much time allocatingenteTprisetaxes
amongjurisdictionsas it did collectingthem.
75. Even where more modern systemsof tax adminitraton have been adopted, as in Hungary, there is still
a strongtendencyfor local governmentsto feel theyhave a primaryclaim on the tax revenuesgeneratedwithin their
jurisdiction. This 'source entitlement approachto the allocationof fiscal revenues is of course familiarin such
contextsas internationaltaxationand the taxationof natural resources. But it is seldomso visible at the local level
as it is in some of the transitioneconomies,in part becauseof the obvious gross inequitiesthat would result from
allowinglocal governmentsto keep large shares of the revenueslevied on firms producing for a nationalor world
market. Of courmse, exactly the same problem arises, or wil arise, in the transition economies,as noted below.
76. Ownershigversus Service Provision. The fundamentalchange in the role of subnationalgovernments

representsa major challengewhichhas as yet not been fullyrealized in most countries in trasition. The key role
of local governmentsis no longer as o but as saervi viders and reguators. The wrenchingreforms most
countries are undertakdngin order to achieve macroeconomicstabilizationand to compress the state budget may
easily produce perverse effects at the local level. As noted below, some central govermments,for example,have
shifted additional expenditure responsibilitiesdownwardsin the hopes of improving their budgetary position.
Revenueassignmentsand/or transfersto local governmentsare generallytightlybudgeted for the same reason. The
UySee Tanzi, Fiscal Reforms in SocialistEconomiesin Transition,IMF, 1991,for a full discussionof the national
tax reforms that are in process in most countries.
- 21 -
result is inevitablythat local govenments ar sorelytemptedto rely on entises for continuedbudgetaryoudlays,

and consequantly to protectlocally-owned enterprisesto ensuretheirprofitability.Unlessa workableframwork
for intergovernmentsl financeis put in place, budgetarystringencyis thus likelyto createadditionalhurdlesto
disassociating localgovenmentsfrom thir enterps.
77. T'heIQ-S4t Oda facingboth local and centra govements are also very difrent from the past.
'Adm trativointeventiods', suchas the detailedregulationof prices,the administationof waitinglists for
housing,and theallocadonof landand mtria suppliesare no longerqriropriate. Instead,whatis nowneeded
is to developand implement an entirelynowregulatorystrctur to provide thoconditionswithin which the market
can do its job e.g. reulation to avoidmonopoly,preventfoodadulteration,curtailfinanca fraud,andso on. In
thisvein,oneessentialtaskof thenationalgovermentwillbe to set up an adequateand clearlegalframeworknot
only for the privatesectorbut also for the deces.alizedlocalpublicsector. Local govermmntsface theirown
unfamiliarregulatorytask, particularlywith respect to landuse, enviromental problems, and retail trade.
78. Notonlyis theroleof govemmentchanging,butits mois changing.Intergovernmental

fiscalreformsare
taking place in the context of an overall shriniing role of government, with all thstresss that a smaller pie
imposes on decision-makers. As shown in Table 3, the size of govenmmentwas much larger in the transition
economiesthan in mulaet economies.In many cases, reforms have alrdy had a major influenceon the overall
"size' of the govanment sector, althoughmuch more some than in other. In China, for example,the government
budget declinedfrom 40% of GDP in 1979to Ims than 20S of GDP in 1989;in Hungary, the similar declinewas
from almost 70% of GNP to about 60%. examples. In Ru via, the geneal budget was over 60% of GDP in the
late '80s; the targetfor 1992is only35%.
79. Although the figures are not alwaysclear, on the whole it appears plausible that the shrinling of the
national level has been generallypamrlled by a shrinkdngof the subnational
level of government:in Hungary, for
instance,subnationalexpenditurefell from 14.3% of GDP in 1988to only 10.4% in 1991. Predictingfuture trends
is difficult becus of the changingnature of the intergovernmentalsystem in most of these countries.
80. On one hand, as already mentioned, central govenumentscontinue to shift importantexpenditureitems

'downstairs' in an attompt to balance the central budget. It is not clear if the final result of this process will be
reducedtotal expenditures,increasedcentral trasfers, or increasedlocal taxes - althoughsome combinationof the
first and second possibilitiesseems most likely in most countres. On the other hand, as a rule public enterpnses
continue to provide a number of basic local public services: they support the construction and maintenanceof
housing, keep kindergates and pre-schools,and so on. Larger entrpses may even have their own hospitals,
retail outlets, and other non-productionfacilities. As privatizationreforms impingeincreasinglyon the enterprise
sector, pressurs to replacesome of these activitiesby local budgetaryoutlayswill undoubtedlybe strong.
81. Thus, at the same time that localities may be reducing expenditres by shedding unnecessaryand
unproductivesubsidies(for example,to housing)as part of the economictranition to a market economy,they may
have to take on some expenditurespreviously caied out by enterprise. Whether such enterprise functionswill
accrue to nationalor subnationalbudgets is hard to predict: it is equally difficultto predict the net change, in the
short term, in the size of the local government sector. In Russia, as one example, the non-production
responsibilitiesof enterpnses are, virtually across the board, likely to be taken on by the subnationalsector, thus
substantiallyincreasingits expenditureresponsibilities. Indeed, such a transfer is de facto already taking place
owing to the financialdifficultiesof many enterprises. Unfortunately,no one has either quantified the dimensions
of this problem or envisaged any workable solution for it. In principle, however, the shift in expenditure
responsibilitiesfrom enterprisesto local govemmentshouldbe matched,at least to some extent, by a corresponding
change in the revenuebase of the local govemment. Otherwise, the budgetaryconsequencesof privatizationare
likely to be very negative. But as a mle, that revenue base is already inadequatefor the tasks demandedof it, as
discussedbelow.
- 22 -
Table 3
Reformand the Size of the Goveommt Sector
(Expenditue as % of GDP)
Country Exp Sub-nationl Sub-nio Total Exp. as Sub-national Sub-national

as % of Exp. as % Exp.u% % GDP Exp. a% Exp.u % of
GDP of GDP of total of GDP Totia Exp.
PrE-1989 Post-1989
Hungary 62.7' 14.3' 25.4' 57.4 10.4 18.1
Poland 49.7 14.7 35.3 40.1 4.00' I0
Romania 45.1T 3.64 11 24.67 6 10.8'
Bulgaina 55.2 N.A. N.A. 43.0 25.0 23.0
CSFR 58.46 20.5 34.5 60.1 20.2' 34.3'
Russia* 1.0' 20.713U 16.0 1L/ 41.0'r 17.0 43.0
China* 271r- 14.3 53.0 22.8 N.A. N.A.
Vietnam 11.3 3.4 31.0 12.3'2 3.9 33.0
Notes: (Seo end of paper).
82. In the past, goverments m the socalis counttiesemployedfixed prncs (includingrents, and other urban
user fees) and wage controls, as essentialelementsin thOeir dstbuti tool kit. Now, however, as part of the
reforms currentlyunder way, the tansitional economicsa required ind to 'get the pices right and to use tax
policies and targetedsubsidiesinstead. As a rule, replacingconmr price subsidies by targeted sbidies (with
a much smaler budgetaryimpact) will requirethe developmentof new infmaion and twcng systems.W Some
observersthink that the targetingproblem willbe relativelyeay in the transion economiesbecauseof the existence
of a large 'control' data bas in such countries. This hope is probablynaive, howevr, given (1) the flawsin that
data base (e.g. in Poland severalmillionpeopleexist *off the books a far as officialdomis concemed; in Hungary
and especiallyRomania, thern e importantlow-incomeethnicgroups Ohat ar smilarly lusive);(2) the dependence
of officialinformationon a local controlsystem that is diintgatng apidly, if it has not already disappeaed; and
(3) the increasinggmwth of the euphemistically-labelled 'infornmalsector.'
83. Indeed,since the breakown of cenal planning,ctrl governmet in most countrieshave surprisngly
little informationeven about the activitiesof local goverments. Only in a few countries (e.g. Hungary, Poland)
do local governmentsnow appear to provide detailed financing reports to the conter. Reporting of data on
extrabudgetaryfunds and borrowingis also deficiejti
84. Public sector pricing is likely to be one of the last areas to be reformedin most countries. Whilenational
price reforms have liberalizedmost private sector prices,1 at the subnatiouallevel governmentsare havinga much
harder time adjusting prces for public sevioces,even for prvate goods provided by the public sector, such as
housingrents (and such ancillaryservices as water and heating) and transport fares. Adjusting these prices is of
I' For useful discussionsof some of these problems, see the followingpapers presentedat the recent World Bank
Conferenceon 'Public Expendituresand the Poor: Incidenceand Targeting': Milanovic,'DistributionalIncidence
of Cash and In-Kind Transfers in Eastern Europe and Russia;' and Jarvis and Micklewright, 'The Targetingof
Family Allowancein Hungary.'
-LPherehas beenmuch discussionabout the 'big bang' vs gradual approachto price reform, but even in countries
as committedto gradualismas China there has been major price liberalization.
- 23 -
course essentialboth for efficiencyard to enhce local rvenues. Doing so, however, implies major changes in
the cost of living and in welfiae, so tat changingsuchpublic sctor prie will entail major, and undoubtdly very
unpo ar, distributionalshiftsif not pro)pey coordinatedwith other refoms. Such shifts, if too wrenching,or too
quick, could upet the filo democcies tat havo only just emergod in most countries.
85. At the -m tm that fiscal i-a is going on, national tax ytems are also undegoing major
changes. Major naionl tax reformshavo been initiate in almwt every country in a effort both to make their tax
sytm more compiblo wit thos of a market ecnomy and to pwvide a soeu source of revenues to finace
essentialste budgetaryoutlays il the face of prvation. While there are considerablevariationsof dotil from
countryto country, as a rule tax reform involves:(i) the replaeint of profit remittancesto govemmentas owner
with corprt income tuation (CIT) to the govement a tax colloctor, (ii) the introductionof a value-ddd tax
(VAT) to replac the classical socialist *tuover tax", which is bet described as consisting of thousands of
commodity-specificrates, or wedgesbetwoe adminisratively-st rtail and wholesaleprices; (iii) the introduction
of personalmcometaxes (PMl)- wage controlswoe the 'mplicit tax' in tho pre-form era there worealso payroll
taxes, which are geneally retained; (iv) the itroduction of propety and land tati01oni rdimentary form; and
(v) the eliminationof a number of ts aimedbasicaly at regulating nterprise behavior (such as the excess wage
tax, wage bonus tax, excem profits tax).2I
86. As theso national level taxes ar being revamped, the system of local fimucing is smultnewouslybeing
changed,both intentionallyand to some extentuintendonally. In Hungay, for example, the VAT and PIT were
introducedin 1990,the year the LocalSelf-GovernmentAct was passed. In Russia, tho CITwas introducedin 199,1,
and the VAT in 1992,the sam- year that the redsign of the intergovrnl fiscal systemwas to be implemented.
In Romania,prepartions are underway for the ino t of the VAT in 1993,and of a global personal income
tax in 1994,with initial local govemmentreformshaving taken place already in 1991. In the Czech Republic, the
VAT is also scheduledto come into offect in 1993, dieo -m year in which a now system of intergovenmental
finance is supposedto be put in place. In Bulgara, the curret VAT implnion date is July 1, 1993MI.In
all of these cases, there appears to have been almost no consideation in the basic national tax reform of the fiscal
needs of the local governments, even though it is clear that, given the fundamentalinadequacyof the revenues
specificallyassigned to local governmentsin most countries (see below), thes govemmentsare likely to be
dependenton some form of 'tax shaing' in some form for yeas to come.
87. As these press become manfest, furtherreforms are assigningto, or sharing with, local governments
taxes whose designis new, whose operationis untestd, and whoseadministrationis inevitablyweak, as discussed
later. Moreover, even though the national level reforms are ofte intendedto be 'revenue neutral', there is little
basis for revenueestimation, andhence little certity as to expectedyields. The potentialvolatilityof taxes shared
or assigned to the subnational sector is compoundedby these countries' overall economic cycicalityll. The
difficultyof ensuring corresponden=cbetweenthe expenditureresponsibilitiesof the local goverments under the
new systems, and their assignedrevenuebase or share, is self-evident.
I'For fuller discussion, see Tanzi, Fiscal Policiesin Economiesin Transitin

1 In Bulgaria a comprehensive tax reform under preparation compnses draft profits tax law, VAT, tax
administration,territorial division, and excises.
J2See D. Go, 'Revenue Uncertaintyin the Transition,' CECPE, RPO# 27118, forthcoming,1992.
-24-
Stabtization: MemMacoeconomic Contextfor D ion
88. The economicreform progms in the trasitioal economieshave had to addre both and
libsaip- n concerus. Integovenmntal fiscal roforms are tWringplace in the context of a w *k overall
nacroeconomicpositionand often a weak cntral govemnmntfiscalpostion. The move to prvate epr p
has contributedto a more vital economy,but hasaccenuad thoproblem of a shriking rovenuebase at the national
level, as many of the smaller new private enterprises elude the grasp of tde tax net and ptesnt major challenges
to complianceand tax adminisa . The filing state-ownedentpise sctor no longer prvitdes th revenuebase
of earlier days. The tax refoms dencribedabove are undoubtdly deirable for maket effici
ency,
but unt
experiencewith the new systemis gained, reven shortfallsseem all too likely.The expediture side of the budget
continuesto be burdenedby heavy outlaysfor both consumerand producersuidies, as well as often generouscash
benefit program and now and sometimes large outlays to support failing enteprises and to restucture or
recapitalizebanks.
89. In Hungry, the consolidateddeficit of the general gove_nxt reaed 4.3 % of GDP, with inflationof
32%, in 1990,tho year tho new system of local self-governmentwas introduced.In the RusasiaFederationthe fiscal
deficit is variouslymeasured as either 15% or 30% of GDP in 19912W, and a likely deficit of over 10% of GDP
is expectedfor 1992;inflation is soaring. In Romaia, the mesurd budget deficit has been rather small, but an
inflationzateof 161% in 1991 reflectsan unstablemacro-economy,and suggest that the fiscalaccountsunderstate
the deficit: the centralbank carres out heavy qasi-fiscal opeaons, and the banking system financesexpenditures
that are fiscal in nature. Only in China, with its gradual economicreform (and continuedtight politicalcontrol),
is there rough budgetarybalance and comparativelymoderateincreass in the prce level (See Table 4).
90. For these reasons, stab and overal

o acro-eco ic concoms often domiate the nationalagenda.
Reducingfiscal iwn%aances - both those at the center, and those potentiallyemaging at the subnationallevel, is
invariably a fimdamal rquiemt of the adjustrnt progrm to which most transtional counties have
agreed. Thecurlmt stuation provides a shp contst withpublic fic in the pro-rform penod, when the
uaitary fiscalsystemimposedclose adherenceto the Plan in the implementationandexecutionof the budget,which,
together with stict controls over the financs of enterprses helped sure broad balance in public sector
finances.A' In paticular, under the old rgim, the finaces of subnadonalgovemment had few macroeconomic
consequences.Revenue sharing served only as an administrativedevice to simplifya system of central resource
allocation.Expenditureswere guided by planningnorms. The result ws tat in an accountingseam at least the
budgetsof the suboationalgovernmentswere alwaysin balance, with any necessaryadjustmentsbeing madesimply
by accountingtmnsfers.
'Ihe difficultyin measring the Russian deficit relates inter aLia to the tretent of the Union expenditure
responsibilitiestakenover by the RussianFederationbeginningin November,1991,followingthe dissolutionof the
USSR. The higher number cited rofers to a 'notional' deficit, which assumes that Russia had financedUnion
expendituresfor the full year.
2"As discussed in the present section, ther are clearly potential macroeconomicproblems that may arise from
suhnationalfinance. Nonetheless,it is importantto note that subnationalgovernmentscannot on their own cause
such problemssince they cannot 'print money (unlessthe centralgovenmet lets them do so e.g. throughlocally-
controlled banks): in the end, it is invariably the validationof subnationaldeficits by central governmentsthat is
the proximatecause of national deficits and the ills that so often ensue. In this, as in -any other ways, national
govenments ultimatelyget the local governme they deserve.
V Of course, one should not exaggeratethe robustnessof the budget in pre-rform times, since there were clearly
many fiscallyunsustainableelementsfrom a long-termpoerpectivein most countries.
-25 -
Table 4
M nmi Indicators
Country Deficit as a Subnational laflation Current Account M3 growth

% of Deficit as a (1991r Balance as % (1991-2)
GDP (1990) 9 of GDP of GDP(1990)
Hunguy -4.3 -.4 32.0 +1.0 l. 29.3
Poland -2.4 N.A. 249.0 +4.8 45.0
Romania +0.9 +.3 161.1 -12.4 100.8
Bulgaia -13.0 _/ -.8 334.0 -2.7 N.A.
CSFR -2.0 I/ +.0 54.0 +2.7 26.7
Russia -20.6 N.A. 382.0 -7.2 I/ 75.6 I/
China -1.8 N.A. 5.3 +3.7 26.7
Vietnam -4.1 N.A. 36.4 2/ -2.0 19.6
Source: Recent EconomicIndicators,IMF.

1. For 1991; 2. Average inflation for 1990. The end of period inflation for 1990 = 67.2%; 3. For 1M.
4. For Bulgariaa more meaningfulmeasure is the cash deficit (net of foreign interest arreanrs)whichwas 8.5%.
0 Most receat year after price libmlization.
91. Shiftina Ex&enditureResponsibilitiessDownstairs". Current stabilizationconcoenshave led the central

governmentto view fiscal decentaization as an opportunityto reduce centra expenditures. This is being done in
two ways: First, by spinningoff expdiiture rsponsibilitiesto the subnationallevel, thereby reducing its own
deficit, and second, as an opportunityto reduce fiscal tnsfs, purpordly to make local governts more
"idependent, but with the welcom side offect of reducing centl outlays. In paricular, a number of countries
are trmsfernng icresing responsibilityfor social expendituresand the social safety net to local govemnt. In
Hungry, for example, the responsibilityfor "welfari expeditures was tansferred to the localities. In Russia,
the contral governmenttrmnsfered social expenditurs equivalentto some 6 % of GDP M992) to the local level by
this means, with the objective of "pushing the deficit downa, presmably in the hope that the subnaional
govemnents would perform the politicallypainfiul utting required - even though the demand for these services
is likely to grow with the worning conomic situation. More rcnly, responsibility for key national,
interurisdictionalinvestments (e.g. in transport)-has al8obeen transferred to the subuationalsector.
92. Even some of the extensive asset tansfers to local govenmments that have taken place in most countries
appear to have been partly motivatedby budgetay concerns. Some of tho 'assets' transferred - notablyhousing
and some entepises - a really liabilitiesgiven the heavy burden of maintenanceand operationof these units and
the fact that rentalincome (lastadjusted in the RussianFederationin 1928!)does not cover evena smallpart of thes
costs.
93. ReducingUntersovemmentalTansfers. The fiscal difficulties at the national level have also led some
Finance Ministriesto focus on reducing intergovernmentaltransfersas one way out. Transfers have been seen as
the ,compressibledimnion" of the federal budget. The principleof "budgetay independence"mentionedabove
has been interpretedto mean that subnationalgovernmets shouldbe financialyself-sufficient,whichin turn implies
that direct transfersshouldbe reducedand even eliminated.In reality, however, in most of the transitioneconomies
centraltransfer to local goverment sector are very large, reflectingboth the rudimentarytax base availableto local
govemnments and the center's reluctance - again to avoid fiscalvulnerability- to part with any of the major tax
-26-
basesand to assignthem to the sublationdllevel. As a nle, however, the amountand distributionof thosetrasfrs
is determinedeach year on a discretionarybasis in accordancewith central fiscal exigencies.
94. Whilesuch budgeta flexibilityis clearly desiable from the cental govenment's short nn view, it is
a mistaketo view t fers as a completey *comprusble' poion of the national budget, as appears to be the case
in countriessuch as Hungary and Romania. Many of the soevicesprovidedby ubational goverments, particularly
in view of the 'pas-down'phenomennnalready discused, contitute estia expeditues for politicalstability
or for futre ecomic deveopmt There is no way that th mny smalllocal govements that havebeen crated
in most countriescon finace the provisionof thoeoservce at an adequatelevel out of their own rsouc. Even
from a short-nm dtbilizaton prspective, an undefunded subnaltionalsctor in the curmt environmentis all too
likelyto resultin a sitution in whichthe only way local governme can copewith budgetarypresre is by using
economicallyundesiable sources of revenuesucha profits derived from the exploitationof incomeming assets
transferred to them and from diret public ownershipof local busine . At the same time, local governments'
open-endedexpediture responibility for 'social asuistmce' is liklUyto sult in emorgencyrecurrec to the ceral
govnent for additional funds, or simply the unnal accrual of aruas tough short-term borowing.
Arrears have been a problem in China in the recent past and are cuntly a major problem in Russia. One way
or another, intergove n transfersare thus likelyto be an importantcomponentof the central budgetfor years
to come in most transitionalcountries.
95. Deficit Control . Placinglimits on the subnadonaldeficits is of course an obviousway in which centml
governmentscan ensure that the subnationalsector does not give ise to additionalor unexpectedm nomic
pressures. Such limits are in place in Russia and China, for example. In an overall framework of hard budget
constraints, such provisionsmay make sense, but this is unlikelyto be true when budget constraintsare still "soft*
in other sectors, notably the aeteprise sector.
96. The combinationof controls and soft constants may give rise to perverse outcomes.WIn response to
the apparent surplusin overall local budgets, for example, Russiahas, as alrady mentioned,sought to squeezethe
subnationalsector by tranferring expendituresdownwards, re-adjusting tax shares and minimizingtraDfers.W
Since local governmentsby law cmnot mn deficits, however, nor for the time being can they borow even for
liquidityreasons,there is almost by definitiona 'surplus' at least sufficientto meet the *cashon hand requiements
of monthlylocal outlays. Reduced revenue shares or transfers can never eliminatethis surplus: on the contrary,
since the systemcanot be in deficit,such measureswill only lead to measued expendituresfar belownormal levels
and increasingcumulativearrears.
97. Budgetarn Autonomys. Althoughlocal govenmunts in most transitional countries (1ruland,Hungary,

Czechoslovakia,Bulgariaand Romania,for example)have thoirown budgets, quite separte from the state budget,
this nominal independenceby no means implies the absenceof central controls. In Romania, local budgets are
subject to the central govemment's implicit approvaW. The central govemmenthas the final say, especially if
k'For examplesof dysfunctionalincentivesfor economicliberalizationas a result of fiscald lion policies

in earlier Chines experience, see e.g. Wong, 'Fiscal Reform and Local utrialization, Modern Clina April
1992, and *Central-LocalRelationsin an Era of Fisal becline,- China Ouarterlv, December 1991.
WThebudgetsurplus, meured on th report csh basis" for the subnationalsector amounted to 1.2% of GDP
in the first quarterof 1992. See Walich et al; IntergovernmentalFinancein the RussianFederations; forthcoming,
1992
to boththe Lawon LocalAdm

W(According ion and the Lawon PublicFinancestheauthoritythatapproves
local budgets is the municipalcouncilfor municipalbudgets,and the district (judet) council for the district budget.
Budgetshave to be balancedby law. The role of the prefect, as reprevntstive of the State, is to ascertain,ex-post,
that local govemments abide by this legal requirment. Ihere is a disguised direct control of local goverment
-27 -
the budgeti unbwaed. In conbt, in Russiasc budgetry approval is implicit in the systemof negotiatedtax
sharingand transfers, whose levels ar conditionalon an approvedlevel of expenditues. On the other hand, China
and Vietnam slill have a unfiedbud, in whichcentraland subnationalplans and accountsae jointly presented.
In China the local budget must be approvod by the provincial goverment, and the budget of the provincial
governmentmust be approvedby the cewr.
98. More generally,eventhoughthe localgovermts have been givenautonomyin pnnciple, in practicethey

lack it in my areas The lalc of tnsparency or claity of laws further e rbat the problem. e rel degree
of expenditureautonomyof localies is often limited. In China, for example,local gove_nts have little formal
or legal independencewith rospectto etheorlocal taxes or the level and compoition of expenditu. In almost all
the countri under eview, ce expenditur noms and enl spendingand wage mndats continueto Carry
weight in many u. Tndee, in Roumaiathes facton stil weigh so heavily that thero is hrdly any real local
expenditurediscreton as yet. In the Czch Republicand Hungry, on the other hand, ther appears to be much
loosercontrolon the expenditure ide. On the oter and, the lack of veue atnmyis seere
w eveywhore. The
center usually stipulates caps on local taxes and in some cas even spocifiesthe tax rates and tax base.
99. Tax Administrationas a Sowce of Vuleabil . Fiscal d talion, and the politicaldecentrlization
that accompanies it in this now environmentha left some cOentl governments c -o caly vulnerble for
another important eson To ensure macroconomic stabiliation, the feded goverment must be able to contain
its budgetarydeficits, whichmeansrelyingon its ownbudgetaryrevenues. In some tansition economies,however,
tax adminition is decenulized, that is, aubnaional governmentscollect revenues on behalf of the center, and
transfer them upwards 'we Table S).PI In Russi, in the CSPR, in China, and Vietnam, all revenues other than
customsand trad taxi ar collectedby local a ons. As the experience of the former USSR (and the
formerYugoslavia)suggstb, delegationof collectionresponsibility,with its potentialfor non-compmiance,or worse,
relianceon contibutions from member publics, are only vible meansof fincing the centrl govermmentso long
as central political control reman strg (as in China and Vietnam). The potential withholdingof tax revenues
from the fedeal budget threatensRussiatoday; even in Chinanon-compliantloCaladministrtn, and the revenue
contracing' system this gave rise to, have been the source of m_cnomically cyclicalfederal revenues. 2-'
expendituremagent becauseall expenditur involvingthe use of centrl govenment resourcesrequres the ex-
ante approval of the Miistry of Finances' local trsury office, and, since fiscal tnsfers representup to 90% of
local resources, the Govemment, legal arrangementsnotwithstanding,continues to actuallyexert a close to total
control.
3 Strictly sealing, the tax dministrationsam doconconat arms of the federal or central administration. De
facto, however, tax officials may have strong ties and loyaltiesto the local governmentswhich often provide their
housing and fringe benefits: ther is conideablo evidence in some countries (e.g. Russia) that local tax offices
respondto local policy.
UvSee Szaparyand Blejer, 'The EvolvingRoleof Fiscal Policyin Centally Planed Economiesunder Reform: The
Case of China,- IMF Woring Paper 407, 1989;and Bah and Wallich, "Intergovenmental Finance in China".
World Bank WoDkngPaPr Series No.863, 1992.
- 28 -
Table5
Tax Admunistraton in Trulstion Soci;at E3c9;lp
Country "Inverted Centmalzed

Hungary x
Poland xi
Romania x2
Bulgaria x2 l
CSFR x
Russia XI
China x l
Vietnam x
Notes:
1. In Poland, all national domesic taxes, and some local taxes and fes are colUectedby treauy officesand then
transferredto state and local budgets. Other local taxes and fees are paid directly to Gmias.
2. In Romaniaand Bulgaia, the local goverments do not have their own administrton. All local taxes are
collectedby local branches of the MOF.
3. The State Tax Service is in the proce of centalifzing.
100. liDcations for The Lonuer Tern IntergovemmenblSvstom Althoughintergovernmentalfiscal reform

is necessuily takig place in this strned macroconomic contextand hence inevitablyreflects changingshort-term
fiscal needs and psures, care must be taken to prevent the stabilition objective from fully dominatingthe
redesign of the intergovemmentalfinances megime. A major challenge facing the trmnsitionaleconomiesis to
developan intergovernmentalfiscal systemwhich is an optimal combinationof mles ad di - one dbt wiU
be both flexible enough to be compatiblewith macroeconomc stabdilizationand the major structural shifts which
are taking place in the economy,and at the sam time providea stable frameworkfor the effectiveoperion of both
central and sub-nationalgovemmentsin the lcnger termL
101. One solutionalong these lines that has beenadoptedin some market economies,for example,pivotsaround
a systemof intergovernmentaltransfersin whichthe totl is flexiblewith cbanging macreconomic circumstnces -
for example, somepercengta of total centraltaxe - while the distributionof this total amongdiffent subnational
govemmentsis Oformula-driven",with the amount received by different localities dependingon such factors as
potential "fiscal capacity' and assessed need. This approachrepreents a compromise,in which the center gives
up some degree of control over its revenue but also insulate itself from ad hoc and possibly escalating demands
from localities,while localitiesavoid discretionarycutbacksin local trasfers to meet stabilizaion objectives,with
all the expenditure dislocations that this implies. In most of the countries under review, thus regularizing the overall
size and distributionof the tanmsfer,like strengtheninglocal finances by broadeningthe local tax base to improve
local tax yields, represent high-priorityreforms. This point is developed further below.
- 29 -
m. WHO DOESWHAT? THE ASSIGNMENTOF EXPENDlTURES

The Efficieno Cae for Deeentized Local Gt.
102. Whilo local governmmntsmay have some effect on stabilizationpolicy and some role in distributive
policyLW, their mn-oreconomicrole is clealy with respect to the alocation of resources. From an efficiencypoint
of view, the basic rulo of expenditureasignment is to assign each function to the lowest level of government
consistentwith its efficientperformance. So long as thm are local variationsin tses and costs, there are clearly
efficiencygains from carrying out public sectoractivitiesin as decenraized a fashion as possible - and these gains
are larger the lower the price elasticityof domad for these services.
103. In principle, therefore, the only services that should be provided centrally - or, in some instances,
regionally - are:
* serviees for which there either are, or are for overching political reasons
presumedto be, no significantdifferencesin demandsin differentlocalides(e.g.
nationaldefense; public health);
* serviceswith substantid 'spillovers betwoenjurisdictionsthat cannot be handledin some

other way such as by contracting,or by grant design (e.g. interlocal tmnsportation;air
and water quality); and
for which the additional costs of local adminisraon are sufficiently high to
smervices
outweighits advantages(e.g. administrationof income taxes).
Apart from the last of these three cases, eventhese soervicscan often be deliveredmost efficientlyat the local level,
althoughtheymay wellbe financedin wholeor part by centraltransfers. In short, in principle,moatpublic services
should be deliveredat the local level, with local decision-makersdecidingwhat services are provided, to whom,
and in what quantityand quality. To put this the other way, the only servicesthat should be providedcentrallyare
those for which there are no differencesin demandsin differentlocalites, where there are substantil 'spillovers
betweenjurisdictions that cannot be handled in some other way, or those for which the additionalcosts of local
administrationare sufficientlyhigher to outweighits advantages.
ExpenditureAssinment in Practice
104. As noted earlier, under the previous socialistregime, since local governmentsbasicallyacted as units of
the centrlized administration,local expenditureswere includedin the unified budget of the central govemment.
There was no question as to who did what. the central govenment did everything, sometimes directly and
sometimesthrough its local administrativeunits. Under the new local government laws in most of the countries
underreview, however, specificexpenditureshave beenassignedto the local level of governmentas shownin Table
6. In the Czech Republic, Hungary and Poland, for example, local goverments have detailed expenditure
responsibilitiesin the area of education (primary and kindergarten), transportation (local or urban streets),
environment(garbagecollection,disposal,industrial waste), housingand related services, and health and welfare.
In Russia, on the other hand, there is no legal definitionof expenditureassignmentsat the local level, exceptfor
WSincelocal governmentsare governmentsas well as service agencies, they are inevitably interested in the
distributiveas wellas the allocativeeffectsof theirpolicies. Incomeredistributionat the local level may be swverely
limited by the openness of the local economy, but if a goverment is not concenied with distribution, it is not a
government at all but simply a service agency. Nonetheless,within-jurisdictiondistributionis ignored here as
unimportantrelative to the importanceof setting up the basic mles of the local finance game so that each local
govemmentis fully accountablefor its actions.
- 30 -
Table 6
Expenditure Assignment to Subnational Authortiei_
Hungary Poland Romnia CSFR China Vietnam Bulgara Russia

Defence No local No No No No No No Military bousing
responsibility
Justicetatemal Enforcement of No Public No No Secwity of

security rightsof national securityis farmingesas
and ethnic provided by
minorities local
branchesof
the Minidry
of Interior
Foreign Econotnic No No
Relations
Education Ptimary and Kindergarten None Partid All schools A Several special
secondary only and pre- tesponsibility in primary througb expeadisues vocatiol
including day care elmentafy the Czecb university level. (epil and schools. Wages,
and high schools chools only Republic Local current) of operation
governmenta pnimary and contmction and
cover all salaries, econdary maintenance of al
mainten nceof schools. Sone primary and
school fciliies, kindergartens. seondary chools.
adminitration and Some technicd Local enterprises
scholarhips and vocational build ome
schools facilities
Culture and parks Supponting Over-lapping Cultural activities No Some museums
cultural activities responsibility such as cinemas with oblast
for cultura and broadcasting significance. All
activities recurrent
expenditures of all
sport and park
facilities Andal
othcr culural
facilities.
,I01 i i,,i ijsi ti'31 ti .1*111i
_G
0 q8t1 .
_ _ _ _ _ _ _ _ ii _ -_1
_ _ _ _ _ t_X_ _ _ _ __Q
1 _ _ _ _ _ _ _ _ _ _ ii_ _ _ _
^:iL_ 12 10t 1i <~~~~~~~

a IaX,1
_ __ __ _ _ _ __ _ __
l - X
- 32 -
Hungary Poland Romania CSFR China Vietnam Bulgau Ruoo

Libannes Partial No (book No No Lal iaui Speci ibaqy
peogmbliy purchase
thimh the aeviccaat tde
eblM levde an
budgs of the mm l ibay
Miniiry of icesat he
OdhnC ) isyo lve-ld.
Poke ccs No - -- Sofiahasa Road(rafic)
Withdie
Wiliti; ed_er
Sanitation Garbaecollection Ownerhip and Rcflu colection Pat of garbage

garageclection) Provisionof
colgcticuratboth
cold water at
oebb and rayou
__________________
~~~~~~all
levels
levels.
Sewge GarbagecoLection Owneri ad Sewage Gar Moatof the
and eleme pwviaoaof colhecti clectioG and epetcioa
clekanig ewageevceo i bwmln die e _pendluStdoe
se_mea rayon leveland
some op_Son
ezpendiuresat
the
_ _ __ _ _ __ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
levd
Publicutiies (ps. Water Supply me Ditict
electricity,andwar) Wdcarspply Subsdi to
snanaeet and ligh heatingand iDn_uctn. hmboJs (nt
rmuenance of wtder
-________________ public
cemteries _ cep_e) at the
. yonlev
- 33 -
_________________ Hungary _ Poand Roasua CSFR Chn VKtnam Bum Russ

Housing Housing Sha Haiwng No Builliqg _ Ma a of Maie e is de
mgement eqpibilty evices _h ea_ oq of
pasurs sa te evof
OdIu2ScI govammsern
r
Pq.tY e _efi
Fuming. bousing.
buildingsl Cqita
subvdifrin foe expenditures
axr
re"ade iacdeunls
_ _ __ _ _ _ _ _ _ _ _ _ _ _ _
-. athesalse
oed.
Priceand ohw subsidies Rest subsdin Rem Eer Dict subsidies Ysn Mas Danc t For .eels, Muss
uWies to to agrullual I dmug. tnot nd
huehold cooetive, sadyto food sac a
nd pb andsbidies to _me-lge bw bred, alk and
transport eutpst t srvie within afinas tedi
a_
,ubsidies e otbIvolved unicipaities. nyon level. Also
in tde Hatig vabgdy remtsbdies
poducfioof isce 1992.
lbcal ubric
Socil Welfae Socialcare Voivodship Welfare Homles Part obist

ffciies sch as r obilitis expeu for disebledad oVeammt
old e an ede who s- o ad o yand
handicapped chids,
hames
rayon kvel
ophanagcs, die n nt of
disbled dec. pwrgan tanded
byupperlevd
New Public enterprie Can esablih new Can ablih Can ablih Cm establish Can etablis aw Can Can esablish Capacityto ivest
(productivesecto) domesticjoint new domesi ne m domestic domneacjoint estblis ne domestic il joist Vema;nts
vente joint ventu dometic jowt vewn vesra nw joist ventos (pSOpewea
joint domestic of privatization
ventures joint proced if ryo
Al th ayonlee
ths aso IincIud
10percentofany
subordination
- 34 -
Hungary Poland Romania CSFR China Viem Bulgaria Russ

Environmant Proection of tae Envimeat Measue to RexOsibl for
nvkionent prtction Outlays. iwmrne aNd el
wehabiritate
the environmental
envkoseit pblems, e.g.
pri rvationof
_ _ _ _ _ ~~~~~~~foca
State-ownedEnterpriss Major local Local Loca Major local Majorownerhip Local -*Gmup C-
ownerhip ownership ownership ownerhip esponsibities of ownership enterses e.g.
responsibilities reponibilities reqnibiliti rsponsibilities lght and heay respotibi locallight
es indusries. lities inuy, housg
Prvide subsidies condructionan
to cderprisefor foodindury.
technia Rayon
upgding d ronibility
buildingof exis only if dhe
temaporary eepid is
warbousel transftrd to the
loal klvel
Nftes:
i Jo Victnam,localgovernmentc are responsiblefor al of th abov activiti s ey
f fi uderthrdit diction. at, each
klve ofgovem tis reponibl for thaa activiie whichfals und
their juisdiction. / '-' impliesno infonnstion. 3 'No' implis no applicableto ubaniol governments.
35 -
socialprogrms (whichaccountfor a majorpat of expenditure).In effect,traditionand inrtda haveeablished
expenditur ssignmetsi Rusia whichi manycases roughlycorrepond to best practice (benfit) areas.
Unformnatly,pehps owingin partto thelackof legalclarityin ths area, itappearstht thes pincipleare being
inreasngly violated in the courn of the on-goingchages in the ietgovemntal finance system.
105. Althoughin somewaysit is not typical,Romuiiaprovidesan intereng exampleof someof the things
thataregoingon in tho a of expenditurassigmentin tho tnsitioatl conomies.At first glance,forexample,
thenewlocalgovommentlaw appeas to asign localgovenmentsexclqsivetechnicaland financialresponsibility
for manymorefimetions(e.g. policing)thanthosethatare actuallyexercised. In fact,however,thisasgnment
is onlyon conditiontht nationallawsor '[centrallgovemmentdecisions'do not specifyotherwise.Thatis, in
practice,localgovenmentsdo onlywhatthecentraIauthoritieswat themto do. Themainfinanca responsibilities
so far actuallyasumd by local goverm ts in Romaniaare basicallyfor vaious mor activitiesin social
assistance,municipalservice,cultwreandarts,publicinvestment, and maintenance.A Arogars localtransportation
anddistrictheating,local govemments simplyserveas conduitC3r the channellingokcoetralgovernment subsidies
to the publicserviceenterprises(regieautonome) deliveringthere two services. Witththe gradualelimnation of
consumersubsidies (to be in
completed 1993)and the tnusfer of legal ownersip of theseentetprisesto the local
govemments (to be completedin 1992)thesituationmaychange,thoughit is far fromclearhow, finacially, the
localgovermentswillhandletheproblemsstemmingfromthenowresponsibilities.
106. Romanais thus unusualin the extentto whichsomeimportantexpenditures usuallyconsideredto have

are still at the centrl level. In particular,and in contrastto most other
strong 'local benefit' characteristies
transitionalcountries,educationand healthremainthe sole reponsibilityof the contrhlgovemmentin Romania,
reportedlyto ensue a minimumstandardin the provisionof theseimportantservices. Recently, however,
increasingfinancialpressureon thecentralgovenmmnthas ledto theproposalfor someof theseexpendituresbe
delegatedto localities.Asis notuncommon in thetransitioneconomies- thoughhighlyunusualin westerncountries
- both publicscurity and fireprotectionar also providedthroughoutRomaniaby localbranchesof a central
Ministry.
107. Overlapping betweenthecenterand localitieswithrespectto cultureand socialassisance,
responsibilities
on theotherhand,arecommonto y countries.Forexample,localmuseumsareoftenfinancedby localbudgets
and nationalmuseumsfinancedby the centralgovemmentbudget;but booksfor bothlocaland nationallibraries
are purchasedby the Ministryof Culture.Similarly,both localand centra governmentsfundprogramsrelatedto
children.Such' '-i.., withdiffeentlevelsof govnmt responsiblefordifferentsubfunctions
h-h-ring
*xtdit
of particularexpenditureactivities,or with more than one level of governmentplayingan importantrole in
financing,regulating,and providingparticularservices-isa commoncharacteristicof multi-tieredgovernmental
structures:it may sometimesbe efficient,but it almostalwaysobscuresthe key questionof preciselywho is
responsible,andaccountable,for what.
LocalPublicEnterprisoOwnership
108. Romaia also providesan extremeinstanceof a problemfound to varyingdegreesin most of the

transitionalcountries:theimportance andcomplexityof thetnangularrelationsbetweenthenewlocalgovernments,
thepublicenterprises providing localservices,and the centrl govenunenLtIn someways,it is not too muchof
a caricatureto characteriz the localgovernment systemnow functioningin Romaniaas amountingto littlemore
in fiscaltermsthana conduitbetweencentrd government and theseenterprises(or relies autonomies,as theyare
calledin Romania).If one asks: 'What do localgovernments per sodo in Romania?',theanswerat the present
timeis: Not much.' Excludingthe rngj localgovermnment expenditures appearto consistof littlomorein most
casesthan minoroutlayson socialassistance,cultur, and gener administrativeexpenditures.Mostservice
deliveryactivitieshavebeenhivedoff to thevariousmge.leavingthe localgovernment as a placethatdoeslittle
morethancollectdog and
licenses registermarriages. These may be activities,
essential but theydo not require
a largerevenuebase- and indeedin many would
instances best be financedby feesand charges (asdiscussedlater).
- 36 -
109. In principle, of course, thern is nothing necesarily wrong with having local govenments that do not
themselvesdeliver out of thei own resours many servicesto local reidens, just as there is nothing wrong with
havinglocal govenments deliver nationally-finacedand naonay-detmned semvicesto local residents(see lator
discussion)or with having most local sevice activitiescaried out by local public enterises with septo budgets
and managemmats.All these systemscan be foundin one country or another and may work perfectly satifctorily.
Indeed, given the extrm degreoof cenhalizion characteizing the trsitional economiesin the past, the present
system in Romania, one of the Ist of theso countries to reject the old system, may be not only acceptablebut
desirable.
110. Nonetess, it is critical to the financialhealth of both the local govommentsand local public enterprises
that the precise elionship between the two should be set out clealy. At present, in a number of countries tha
lines distinguishingactivities that should be performd by local gov _ents, local public enterprise, and
commercialenterps are undesirablyconfused. Unlessthese questions armdisentangled,senous problemsmay
occur both in setting up properly functioninglocal governmentsand local public enterprise and with respect to
privatizationmore generally. A few examplesfrom fieldwork in several countries illustra the point:
In some areas, local governments are responsible for certain activities (e.g. parks); in others, local
enterprises;
Some activities(e.g.bakeries) are commercialin some areas; others are nru as adjuncts to local public
enteprises;
It is alsoimportantto clarifythe relationshipsbetweendifferantpublic enterprises e.g. at regionaland local

levelsas well as betwen enterpriss and governments,especiallywhen an enterpnse is locatedin one area
and serves other areas as well.
:11. On the whole, municipalenterprises shouldbe confinodto traditionallocal public service activities(water
and sewer, heating [traditionalin eastn Europe, at least], tansit, perhps solid waste collectionand disposal).
All essentiallycommacal activitiesnow conductedby suchenteprises such as bakeries, laundries,and construction
companiesshouldbe sold off as soonas possible. Finally, in someinstancescertain activities(e.g. parks) for which
there seemsno appaent rationale for adoptig the enterpnse form of organiation should probablybe folded back
into the local governmentproper.
112. Where it is decided to organizesome local public service in the form of a public enterprise, any subsidy
paid from generadlocal governmentrevenuesshouldbe set on an ex ante basis to establishthe hard budeet constraint
needed for efficient operation. Moreover, such enterprises should be subject to all normal tax laws (central and
local), and in general they should be run like commercialenterprises e.g. setting prices and terms of services -
although presunably within a natioally-esblished regulatory framwork (which as yet does not exist in most
countries). At present, local authoritiesessentiallyhave eitherno guidancein setting enterpriseprices, or too much
guidancein the sense that these prices are still being set by centrd authorities.
113. The budgetsof local public enterprisesshould alsonot be consolidatedwith local governmentbudgetsper
se: only any explicitsubsidyand any profitstransfer should appearin the goverment budget.Of course, enterprise
budgets should be annexed to the local budget and submitted - not for approval but for information- to the
appropriatecental agency, as noted earlier. Mhenumerous somewhatanomalousdecentralizedinstitutionsfound
in manytransitionaleconomies(e.g. Hlouses of Culture) shouldeither be organizedas enterprisesand treatedlike
other enterprisesor, if theyare to be treatedas componentparts of local govenments, consolidatedwith other local
expendituresand revenues.
114. It is especiallyimportant, both for good govemnment and for good business, that local governmentsshould
get out of commercialbusinessactivities as soon as possible. The businessof local governmentis not business,and
the sooner local governmentsbegin to concentrate on their central task of providing as efficientlyas possiblethe
-37 -
local public servi&astheir inhabitats desir the better it will be for both busine and the local inhabitants.
Although intentions in this respect are cleady good in most countries, there does not apper to be adquate
recognition of the danger of d lintion becoming an obstacle to the privaizton process, as is already
happeningto some extent in Hungary, Bulgaria, Russia, and pehas elswwhere(see later discussion).
115. As emphasized earlier, countries tat decentalize subntal expditure responsibilities to local
govemnmts that either do not have access to adequate local tax sources, or that are not capable (technicallyor
politically) of admnting thos taxes to which they have access, run this risk, especiallywhen sver fiscd crises
reduce central tanse to local governmnts. Whln the major source of governmnt rvenue has traditionallybeen,
and is still to a lar extent, entrrse, the temptationfor local govenmes to extractmore revenue from those
enterprise in thdircontrol can be overwhelming- and will invariablybe most unforunate from an efficiencypoint
of view.
Provisionof Social Services: Shiftinathe Safety Net 'Downstairs'
116. Most publicenterprisesare engaged in providmgsuch 'hard servicesas water, electricity,and transport.
In addition,however, particularlyclose attentionneeds to be paid to the role of local governmentsin te transition
countriesin deliveringand financingservices in at least three important 'people' areas - social assistance,health,
and educadon.
117. First, while the present role of municipalitieswith respect to socid assi varies from country to
country, being qmuterestrictedin Romania, for example, but much broader in Hungary, this are is all too likely
to becomea major problem in all countriesin the futureas a result of two unpleasant but likely developments- the
weakening of the naonal cacity to minin ociadprotection at the present level and the growing need for such
assistance as a result of economic restuctunng. Even if the need for this bottom layer of the social safety net
becomesgreater, however, it is far from obviousthat the municipallevel - while it may be the appropriateexecuting
agency - should be responsiblefor financins such assistance. In many western countries, for example, it was
preciselythe unsusamablyheavy loadon municipalwelfaresystemsresuldngfrom the interwar depressionthat lead
to social welfare becomingmainly a national, rather than a local, responsibility. It is thus ironic, and potentially
dangerous, that so many transitionalcountries seem to be moving in the oppositedirectionjust as the need for such
services is rising sharply as part of the adjustmentprocess.
118. Second, health and education are already controlled in large part at the local level in a number of
transitionalcountries. Experiencein many countries around the world shows that people are willing to pay local
taxes to have their children educated. Hence, carrying out some educationat the local level is one instancein
which a devolutionof expenditure responsibilitymay, at least in principle, be matched by a corresponding
devolutionof financialresponsibility. On the other hand, there are also strong reasons for centralinvolvementboth
in ensuring adequateinvestmentin human capital throughoutthe country, regardless of local fiscal capacity, and
to some extent in financing a higher level of such expenditurethan localitieswould be likely to do on their own
(See also the later discussionof trsfers).
ExnenditureAssinmmentin Princiglo
119. . In the long run, a desirable assignmet of expenditures to local govermmentsin many transitional
economiesmight be somethingalong the lines shown in Table 7, perhaps with the local' share being further
dividedbetweendifferent tiers of local governmentalong benefitarea lines. As Table 7 indicates,thereare only
a very few functions- national defense, purely local services - that should be clearly assigned to one level of
governmentor the other. Most expenditure functionsare likely to be shared between levels of government,as
indicated to some extent by the breakdowns shown for a few of the functions in the table: similar detailed
subfunctionalcategories could of course be set out for e.g. health and tansport.
- 38-
120. It is also importa to unstand that Table 6 refer to the level of govenumW dt a paticular
service. Who finM it is a diffent quesin. Moover, ther is no prsmption that any or all of thee ser
must be Qgrducdby public setor agencies, althoughit is atuud that the public sector ust indeed to some extent
be responsble for thoe
pgMjg of such services if anythingclos to the opbtal mnmt is to ' a provided. Wbat
matters to people in what servces ar provded, not who providesthm. Ilded, may 'public sector' servicesar
ae-tuallycaried out by non-governmet ntities in diffrent comtries.
TABLE 7
A PossibleEx=<diWurAmsin=
Local Center
NationalMuseums X
Local Musem X
Communitycenters X
Social Assistance X X
MunicipalServices X
Transporaon X X
Telecommunications X
Health X
Educatic:n X
Kindergarts X
Primary Schools X
SecondarySchools X
Higher Education x
Protection:
Local Police X
Fire Protection X
NationalPolice x
NationalDefense X
121. To iUustate, aternatives to direct publicproduction by genal-pose local goverments that may be found
in one country or another include:
* Variousforms of contractingwith prvate firms e.g. compeotivebidding for time-limitedfranchise (often

with the local govenment contiuing to own the assets, but turing to prvate sector firms - or possibly
cooperativesof former public sector workes - to manap them);
* Establishingstandardsfor required sevice and leaving it up to consmers to select private vendorsand to

purchase the required level of service;
* Issuingvouchers to (some or aUl)citizens and allowing them to purchase the required service from any
authorizedsupplier,
* Contracting with other government units (e.g. disticts, other municipalities, enterprises, central
goverment) to provide the serice;
- 39 -
* Producig some seres, whilopurchang othm fiomothergovenmentagcies or pnvat firms;

* Estabishingon a voluntarybasis limited(Oneighhorhood)
govenmnts employinguser charges(and
posbly limitedtax powers)to gsneate rveaue;
* Eblisbing limited-purposind dentgovmentl unts cverng momthn onegeneral-pose local
govemn (spca disict) finnced by bnefiay chag of onemo or another,
* Establshing specia tax (imprve_t) ars within a given g -puros loca governt to provide
servicesto areaswhichare willingto pay for thm
122. Of course,the mer possibilityof suchvariedaugr nme to provideparticularsavica doesnot imply
eitherthat thesearangments will be easierto managetha a mr conventionalorganizatonof publicsector
activitiesor that theywill yieldofficientand equitablesemce delivery. Noeltb s, one of the morestriking
omissions in theon-goingdiscussionof localgovemmentin thetranti eonomisis the fidure to keepstraight
thecnticaldistinctionsbetweenwhoownswhat,whoDaynforwhat,and who4 whatLContraryto themeage
manycountriesin transitionsem to have absorbedfrom their prior expenence,namely,that ownership- is
eveything,in thepubhicsectorwhoownsan assetdoesnot necessrilydeterminewhoregules theuseof theasset,
whoopoaes it, or whofinancesit - and thesefa ar, as a nle, muchmormimportantin determiningbothwhat
is done,and whatthe fiscaland otherimplications are, thanis ownershipalone.
123. A striidng feature of local government finances in most trnsitional eonmomiesis the prolifeation of
extrabudgetaryfunds (EBFs). In Hungary, in additiin to the lag national-levelEBFs for social security, housing
and solidaritya, thee are a large number of off-uidget accounts at the local level to which oarmaioe finds
devolved. Such'inherited emarndg' is alsoa promint faotur in th CSPR. In Russ, the rightto set up these
fnds was accompaniedby the assgnmentof a number of non-tax rvenu sou to local governments.Bulgaria
and China so permit suchfunds to be st up, and in the latter thoy conatitutean importantparallel budget at the
subnationallevel.
124. The geness of theso extabudgtay funds is complex. In par they rep t an attempt to escape the
"control mehanism? of tho cantrl govenment. Insmuch as local budgets in some countries are not really
autonomousand remain subject to smutiny from above, EBFs ovade this scruy on the expenditureside. Ia
countrieswhoe the loca govenmments ae transfer- dependent,and whr transfers are allocatednot by formula
but by bargainingand negotiaton, puttig rnsourceaoff-budgetinto EBFs is a meansof concealig the true resource
base of the localityfrom the center. Some localgovernmmatsthemslves claim dtt oxtrabudgetaryfundsreduce the
uncertaintiesof the budgetaryprocess. Moreover,to tho oxtent the rvenue system remainsbasicallyunitary - that
is, the tax base and tax rate of most, if not all, taxes are definedby the center, with revenuesallocatedto the local
level, EBFs may provide the only window for any local revenueinitiative.Lastly, in countries such as Russia and
China, whero local governmentsmust share any revenuesthey coUlct with higher level govemments,EBFs, being
off-budget,are not thus shared, a fact which gives them considerableadded attraction from a local point of view.
125. On the other hand, such funds present serious problem for effective budgetary magement at the
macroeconomiclevel. They also reduce the tanwparencyof budgetaryoperationsand complicateassesment of the
impact of fiscal policy. In practice, extrabudgetay funds function as parallel accountingsystems which, for all
intentsand purposes,basicaly constitutefull paradll budgetsnot subjoetto the stricturesof conventionalbudgetary
procedures. In addition to being a clearly inefficientbudgetarypractice that permits public sector operationsnot
legallyapprovedthrough the proper channels,the loss of controland informationasing from the presenceof such
multiple budgets is bound to undermine the ability of the government to use fiscal policy as an efficient
macroeconomicinstrument.Withoutdiscussingthe specificmerits of each EBF, as a geneal rule legitimaterevenue
-40-
sourcesshouldbe fullyincorpoatedintothoereulr budgetto providefullaccountability of fiscalopaions, and

if this is not possble, fullinformationaboutthe sours and use of extrabudgetayfundsat sub-nationallevels
shouldbe proWvided to the centrd fiscalauthorities.
THROUGHUSERCHARGES
SERVICES
IV. FINANCINGLOCALPUBLUC
126. Th prvious sction notedthatin mosttrational cownomie manylocalpublicsectoractivitiesar caried

out throughpublicetaps. Such terpis ar coma., for example,in trort, in houng, in garba
my sch entrprisesae likely
collection,in watr and sewap, in electricity,anda on. Evaewithprvatiuzato!a,
to continueto existin mostcountrias.Thequesdonof howlocalpublicentepris armto be fiaced is therefore
of criticalimportaneto thehealthof thelocalpublicsector.Theanswersuggestedby thebenefitapproachto local
governmentfinnce outlinedearier is tht as a genal rule, such enterpises should be self-financing,that is, they
shouldbe paid for by those who rceive the servicesthey provide. To put this another way, the first rule of local
financoshould be: Wherever possible, charge.' For efficiency,chages should be levied on those who receive
the benefits: the direct recipients, whether businese or 'thingas (real property) should thorefore be charged.V
127. Although the ultimate incidence of such charges is of no more interest in principle thn the ultimate
incidenceof the price of cheese, studies in difforet countri have shown that the distributive conseqm c of
charging for local public services is not necessrily regressive. In any case, attemptingto rectify fundamental
distributionalproblemsthrough inefficientlypricing scarce local resources is almost always a bad idea, resulting
in little if any equity being purchasedat a high price in efficiencyterns. As noted earlier, however, particularcare
will need to be tken in adjustinglocal public serviceprices (includingrents) in the transitionaleconomiesbecause
of the importantrole such prices played in offectingdistributionpolicy under the old regime.
128. While user charges are most likely to be viewed by hard-pressed local officialsas a potentialadditional
source of revenue, their main economicvalue is actuallyto gmoto economic-efficienv by providingdemand
informationto public sector suppliersand by ensuring that what the local public sector suppliesis valued at least
at (marginal)cost by citizns. This efficiencyobjectiveis particularlyimportat at the local governmnt level since
the maineconomicrationale for local govemmentin the first plac is, as noted earlier, to improveefficiency. There
is thus a presumptionthat, wheneverpossible,local public services should be charged for rather than given away
(unless, of course, they are pure local public goods or the explicit intentionis redistributive).
129. For example, when local govemmentsprovide services though local public enterprises such as water,
power, gas, and public transit, these services should generallybe charged for at prices that will recover the cost
of providingthe service from the uses or buyers of the service either i;mmdiately or over time.2 The price of
water should reflect the cost of piping it to the homes (or standpipe)as well as the cost of mantaining the pipes,
treating the water, and so on. Bus fare should cover the cost of purchasing buses and maintainingand operating
them. In principle, such prices should be set at the competitiveprivate level, with no tax or subsidy element
included- exceptwhen doing so is the most efficientway of achievingpublic policy goals (and even then, as noted
earlier, it is best if the tax-subsidyelementis accountedfor separately).
130. Unfortunately,in most countriesmuch less use is madeof charging at the local level than seemsdesirable,
and many of the charges that are leviedare poorly-designedfrom an efficiency point of view. Weak accounting
IZ'I view of the substantial importance of locally-providedintermediategoods to business, some limited local
taxationof business may thus be warranted,but, as emphasizedelsewhere, it is importantto constrin the ability
of local governmentsto *export' tax burdensto non-residents.
-there are, of course, many well-knownqualificationsto this statement,but it is sufficientlyaccurate for present
purposes.
-41-
systemsthat do not clearly identifycosts is one reason for the prevalet under-pricingof these semces. Another
rason is the historicalhangoverof the previousdistributionalrole of such prices. And a third is the simplepolitical
difficulty that newly-electedlocal goverments have inposing subsmtnt increases n srvce prces on those
who have just electe them.
131. Lca governmentsalso obtain revenuefrom *som&ofes uch as licens fes (mariage, busine, dog)
and various smallcharges leviedessenially for performig specific services - reg ng this or providinga copy
of that - for identfiable individuals. In effect, suchfes costitueost imbement from the private to the public
sector. indeed,in some budgetary systems, such cost twcoveries are neted out and only net (of recovenes)
expenditures sowd Chargig people for somethingthey are requiredby law to do may not always be sensible
- for example,if the benofitof (say) regisation is general and the cost is specific - but on the wholethere is soldom
much harm, or much revenue, in thus reovering the cost of providiog the service in question.
132. Anothercategoryof charge revenueencompasseswhat may be called *wgafic benefit taxes (or chures).'
Such revenues are distinct from service fees and publie enterpise pries becau they do not arise from the
provision or sale of a specific good or service to an identifiableprivate individual. Unlike 'prices' which are
voluntarilypaid - although like 'fee's which are paid for services that may be required by law - taxes reptesent
compulsorycontrbutions to local revenues. Nonetheless,specificbenefit taxes are (at least in theory) relatedin
someway to benefitsreceivedby the taxpayer. In contast to such genel benefit taxes as fuel taxes leviedon road
users as a class or local taxes in general viewed as a price paid for local collectivegoods (see below) - specific
benefit taxes relate to the specific benefitssupposedlyreceived by specific taxpayers.
133. For example, if as a result of a new road, better street lights or a new sewer system, property values
increase, or busine sales dse, a 'benefit levy' might be introduced. This could tameseveral differentforms; (a)
a 'special as nt'; (b) a 'landvalue incrementtax'; (c) an 'improvement tax'; and (d) a 'supplenentary tax',
etc. Most such chargs are imposedeither on the assessedvalue of real property or on some charateristic of that
property - its aea, its frontage, its location. A common benefit-relatedcharge is the developmentcharge (or
'beterment tax') - a lump-sum charge designed to recover the cost of infiastructr developmet from
beneficiaries. Thus, those whose land is near newly-installedstet lightingwould be charged to help pay for the
cost of the iation. A developmentcharge may cover only ono project-for example, a neighborhoodroad
pavingscheme or the constructionof a sewage canal-or it may cover the full developmentof a new area. Such
a methodof financingmay be not only efficient but equitable, since those who benefit from developmentpay for
its cost. Two paricularly successfulsystemsof developmentcharges are land readjustmentin East Asia and the
'valoriation systm' in Latin America. While such shemes are most obviouslyuseful in rapidly-growinguban
areas, there may nonethelessbe some role for them in at least some cities in the transitionaleconomies.
134. In any case, the importanceof charging for public servicesin these various ways is, as emphasizedearlier,
much greater than the relativelysmall amountsof moneymost countries can or do collect from such levies. To the
extent that a local governmentis viewed primarilyas a provider of services, as it should be, and the benefits of
those services can be attributedspecificallyto individualcitizens, properties, or busneses, the appropriatepolicy
is clearly to charge the correct (roughly, marginal cost) price. Only thus will the correct amounts and types of
service be provided to the right people, that is, those willing to pay for them. Correct pricing in this way helps
public officialsmake sensiblejudgementsas to how to matchscarce resources with rising demand. When the true
cost (and the relateduser charge)rises so high that demandfalls, the costs of the service exceed its benefitsin the
opinion of the public. The signal is quick and clear. A decision is then needed as to whetier to: (a) increasethe
price and serve fewer people; (b) cut back on the cost and standardof service to keep it affordable;or (c) subsidize
the servicefrom generalrevenues.In the absenceof such price-generateddemandinformation,supplydecisionscan
only be made by bureaucraticrather than market-orientedprocesses.
-42-
V. THE CHOICEOF SUBNATIONAL
TAXES
13S. I es
asil collectivepublicgoods forlocalreiden
pupose of loal taxs to fineo locally-provided
If such goods a tuly 'public in te ss of accnungequallyto all rsiden of the jurisdiction, Mnif
edisribution to ote tha natna standardsis not aim of local publicpolicy,ad if admiisrative (ad
compliac) cost am leftout of accm , thobe ource of localmven migh prahp be an equalper cap levy
suck theo2LU, whichaWshs dkevitu of beg economiallynautrl or officiet in the senseof giving
to no exces buni
136. At loest an ounty, Huny, aclly has introdced a form of localpoll tax, called eo ta
In view of the high propot of publi housingin Hungay - a proportin dth will undoubtedlydeli butis lbly
to remai sigficant - an argumt c indeed be o for levying aomeform of local poll tax an public hosing
tents.W Altough tis taxmy provemor difficultto administerin Hungary'si ney mobilesocie than
sem to be geneallyu_ lim h a tax, whileit is n r liely to yieldnuch rven, maythusbavoa minor
roleto playin localfinc Fromthispersctvo, however,it wouldbe prefble if te rate of sucha taxwo
set locally rathe imposd unifoimly by n al legiooas in Hunguy. Although such differuatal local
poll toxesare eay to ovadeby movig, tis pobiliy may be viowed - creaing a not undesiable check on their
rates. Of course, venthose who do not floemay be had to tax: the low officincy costs of th, poll tx m likely
to be purhasd at the expns of high admiittve _d compliancecosts.
137. Morefundamtlly, thoa is good n to beive tat in manycoa somelocal renidets - porty-
owners,peoplewithschoolag cbildn, or whoever- benfit morethanors frm theprovisionof localpublic
goods. Thatis, whilethee maybe no ron to levy cfic benefittaxs thre maybe goodrson for somelocal
residents to pay m than otrs
138. If, for example,the deand for local public goods is incom laadic,a benefit cas can be made for a loal
income tx - or, more feasiblypehap given the high adminative costa of separatelocal icome t , a local
surhage on the cental income t& If the enoymet of such goods is amocied wi consuptin (rahe than
reddence), a benefit cas can dmilady be made for a oxl tax an consumption,which would in practice almost
cerinly have to take the form of a rotil sales tax. And finally, if the benfits of local public goods ar enjoyed
in proportion to the value of real property thereis obviously a cas for a local propety tax.
139. Under-flmdnd. .md.EimJIy D _tdm? In practice, th fisca sitaton ficing most central govermnt
in transitional count is such tht vitually no major revenues we ceded to subsional govenments. Many
couns - esily those ith a stonmgregional tradition suck as Rsia and the formor republics- aig the
tax revenues (other than the coporate tax) from locally owned enterprs to the local government, givUg them
strong cntlisntc incetives to protect ad maitai the monopolyposition of ter entepis. Thi is likely to
havecosequences for futre economicgrowth,especiallywhere localitiesarn swcessuy 'protctiniat. In mny
of the countrie surveyed, the strong financialinterest of local governmentin locally owned enterpris rm .
This is a holdoverfrom the old fiscalsystemunder whichenterpnse profitsaccrued to the level of govemnt that
owned them (i.e. profits of locallyowned entuprises weA to the local govenment; profits (and later, taxes) of
centrally owned entprises went to the central government; and provincial govemments received the profit of
their enteises).
tt is of course asumed that any desired general income support is providedprimarly through national sources:
as argued earlier, local goverments shouldnot mispricetheir servces for distributivereasons.
&Such a surchargeis, for example,proposedin Bird and Wallich, 3FinancingLocalGovemmentin Hungary.a and
in Wallich, Fiscal Decentralization:InterovernmentalFinance in the Rusian Federatio.' See also the lter
discussionin this section.
- 43 -
140. Of all the countries surveyed, only Hungary and Russia give subnationalgovernmentsas large share of
certain non-enterpriserevenues. In Hungary, local governmentsget a share of the personal income tax (PMI -
initially 100%, then 50%, now 25%. In Russia, subnationalgovernmentsget some 20% of VAT, 100% of PIT
and 60% of the corporate profits. More generally,in most of the former USSR republics-the present CIS - the
personal income tax is assignedto the subnationallevel, a holdover of the old Union days in which the PIT was
a "localtax' assignedto the oblast level and passedthrough to rayons/cities. Generally, subnationaigovermments
in transitioneconomiesare under-funded,and overly transfer dependent (See Tables 8 and 9).
141. The only 2tentiallv (but not preseatly) significant tax these countries have assigned directly to local
governmentsis usually the 29z txJu,as in Hungary,Bulgaria, Poland, Romania, and Russia. In Romania, for
example, the property tax is levied on all privately owned real estate. The tax rates are differentiatedand, in an
undesirablecarry-over from the old rgime, variedwith the 'social characteristics' of the property owners: 1% on
the value of the property in the urban area ownedby workers and employeesof commercialsocieties; 1.S% on the
valueof propertyownedby individualbusinesslikeindependentprofessionals,self-employedindividuals,and family
associationsof ro more than ten employees;a lower rate of 0.7S% is levied in rural areas. The taxablevalue of
property is reportedlybased on inranco assesmnents,sales contractsand constructioncostsand appears as a rule
to be grosslyunderestimated. Moreover,all buildingsownedby public institutionsare exemptedfrom this tax and,
in another featue seeminglyguaranteedto keepthe yield of this tax low in a country privatizingits housingstock,
all housesand apartments bought by individualsfrom the state are exempted for ten years. Even apart from this
serious structual defect, the difficult and costly task of updating property valuations (and keeping them current)
means that while this tax may become an important source of local revenues in the long run, it would appear
mistakento rely on it for substantialrevenue in the near future.
142. Matters are not much better in Hungary, where the property tax is levied at a more uniform rate - 3 % of
"correctedsales value' (whichusualy meansabout 25% of market value) - or, alternatively,at a flat rate per square
meter. As a rle, the only informationlocal governmentsseem to have in practice on whichto base such taxes is
the area - and evenwith respect to area the exemptionsspecifiedin the law (and carried over from the old system)
ensure, as in Romania, that not much revenue is likely to be collected from this source for some years.3"
Substantialnational assistance in developingan adequatevaluation base seems necessary if the property tax is to
becomean importantcomponentof Hungarian local finance.
143. Another tax granted to local governmentsin some countries has been the old tax on individualbusiness.
essentiallyan income tax on small private enterprises. This tax has been assignedto localitiesin Romania, for
example,on the grounds that it is easier for them than for the central governmentto check and collect such a tax.
With increasingprivatizationin urban and rural areas, it mightappear that considerablerevenuemight be expected
from this tax. In fact, however,even in cases where, as in some larger cities, there are already computerized
systemsof informationon tax bases, tax rates, exemptions,populationand familystructures, it is by no meansclear
that it will be easy to keep these data bases up to date, as economicactivity moves increasinglyout of the direct
control of the govemment.For small communesin rual areas, family visits by tax collectorsare now the most
commonway to collect direct taxes in many transitionalcountries. Again, however, it is unclear how much weight
this system - whichgrew up under and was long supportedby the totalitarianregime - will be able to sustain in the
more democratic, and less controlled, circumstancesnow prevailing. In any case, with the growth of the private,
and especiallythe informal, economy,the level of evasionwith respect to local taxes seemsmore likelyto rise than
to fall over the next few years. In reality, experiencethroughoutthe world suggeststhat the only kind of business
taxes local governmentscan collect efficientlyare low-rate and relatively crude levies.
L There is little to be said for using area as a basis for urban property taxes. The case in differentin rural areas,
however, where taxes such as that in Romania- a differentiatedrate ranging from 35 lei to 400 lei per hectare,
dependingon ferlity and location - may make sense, provided the present fixed per hectare levies are adjusted
annually(or more often) in accordance with an approprate price index. Unfortunately,those who have received
land under the land reform are exempt from land taxes for 8 yean in Romania, thus severely weakening this
potentialsource of local revenue.
- 44 -
Tabl S
Revme A _ime to Submiol Authorties
_____~~~~~~~Ri
Hungar Polan Romania
-_lonkR.
CzechR. Slovak a.asiml _~
Vi etam)_ --
guia
u -__
11,1,inse
(1991) (I991) (19"l) (1991) (1991) (195) (1991) (1991) (19)
Trover tax(erVA) 0 L0 N.I. 100% 100% O /fi 20%
CTruoqte incoAtax 0 X 0 N1 N.I. IOO% 100% 50X/ _ _60%
Peuionl omeitax 25% I/ 30%_/ 0 N.I. N.l. 100% 100% 70% 100%
Tx onPpetty of 0 100% /
95%2 SO IO
Lcally OwnedSOEs -I 0
tO 0 V-. 100%
_ _
Exciwes 0 0 N.I. N.L 0 0 61.1%
Wagetax 0 30%I/ 0 0 13% 0 0 0
Smalland MediumErAJ N.A. 50% 0% N.L N.L 100% N.L N.L NA |
Collves bnome tax
Local Taxe: _
- real es tax 100% 100% 100% 100% 100% N.L NJL 100% 100%
- autolroadtax 100% IOOX 100% N.L N.L N.l. Nl. N1 100%
- s1mabusisa 100% 100% N.A.L N N.L NX. N.L N.A.
- agriculturaltx 0% 1OX% 100% N. 100% 100% 100% 100% NA.
- pontax 100% N.A. N.A. NAL N.A. N.A. N.A. NA. N.A.
NatumlReaauce Tax N.A. N.A. N.A. N.A. NA. 100% N.A. NA. 100%
Odiet touim & mis. Mile. NA. N.A. N.A. N.A. NA. 21 misc.
misc.
taxes
Soume: Misc. BRD domaents. See biligaphy.

Noe: jI Reducedfom 100%(1990);50%(1991).;3/ Unot 1M tax eform.;I/Alo coaidad as propertyincome
tax.;ifTobe abo le in 992.; 11To be usneftmnedistoa mmr effective
PN in 1992and haredaccordingto a 85X (cexral - 15%(ocal) traio.;
_I To be inrodud in 1993.;2/ 100%of ech tax acnus to die le of goveetneutownigth end
by provinciafyownedSOEs accruesto thepvincial budgd.; I/ 40% from unkiipalenterprisesand 10%from eprie; eg. VAT pdd
stente pries.; NJL:No hfomation.; N-.: Not aplicable.
- 45 -
144. Just such a tax ha been creted in Hungay in the local tax on busnor. Unfortuately, this is
a bad tax. Although the pesnt low rate (3 per mill) on business tumover is not likely to do much ecomiic
damage, such a cumlative busuie tax of course poduces preciselythe s- kind of tax cascading an Hugary
has tied to eliminateat the naiooal level by adopting a VAT. This problem will become more srious if, a is
lky, the increaing prsu on local finace leads to inea in tax rates in the future. Moreover, a premuly
designed(with the explicit exclusionof retil sleo fiom the tax base) this tax encoraes both tax exportingand
local attaemts to maipuate the tax sysmI for incentivereasons. Tax exporting is of cours the antitheis of
aional local finance,and local fisca ;entives to productionhave a dml record tumghou the world, so nehr
of thes featuresof the tumover tax is desirable.
145. Another bd tax that has apped in a umber of countries is some form of local Ourist. Touit
establihmt (inluding second homs and cotta) should of couse be suvi*actto geneal local bunes and
property taxes. And of coumesshould be chaged fully for the costs they im on the envinmet ad on local
public services. Buttde seem no reason to encouragefiscal irsponsibility by maling it especia y aractive for
local govemmentsto imprse taxes on noresidm: as a nle, they need to be restrained from such actions, not
encouragedl.
146. Another possible local govenment tax, in contrast, deserves more atenion than it seem to have so far
receivedin most trsnsitionaleconomies:the taxation of motor vehcls. At present, it appers that only about 7 %
of vehiclesin Hungary, for inance, ar subject to the relativelysmall and not particularlywell-desiged *duties
leviedby local gover_ments., Many counties similaly have widepd exemptionsfrom their exing vehicle
taxes e.g exemptingnot only aU vehicles used by public institutionsfinnced by the state but those used by war
veteran and so on. None of these exemptions,of cours, mo.m any sene given the basic benefit rtinalo of this
tax. As vehicleownersip beomes more widespread vebicletaxes (tedesignedto be less vulneable to inflation
and perhap more in line with envirnmeta needs)might in timo becomean importantrevenue source for at least
the larger localities. In principle, ul vehicles should be subjwctedto taxes designed to some extent to offset the
social costs attributable to vehicles. While such taxes should probably be dosignd and imposed uniformly
thrughout the nation to avoid obvious administatve problems,there is no reson why these revenuesshould not
be assignedin whole or in part to the local governts in which the vehiclesar registered.
147. Although both vehicle taxes and espocially the property tax clearly have significant long-term revenue
potentid, particularlyin larger urban areas, on the whole the local tax packge bestowed on local governts in
most transitionaleconomiesleaves much to be desied. Often, for example, local tax rates and tax bae are
estblished in central govemmenttax law, and at most a range of rates is providedwithin which local govenmts
can choose. Many of the taxes thus establishedhave rates fixed in nominal cunmcy, thus making them highly
vulnerableto erosion throughinflation. Others are minor nuisancelevies whichmay weUyield less in revenuethan
they cost to colect. None pi .ides an adeque fihcd foundationfor an active and viable local goverment
WIthas been argued that tounsm taxes (krtaxeonor taxes de tournse) should be seen more as a globalizeduser
charge than a fiscal levy. Resorts with annuallywidely fluctuatingpopulationshave to assume expenditure,both
for operationsand capital investment,of a type and amountthat cannotand shouldnot be recovered from the (often
few) permanentresidents but from the acual beneficiariesof these services. If not, the unavoidableconsequence
is that resort areas turn to the central governmentwith a demandfor increasedfiscal resource transfersand that at
the end tourismactivitiesget, in one way or another, subsidizedby the centralgovenment as, for instance,happens
in Turkey.
ain Romania, in another undesirablehangoverfrom the old regime, initialy, some motor vehicle taxes (levied at
flat ates, by type of vehicle) went to the central govenment, depending on the nature of the taxpayer. Since
August 1992,these now go to municipalities(individuallyowned cas) or to disticts (cars owned by corporations).
-46 -
148. In momscounries(meTable5) at preen, all theceal govenumt ta ar collectedby a _ dJj

tax Althoughsich dualloyaltiescanobviouslycaon problems,it soem unlikluythat bhvig locl
taxescollectedby localgovernzmntwillhelpmtters muckin the r fture, givente sevee so of trid
taxadui_nor. In Vietnam,theMOFha eblished a Geeral TaxDeatm_t to ovetseetheaciVitie of a11
provincil tx d_etmu. in Chinatha amrtwoorguai-mn: theTaxBureu andtie Fiac Depatmn which
havesparae d with detiled functio at eachlevelof city, countyand province. Rom is ao an
_atmt
theprocos of crating twoorgatios to hndlo fince and tax admintaon at thelocal a woea ntn
level. in Russia, tax admnitration it responsibilityofte Stae Tax Semrice(STS) whichhu an orgaional
sucur at all throo lvela of govem t, andis an saunomou agecy with initial rning.
149. A *uy JItrx miht pehop be dfnd one (i) aseosd by local gover t, (ii) Atrte decided
by local gvota, (iii) collectedby local govanins, and (iv) with its poceeds _auingto local gov bnts.
In the ral woldd,however,manytaxespossessaoly one or two of hes charactstics. In sm cases, for
oxample, a tax accros in wholeor in par to localgovemnts, but its re ae sotby thontionl gpv _nit,
whichalso asoss nd coUects it. Insted of a local tax, for many puposes suh a tax mightbe conidered to
be reallya cental gvenint grant allocatedto localgove_nnts in prportan to the amount of natiOnain)ome
tax colectedlocally. On the otdr had, whatlookslike a cntral tax and a rlated tafr proa m reyaly
be a localtax. Somei ovrl trasfe, for example,in effect smplyrer taxesto theregios in which
theywerocollectedin the firstplace. If thelocalgovermet detmn thetaxbaseand rate and receivesall the
rvenues, the only role th cntal goverment plays is a a collection agen: it may even be rembud for its
worc, for ince, by beiongallowed to reta a sal percentage of coUctions. Proe bly, the cntrl
gover_it ha a comparve advantae in tax collection,and the local govenunet has conacted for its service
in this reopect. In this case, thereis no intergovenmntl tranfer at all, exceptin tho narrowest accountng ss.
Characteristicsof a Goo Loca Tax
150. The followingare amongthe chaacteristics that may be sought in an 'idedl local tax:
[1 Thetax base shouldbe reltively immobile,to allow locd authoritiessoms loony in vaying rates
without losng most of their tax bas.
[21 The tax yield should be adequate to meet local needs and sufficiently buoyant over timo (i.e., it
should expand at least as fast as expenditu).
[31 The tax yield should be rlaively stable and predictable over time.
[41 It should not be posible to expost much, if any, of the tax burden to non-residit.
[51 The tax base shouldbe visible, to aesur accountability.
[61 The tax should be perceived to be reasonably fair by taxpayers.
[7 The tax ould be relatively easy to admns efficiently and effectively.
Not everyone would agroe that all those charcteristics are necessarily desirable - e.g. is it unequivocally good that
local govements should be insuladd from either the tax base consequcs of their tax rate choices or from
inflation? In any case, it is unikcely that sufficient taxes possesng most of the c _aesticswill be md
available for locl use. In fact, as already noted, the so-cailed "local* taxes in most trnstion countries are
controlled in varying degrees by cental governmmets e.g. with respect to the deermintion of tax base, tax rates,
and/or tax aminisa Noneteless, unless ther is a significant degree of 122 freedom of choice with respect
to some or all of these matters - so that local govenments can change the level and conosidtion of their rovenues -
it is not very maningful to talk about *locl autonomy. a
-47 -
LAA vt Taxes
151. Howdoe te pmprty taxscorein the tus? In t firs pace, a experica in a numberof cowies
ha sow in nrcet yea, theres oftenwideopredresistanceto thepoperty taxwhe toomuh veue is sought
fromthis soumc. One mum is prcisely becauu it is a very visble tax, for vl reaso. First, ulie the
incomntax, thepopery tax is not deduct at ouce but generallyba to be paid dicly to the nundiciplityby
txpaye i periodiclumpvumpayments.Taxpayeowhopay taxesdirectlyto goveant tendto be mo awar
of thesizeof thiir tax bill th thos whos tke-hom pay is eduoedby wedky or montly tax deductions.Mm
need to make such penodic larg paymint ny woUadd to th accouiHtbty a reposbility of local
govemint, but it alo gry incrums the nmativity of taxpayer.to evennoinal inces in taxe.
152. Seooodly,the ineltcity of the propertytaxha a smilar effcL Sincethobas of this taxdoe not u
a rulemcra autonacally overtime, the penodc nominalincs i propertytax billsneeded to mintain
el reoven whe price levelsrie reqr inad taxrat. In tunn of politicalaccountability, tis needto
confrontthepeoplewith thecostof govenint againreprents a vitu of the propertytax; agin,however,tho
downside(fromthe local government's pointof viewat least)is the heightmedvisibilityof nominaltax increases
and theaccompanying politicalresstance.
153. Thirdly,localproprty taxesof coursefinancesuchmunicipalservicaeas ducadon,roads,and grbage
collection.Thequantityandqualityof thee servicas(or theirabsnco) is thus readilylinkedto theprperty tax.
Whenpotholesdevlop in their eet, taxpayersarend daly quickto questionthe taxesthat supposy
financestret pir. Once again, the very faor thatmakes the property tax a good source of local goverment
revenmu in piciplmoa it ospecially
vulnerble to politicalrestance in ptce.
154. Oter problems result from prperty tax administrton. As a nle, property is supposedto be aaosed an
the basisof its maret value,usuallydefineda theprce struckbetweena willing buyerand a willingsller in
anarm's lengthtrnsaction Exacdyhowthis is to be implemented in countriesin whichprivatepropety is in the
procems of beig establishedin law for thefirst timein decade is not anirely clear. At the veryleast, therwis a
substantialvaluationtask requiredbeforemuchfiscalyieldcanrealisticallybe expectedfrompropert taxs in most
transitionaleconomes.Moreover,evenin countreswithwell-developed propertytaxsystems,discrepae usually
aise beweenmessed valuesandmarkt valueswithinclassesof property,betweenclassesof property,and across
unicipalities
forbothpoliticalandtechnicalreasons.Sincetaxpayerscaneasilycomparetheirpropeatytaxeswi=th
thoseof milarpwpertiesin theirneighborhood, suchdiscrepancies leadboth to spocificasse appealsand
to generalparsr for tax reief.
155. In short worldwideexperiencesuggeststhatthereare at leastthre substantialconstraintson the useof

propety taxesfor local finance:
*First,althoughtheadministration
of thetaxcancertainlybe improvedin thetannsition
countries,
therewil alwaysremainsver problemsin administeringit in a horizontallyequitablefashion,
particularlywhenpricesarechangingrapidly.
*Second,the temptationto the inevitablyrather fragilonew local goverments to indlge in
politicallypainlessbut economicallyinefficient'tax exporting' meansthat severeconstraints
shouldbe placedon thedegreeto whichlocalgovermentsare permittedto tax businesesmore
heavilythanresidentialtaxpayersif thepropertytax is to be an economically
desirablesourc of
localrevenues.
wrhird,both becausnof its fults and its virtues,heavyrelianceby localgovermentson the
propertytaxprobablyensuresthat,as is thecasein mostcountrieswherethepropeftytax is the
mainlocaltax, theywilleithercontinueto be heavilydependenton intergovenmntl grantsto
finance their activities, or they will not carry out many actvities.
- 48 -
156. D it these politicalandadminisaive problem, thepromty taremainsa potaly sigificntsourc

of revenuefor local govemmentsin trnsitional countries.Moreover,there ae good reasonsfor taxingreal
proprty both a a local tax and as a tax in genal. A tax on rel propety may, for example, mao good ose
as part of th tax system as a whole: although relivoly expensiveto adminiter, such a ta scores quite woe in
terms of both its efficiency and its equity aspects. Moreover, if loeved at the local level, a pperty tax can, as
empized above, seve as a good men of fnacing local public goods - subject, however, to three imp_ont
provisos:
(1) An adequatenationl fmwork law is estblhed (to prevent unwarrantedlocal manipulatio

of the bas and rate suctu and in particularundue loadingof the tax burde on nonsidmns).
(2) Loc govermmnts ore providodsufficient technical upport to carry out their ed of the
administrativeprocess (which should probablynot include valuation).
(3) Local gov are pemitted (withinthe paramte set out in the frameworklaw) to vay
remnts
the tax rate anmually. Such rate flexibilityis esetial if the tax is to be adequatelyresponive to
local needs and decisions.
none of the proprty taxes recentlycreed in the trnsition ecanmies apper to atsfy thes
Unfortunately,
critnia.
Own SubnationalTaxes versus Shared NationalTaxes
157. On the whole, if there is an obvious alternative(or supplement)to propety tuaxs it is likely to be some
form of local income tax, probably levied as a supplementto national income taxes. The cas for rdying on
propertytas, especiallytaxes on residentialproperty,as an importat soumr of local govemmentfinance
strong. Since, however, property taxes can only be pushed so far, if more local 'own-source' rvenue is desired-
- either to expandthe size of local activitiesor to make local govemmentsmore self-reliant- thr is much to be
said for supplemtary ('piggybacked') local income taxes.
158. Of cours, property taxes do not even come close to financing the needs of the subnational sector where
these have sgnificant expenditue responsbilities. Thus, in some countries, notably Russia - wher the subational
sector fina close to 50% of public expenditur - access to major national taxes, probably the pernl inconm
tax, will be needed.
159. Like the property tax, such a tax, if appropriatelydesigned,would be visible and hence in principlestisfy
the criteriaof politicalresponsibilityand accountability. However, the fact that income tax revenuestend to grow
with less political fuss, while presumably good news for local officials, suggests that ineas reliance on local
income taxes might be viewed with mixed feelingsby some. On the other hand, since an income tax is usually
perceivedas more progressive than a property tax, it scores higher than the latter on equity goWunds.
160. The principal argument against local income taxation is administratve. Most of the other arguments
sometmes raised against this proposal have no merit. One such argument is that there is sometiing inherny *bad
in efficiency or equity terms about allowing local govemments direct access to the income tax. Some local
governments, for example, probably have no income tax base. It might therefore be argued that it is somehow
unfair or inequitableto let those local governmentethat havesuch a base exploit it. But the logic of this agumet
is not apparent(and of course the basic unequaldistributionof local tax resources is prociselywhat the equalization
grants discussedlater are designedto deal with).
161. Similarly,arguments that local income taxes reduce national fiscal flexibilityor induce inefficientfiscd
competitionor inefficientresource allocationare at best incompleteand in general misleading. The functionslocal
govenments are supposedto carry out are essential,these expenditureshave to be financedsomehow,and the local
- 49 -
incometax approachseemsmore likely thanmost alternativesto free the national governmentfrom the responsibility
of financingsome such serviceswhile still leavingit a free had to alter its own tax system as it sees fit.
162. Moreover, assuming the benefit' model of local governmentpostulated abovs is followed, such taxes
simply constitutethe price of local public services and have no adverse effects on resource allocation or on fiscal
competition. On the contray, their allocativeeffects are desirableas is any competitionthey may induce among
local goveombt in lower-cost provisionof desired public services (includingthe intermediateservices entering
into production functions).
163. Whatevermerits swh argumentsmay have, the prevalenceof plans for local saresh in nationalincome
tax revenuesin countries such as Hungary and Russiaprobably reflects more the influenceof the historicalpatern
under whichthe revenues(like the expenditures)of both levels of governmentwOerunified than of economiclogic.
In Hungary in 1991, for example, localitiesreceived S0% of the revcnuescollectedfrom the personal income tax,
allocatedon the basis of residence.These funds accountedfor about 13% of total local revenues.
164. An alteroativeto the Hungariansysem would be to allow local governmentsto impose their own income
taxes in the form of a surchargeon the national tax.2' This option has two unique advantages:
*It is likelyto lower the level of incometaxes in Hungarywithout exacerbatingthe budgetdeficit.
*It is also likely to induce more efficient local expenditurethan would otherwisebe the case.
Suchadvantagesdo not come costlessly,however. The major argumentagainst this approach is undoubtedlythat
it would be relativelycostly administrtvely and in particularwould render the alredy difficult task of efficiently
administeringthe national tax system even more difficult. Nonetieless, the prospct of achieving both a more
efficient (and democratic)systemof local revenuesand expendituresand a more efficient (and lower)level of total
expendituresand taxes is sufficientlyatrctive to suggest that this option deservescareful consideration.
16S. This proposalcould be simplyimplementedin Hungary: for example,cut the national incometax by SO%-
the share going to local governmentsin 1991 - and allow local goverments to imposea flat-rate surchargeon the
(remaining)national income tax, to be collectedby the national govenment and remitted to the local government
in question. Local govemmentswould thus continueto have access to PIT revenues, but now they would have to
face local votersand justify their local incometax rates. In all likelihood,many local governmentswould initially
choose to levy lower taxes (surchargesof less than 100%). The result would then be to lower both total taxes and
total expenditures,since the total amoumtlocal governmentshave to spend would be lower to the extent they levy
less than a 100% surtax.2y In order for local governmentsto justify levyingtaxes as high as (or higher than) those
foregoneby the nationalgoverment, they wouldpresmably have to be able to demonstrateto their voters that they
are spendingthe moneyefficientlyand in a dsired way. Suwha check on local spendingseems especiallydesirable
in a situation like that in Hungary tcday in which central transfers are likely, for political reasons, to remain
basicallyunconditional.
166. Allowinglocal govermmentsto imposesuch incometax surchargeshas at least two major advantages:
IfSeeBird and Wallich,'Financing Local Governmentin Hungary.'
M'Ofcourse, thisexpositionabstracts fromthe possibilityof local borrowingas well as local exploitationof business
enterprises. As noted elsewhere,however, local borrowingshould in any case be tightlyregulatedand controlled,
and local recourse to businessenterprises for revenuepurposes should be heavily restricted.
e so -
(1) Local govermments - at least the richer ones - will haveaccess to a broad tax base so that they
can more adequately finance out of local resources the extensive range of services they are
supposedto provide.
(2) Whethertaxes go up, down, or stay the sam, the accountabilityof local governtnts, and
the efficiencyof their expenditurs, should icrease.
Assuming the functionslocal goveranmts are supposedto carry out are essential, the local income tax approach
seems more likely ta most altenatives to free the nationalgovernmnt from the responsibilityof financingsome
such serv-ces while sdll leaving it a free hand to alter its own tax system as it sees fit. The altenatve of local
surchargus on value-addedtaxes or on enteprise income taxes are much less desirable on both techmicaland
economicgrounds.
VI. THE DESIGNOF IPTERGOVERNMENTALTRANSFERS
167. Russia, China and Vietnam are the only transitional countries with no explicit arrangment for
intergovemmentaltransfers. In all three countries, there exist upwartdtransfers' from provinces to the cenul
government. In most other transitionaleconomies,while transfers ar generallythe most importantsourceof local
governmentrevenue, mostsuch intergovernmentalfiscalflowsare basicallydiscretionaryand "negotiated in natue.
The exceptionis Hungary, whero there is an explicit formula for the so.called normative grant." A somowhat
similar grant formula is currentlyunder considerationin the Czech Republic. In Russia, where there is an ad hoc
intergovermental grant, there is no evidencethat it is equalizing.
168. In themselves, transfersare neither goodnor bad. what matte from a policy perspectiveare their offelc
on the outcomesof interest, such as allocativoefficiency,distibutional equity, and macroeconomicstability. An
additionalessential element of an adequatesystem of transfers from the point of view of both centrl and local
govemments is Sau!parm=xyas when the transfer system is driven by rules or a formula rther than by
discretionarypoliticaldecisions. It is clear from Table 9 that in most countriessubeationalgovements are highly
transfer-dependant.On avenge, some 70% of revenues,and oftenmore, comesfrom trnsfers. Generallyspeaking,
the concept of grants for capital finance,matchinggrnts and other standard approachesare little known.
Closing the Fiscal Gan
169. Intergovernmentalfiscal transfersplay several disit roles in most countries. In the first place, such
transfers generaly constitutethe principalway to achieve what is sometimescalled 'vertical fiscal ballance, that
is, ensure that the actual income and outlay of subnationallevels of govermmentare approximatelyequal. For
various reasons, both economic and politica, central governmentsusually have much greater revenue-ising
capacity than do local governments. One way to transfer some of the revenues accruingto the center to finance
the deficits of lower levels of governmentis through intergovemmentaltansfers. Of course, such fiscd *gaps'
may alsobe closed, and verticalfised balancerestored, by trsferring revenue-raisingpower to local govermments,
by transferringresponsibilityfor expendituresto the centralgoverment, or by reducinglocd expendituresor raising
local revenues. In virtully all countries, however, there invariably remainssufficient mismatchin the revenues
and expendituresassignedto differentlevels of government for an important 'balancing' role to be assigned to
intergovenmental fiscd transfers.
EqaBlization
170. Equally important in determiningthe role and design of transfers is what is called 'horizontal fiscal
balance"or equaization. This is a controversialpolicy objectivein many countries,and the designof such transfers
requires carefil thought. The basic economiccase for such 'equalization' transfers shculd be understoodclearly.
On the one hand, such a trnmsfermay be neededto enable 'poorer' local governments- 'poorer' in terms of their
capac;ty to raise resources out of local taxes imposed on local residents (and not in terms of how high the private
- 51 -
Table 9
Stuc,ture of SubnationalGovernmentFinance
(recent year)
_ _
Hungary Poland Romania Czech R. Slovak R. China* Vietnam Bulgaria Russia

Own Resources 18% 50% 25% 2I 9% 71% 15% N.I. 4.4% -
SliaredTax 13% 25% 0% 6% 4.7% 85% N.J. 49.4% 95%
Total Local Resources 31% 75% 25% 15% 76% 100% N.J. 53.8% 95%
Transfersfrom Centrl Government 68.5% 25% 75% 85% 24% 0% N.J. 46.2% 5%
SouceMic.IBRD documents. See biblio-gf-apu -
Notes: 1I 51.4% as grants and 17.1% as Social Security Fundstransfers.;2/ Expectedto declineto 20% in 1992.; ' For 1985. Also transfersnot separate
from aggregatefigures; deemednegligible.;N.J.: No information.
- 52 -
incomesof thoseresidentsor the output of the localitymaybe) - to respondadequatelyto centraltransfersintended

to generatethe appropriatelevel of public goods (as discussedbelow). On the other hand, an equalizationtransfer
may be needed to enable govemments that are poor in the sensejust defined to provide an adequate 'minimum
bundle' of local public services to citizns. The argument in this case, similar to the familiar 'basic needs'
argument, is broadly that all citizens should be entitledto some basic level of such services, regardless of where
they happen to live.
171. If horizontal fiscal balance is interpreted in the same 'gap-filling" sense as vertical fiscal balance,
equalizationimplies that sufficient transfers are needed to equalize revenues (including transfers) and the actual
expendituresof each local govermment. Such "fiscal dentistry" makes no sense, however. Making up al gaps
betweenactual outlaysand actual own-sourcerevenuesfor all local governments,like equalizingthe actualoutlays
of local governmentsin per capita terms (that is, raisingall to the level of the richest local govemment),ignores
differencesin local preferencesfor public and private goods and thus vitiatesthe basic economicrationalefor local
govemmentin the first place. Moreover, such extremeequalizationignores local differencesin needs, in costs, and
in own revenue-raisingcapacity. Finally, equalizingactual outlaysclearly discourages both local revenue-raising
effort and local expenditurerestraint, since under this system those with the highest expendituresand the lowest
taxes would get the largest transfers.
172. For these reasons, in all countries with formal systems of equalztion transfers, the aim is either to
equalize the capacitvof local governmentsto provide a certainlevel of public services or the actual performance
of this level of service by local governments. The performancecriterion, which adjusts the transfer received in
accordancewith the need for the aided service (and whichmayalso allow for cost differentials)is in principlemore
attractive to central governments- or those concemed primarily with the provision of certain services such as
educationor social assistance- since the level of serviceto be funded is determinedcentrally and the transfer can
be made conditionalon the provision of that level of service. Unfortunately, this approachsuffers from the same
disincentiveeffect on the revenue side as equalizingactual outlays, since that goverment which tries least again
gets most - unless adjustmentis made for differentialfiscal capacity.
173. In principle,then, any sound design for equalizationtransferspivots on some notion of revenuecanacity.
At one extreme, the aim might be to provide each local governmentwith sufficientfunds (own-sourcerevenuesplus
transfers) to deliver a (centrally) predetermined level of services. Because such capacity-basedtransfers are
generallybased on measures of ootential revenue-raisingcapacity and not on actual revenues, no disincentiveto
fiscal effort is created by this approach. Differentialsin the cost of providing services may or may not be taken
into account. Of course, transfers based solely on capacity measures do nothing to ensure that the recipient
govemrnments will in fact use the funds they receiveas the centml govemnmentmight wish - unless, as notedabove,
receiptis conditionedon performance (and complianceis monitoredin some way).
Fiscal Effort
174. While the evidenceon the effects of transfers on local fiscal effort is far from clear in any country, there
is some empiricalevidencethat transfersoften tendto discouragesuch effort. Nonetheless,it is generallynot a good
idea to includean exlici fiscaleffort elementin a transfer formula. Suppose,for example, that a transfer is made
directly dependent on the relation of the effective tax rate in the recipient municipality to the average national
effective tax rate. One problem with this proposal is that the measurement of fiscal effort is considerablymore
complexthan usually seems to be realized - even if, as suggestedearlier should be done, local govemmentsare
largely restricted to non-exportabletaxes. If, for instance, tax bases are sensitive to tax rates, then the usual
measuresoverestimntecapacity in low tax-rate areas (and hence underestimatethe effort needed to increase tax
rates) becausethe base will declineif the rate is increased.
175. More importantly,putting too much weight on fiscaleffort in allocating grants unduly penalizespoorer
areas, in which, by definition, a given percentage increase in effort (as usually measured)is more difficult to
achieve. The problem givingrise to the need for equalizationin the firstplace is that the capacity (tax base)of poor
-53 -
areas is too low, not that their tax rates are too low. Most fiscal effort measures inevitably reward the richer
recipientgovernments,which find such tests easier to meet. Imposingsuch an additionalpenalty on poor regions
in a transfer program that almost invariablyfalls far short of fully equalizingfiscal capacity seems hard to justify.
MatchinoGrants
176. A final componentof many transfer systemsis what is called a matching(conditional)grant, in whichthe
central governumt pays only part of the cost of crtain expenditurescarried out by local governments. (At the
extreme, where the cental governmentpays all the cost of work carried out by a local governmnt acting as its
agent, what is really happeningis of course simplycgotreimbursement.) Severalr-ionales for such transfersmay
be distinguished,each with differentimplicationsfor program design, as discussedbelow.
177. Matchinggants in principlehave importanteconomicand fiscal advantagesin terms of both allocative

efficiency(spillovers)and the efficient use of scarce centrl govermnentresourcesto attain desiredlevels of certain
services. In addition,while of coure renderinglocal govemmentsmore susceptibleto centralinfluenceand control,
matching grants may have the inportant political advantage of introducing an elemnentof local involvement,
commitment, accountability,and responsibilityfor the aided activities. Moreover, properly-designedmathing
grants may contribute to equalization(horizontalfiscal balance) and, like all other transfers, they help to resolve
any basic fiscalmismatch(vertical fiscalbalance)problem. Suchgrants may be particularlyimportantwith respect
to capital investmentprojects (where they may substitutefor, or supplement,subsidizedloans).
178. The rationalefor matchinggrants with the strongestbasisin the economicliteratureis that the benefitsfrom
the local activityin question may spill over to other jurisdictions, that is, provide benefits to localitiesother than
those which decidewhbther to undetake the activity. Since such externl benefits will not be taken into account
by any particularlocal governmet in deciding how to spend the funds at its disposal, in general too little such
activitywil be undertakenunless the local governmentreceives a unit subsidyjust equal to the
externality-inteansive
value at the marginof the spilover benefits.
179. The correct matchingrate (m), or the proportionof the total cost paid by the central government,is thus
in principle set by the size of the spillovets. This rate may decline as the level of expenditure rises, if the
externalitiesdiminish. It may also vary across localitiesif there are reasons to expectgreater externalitiesin some
places than in others or if there is reason to expect a higher local price elasticity of demand for the service in
questionin some a&ss as opposedto others. Basicaly, however, a matchinggrant program designedto encourage
the optimalprovisionof public services would be expectedto vary primarilywith the nature of the activity, that is,
dependingon the level of associatedextenalities.
180. Sinceno country has achievedfull equaeizationof local fiscalcapacities,a uniform matchingleveloffering,
in effect, the same price to differentlocal governmentswil discriminateagainst poor regions. Indeed,even if
revenuebases were fuUyequalized,thetr mightstdl be groundsin terms of need or cost differentialsfor including
an equalizationelementin matchinggrant formulas. For example,per capita grants for roads in sparselypopulated
and mountainousregions should generallybe larger becausethe per capita cost of achievingany particularstandard
of road service will obviously be higher.
181. A quitedifferentrationale for matchinggrants mayarise from the existenceof a centml governmentbudget
constraint. If the central government wishes to use its scarce budgetary resources to attain given standards of
expenditureon certainservices providedby local governments,it should pay only as much of the cost as is needed
to induce each local govermmentto provide that level of service. With a grant of m percent of cost, the effective
price to the localityis 1-m. To ensure maximumtotal (local plus central) expenditureon the service in question,
given the size of the centml govermmentcontribution,the optima way to allocate a given total transfer among
localitieswill then be inversely to the price elasticityof local demand for the service (assumingno cross-price
elasticityeffects).
- 54 -
182. Matchinggrants should as a rule be inverselycorrelated to the income level of the recipient govenmment.
The purpose of such transfers is essentiallyto ens that all local govermmnts, regardless of their fiscalcapacity,
provide a similar level of certainspecifiedpublic services to their residents. The idea is simply to set the pQce of
the service (1-m) to each local goverment in such a way as to neutrlize difference in capacity by varying the
matchingrate (m). Th higher the incomeelasticity,the higher the matchingrate needed for low-incomerecipients
(to offset the higher local expenditues on the aided sevice in higher-inokomareas), and the higher the price
elasticity, the lower the matchingrato needed to achieve a given level of total expenditures. In practice, there is
thus a case for varying matchingrates inverselywith income levels even when only the incenutiveoffcts (and not
the distributionadeffects) of matchinggrants are considered.
183. Unfortunaely, neither thoory nor availableempitical studies provide clear guidelines to determinethe
precise matchingrate appropriatofor particularexpenditureprogrms, lot alone how those rates should be varied
in accordance with the charcteristics of differentlocal governments. Nonetheles, a possible approach might be
to considerthe matchingrate for each progrnmas havingtwo components. The basic matchingrate for eah service
would then reflect the degree of centrl goverment interest in the provision of that service whether that interest
is motivatedby concem over spillovero,the *meritgood natre of the activity,or simply the desire to implement
some plan). This basic rate could then be increasodinverselyto a uniformlydetermined measur of capacity. Th
matchingrate facedby any particularlocalityfor anyparticularprgram would then be higher the greater the degree
of central interest and the lower the (expected)degree of local enthusiasm(price-elasticity)and ability (income-
elasticity)to support that progrmL Though data limitationsmtn that such refinmets are most unlikely to be
relevant in any transitional economy for many years, these principles should nonetheless be kept in mind in
developingan intergovernmentaltrasfer system - or any systemof subsidizedmunicipalcredit, where exactlythe
same considerationsare relevant.
Other Obiectivesof Transfers.
184. In addition to the economicar _ments

for transfers discussedabove, there are of course also important
politicalargumentsfor transfers in all countries:
* It may be necessay, for example, to transfer some resources to jurisdictions that do not, strictly
speaidng, need them in order to make it politically feaible to transfer needed amounts to other
jurisdictions.
* It may also be essentialto transfer resources simply in order to keep some economicallynon-viable
loc,t govenmments (eg., small riral governments)alive for politicalreasons - to salvage regionalpride,
to providejobs for local supportes, or for some other mason.
In both these cases, the main design problem is to minimie any collateral damage to the presumed economic
objectives,both by achievingthe politicalends in as cost-effectivea way as possible and by trying to ensure that
the design of such transfersoffsets the good features of other trasfers as little as possible.
185. In summary,the main substantiveaim of a well-designedtransfer program is to 'get the prices right' in
the sense of facing local decision-makerswith the full consequencesof their actions. The first step in getting the
right incentives from intergovemmentaltransfers is, as argued ealier in this paper, to establish the local public
financesystemitself as much on a benefitbasisas possible. Ideally, local own-sourcerevenuesshouldcomeentirely
from local taxpayers. Local governmentsshould not have acces to taxes that they can export to non-residents
(exceptto the limitedextent such taxes may offset the provision of local public goods that lower productioncosts).
5ss
-
A Caso for Conditionaliq?
186. Given such a system, the next step is to tecogpi that (in a non-fedral systm) local auorities must
fundamntally be responsibleto the ceaml authoritiesor, mor accuatly, to taxpyers at Inrge, when they are
spendingcental funis. Ther is thus il prnciple little role for completely unconditiona trnfe - except,
perhaps, to the extent that such "tnsfers are not really tuane at all but rather simply centm ral
for locally-executedcentral prjocts or else result from centrl colloctionof local taxes (for example,because it
wouldnot be cost-effectiveto set up separatelocal tax adminisratons). Since any other unconditionaltransfen in
this system are essentially motivatedby politics, the concern these case is primaily to limit the damage done
to policy outcomes - for example,trnsfer. that simply finace local deficits or that age entirely disretionay in
naure are invariablybad.
187. On the other hand, transfersintendedto encourage"rending on a specific local service, whether because
it generatesexteanlities or becauseit is more efficientto a - 'he service locally, shutld genenrly require
somelocal contribution(matching)and shouldof coursebe cono..Aonuon the performanceof the servicein question
in accordancewith specifiedstandards.A' Both the detemination of the appropriatematchingrates and the extent
of centralsupport and monitoringof local performance(see also the ealier discussionof this subject)require close
studywith respect to each substantivearea in which such grnuts (or their analyticalequivalent, subsidizedloans)
are to be established.
188. So far there appears to have been little considertion in the trasitional economiesof what would appear
to be a strong case for some degree of conditionalityin centa transfers. In Hung=v. for instance, which has by
far the best-developedtransfer system, the normatve gnat has three importantcharactiics. In the first place,
the total to be distributedto local govemmentsthroughthis grant is tirely discretionary. Secondly,the fomula
for distributingthis total in turn contains two components. The first conponent, which may be thought of as the
equalizion componentof the grnt, is essenially an equal per capi tranfr. Tho secwod,and lager, part of the
grant, however, is distibuted basically in accordance i largely based on a mease of expecdituro needs',
particularlywith respect to education. Finally,the grant is completelyunconditional:local govenments can spend
the money howeverthey see fit.
189. In principle, there seems little rationda for bestowingsuch large unconditionalgrants on so many small
localgoverments. As with some other featues of the Hungarin local govenment financesystem, it appears that
the understandabledesire to reject completelythe centrally-dominatedmodel of past intergovemnmenal rlations may
haveresultedin, so to speak, throwingout the babywith the bathwater. The centrl government- that is, taxpayers
in general - have a legitimateinterest in what is done with grants to local govemments. Moreover, the country as
a wholealso has a legitimateconcer to ensurethat services such as educationand healthare availablethroughout
the country at least at minimumstandards. There is theroforea strng case for at least limitedconditionalit, for
instance, by requiring that the grant funds should be spent on e.g. education or hellthU' or requiring local
govemmentsreceiving such grants to provide servicesof at least a specifiedquality and level.
LNo matter how closely supervised,the inherentfungibilityof financial transfers means, of course, that there is
no guaranteethat conditional tr usfers will not in whole or part simply replace locally-financedexpendituresthat
would have been made in any ca . On the other hand, the availableempirical evidencein a number of countries
suggests that to at least some extent conditional funds do indeed astickl, that is, result in an increase in total
spendingon the aided function.
Wlofcourse, such legal requirementsare inevitablyto some extent only pro forma. The fungibilityof moneyand
the ability of local governmentsto alter other expendituresand taxes mean that requiringa grant to be spent on a
particularactivity does not necessarilyimply that tota1(centrally-finded plus locally-funded)expenditureon the
activityhas gone up proportionally. Indeed, in most cases it will not.
- 56-
190. Compliancewith any such conditionscould be monitored,as suggestedearlier, through requireme for
uniform and timely local financialreportingand tirough periodicnational inspecdonsand audits of local facilities.
Althoughtn the current situationin Hungaryit is probablypoliticaly inadvisableto alter the presentunconditionality
in any importantway, at the very least the national govenment should make evely reasonableeffort to improve
local financialreporting - for example, making the provision of such reports a condition for receivinggrants - as
well as attemptingto improve its informationbasn on what is going on with respect to the provision of local public
services.
Deterininninthe Size of the Ora.
191. Anotherfeature of the normativegrant that requires further considerationis the determinationof the tot
size of the grant. At present, this determinationis in effect entirelyup to the central govemment. Whilebudgetay
flexibilityin this respect is obviouslydesirablefromthe centrl goverment's short-runpoint of view, it is a mistake
to view cental transfers to local govements as constitutingan entirely 'compressible' portion of the national
budget. On the contrary, many of the servicesprovidedby local govenmments constitutoessentialinftucture for
future development. Moreover, since it is unlikely that most of the small local governmentscreated in Hungary
can ever finance the provision of such services at an adequatelevel out of their own resources, either now or in the
foreseeablefuture, centrl grants to local governmentswill remain an important, and largely non-compressible,
expenditureitem in the centralbudget of Hungary, as of most transitionalcountries.
192. In these circumstances,there may be much to be said in favor of establishinga formula-driventotal e.g.
as a specifiedpercentage of national revenues or of some particulr national tax or taxes. This procedurehas
subs antial advantagesfrom the point of view of both centra and local budgeting. Since the amountof the local
transfer is determined,the central govermmentis to some extent insulatedfrom pressure to increase it support of
local govenments. On the other hand, local governmentscan budget with much greater certainty whenthey know
that the total level of central support wiUvary with e.g. income tax coUlections(distributedin accordace with a
known formula) than when they are totally at the mercy of discetionary contral policy. Again, in the transition
scialist economies,the grants are typically fixedin an ad hoc way, both in aggregatevolume, and in terms of the
allocation across individual governent units, which are negotiated and bargained between center and locals.
Hungaryis the only exception here: the grant volumeis detemined annuaUyby MOF, but its allocationis by a
complex(somemight say overly complex) formula. This complexitycan be explainedin part as a reactionto the
imprecisionand bargainingby which the previousallocationsystem worked: over-precisionis an attemptto move
as far as possiblefrom this approach.
Distributionof the Grant.
193. Howeverthese two points- the unconditienalitof the grant and the determinationof its total - are settled,
the distributiveformulaof the grant also needsto be addressed. As noted above, at present there are two elements
in the presentHungarianformula: equalization(per capita) and needs. As agued above, a third essentialelement
in any general grant formula is to make some explicit allowance for the revenue-raising capacity of local
governnents. That is, the basic formulaof the normativegrant should be altered to some version of the following
general formula:
Gt = eE1 - tRl
whereG is the amountof the grant, i refers to a particularmunicipality,E to some measureof 'need' (for example,
the presentnormativegrant formula, population,or anything else that seems appropriate), e is an assumedlevel of
expenditurefor each unit of measuredneed, R is a measureof revenuecapacity, and t is the assumedrate at which
this capacity is tapped (or taxed).
194. Asuming for the momentthat the tota G to all i is equal to the prest normative grmt and t the
'needs" measure is that used in the present formula- as noted earlier, both of these assumptionscan be changed
- 57 -
in any desired way without affecting the general argument! - the introductionof the capacity element has three
importanteffects. First, it will shift grant funds from high-capacity to "low-capacity recipients. Second,it will
stimulateall recipients,regardlessof their estimatedcapacity, to tax that capacity at the assumedrate (for example,
the national average rate), becauseif they do not do so, the gant they receive will be reduced precisely by the
amountthey fall below the assumedrate. Finally, on the other hand, if any recipient choosesto levy higher taxes
than those assumedin the grant formulait gets to keep all the extra revenues- that is, it is not 'taxed by having
its grant reduced.
Measuring Canacit and ExpenditureNeed.
195. Clearly, the most critical element in this formulationis the measurement of capacity and need. Such
calculationsare difficult in any circumstancesand perhaps particularlydifficult in the changing circumstancesof
the transitionaleconomies.Nonetheless,some relevant informationexists: for example, on collectionsfrom the old
duties, on PIT liabilities, and on the 'turnover' basis for the business tax. Moreover, there appear to be
considerableinformationin some countries - notablyHungary and the Czech Republic- on the basic characteristics
of the new local governments:their population, its demographiccharacteristics,their economicbases, and so on.
It should thereforebe possible to construct a number of estimatesof revenuecapacity (defined so that the estimate
is not subject to local manipulation)W'and expenditureneed in order to experimentwith variations of this formula
and to determinewhether the resultingdistributionof the gant makes sense in light of needs and the aims of the
new local finance system.&
196. In contrastto the relativelywell-developedHungariangrant system, in Ro for example, the present

tmnsfers to local goverments are entirely discetdonary and negotiated. The same is true of Russia, and most of
the formerunion republics, now the CIS. On the other hand, like the Hungarian case, this transfer appeas to be
lump-sum,that is, it would appear that local govermets can spend it, once they get it, free of all conditionsand
constraints. In fact, however, as under the old regime, the local branches of the Ministryof Economyand Finance
continueto supervisethe implementationof local budgetscloselyand to make sure that the targets specifiedin local
budgets are achieved.Most of the transfer has in recent years been used to cover the losses incurredby the regies
owing to the large heating and transportationsubsidies. In 1991, for example, these two subsidies accountedfor
more than 32 billion lei out of a total transfer of 65 billion lei. By September1993, however, these two subsidies
are supposed to be gradually phased out, folowing a 25% subidy cut on May 1 1992, and similar cuts on
September1,1M, May 1, 1993and September1 1993. Coreponding to the cut in subsidies,the cental trnsfer
to local govemmentswill appatly be cut proportionally,althoughthat this would happenwas by no means clear
to some local govermmentofficials.
197. The basic criticismsthat may be made of the present transfer system in Romania(and other countriesas
well) are thus three. First, it encourages recipientsto be wasteful and lax - and thus run bigger deficits and get
bigger transfers. Second, the system may of course also be criticized for its lack of transparency and for the
considerableamountof negotiationthat may (at least potentially)be hidden in it. Third, even if these defects were
corrected- as they largely havebeen in Hungary- it would basicallydeliver funds derived from taxpayersin general
tihe same is true with respect to the expenditurelevel per unit of measuredbase: it may, for example,be taken
as the average national level of expenditure on the chosen item, or as some desired or minimum level of
expenditure.
LwNote that the estimatesof tax capacity required for implementinga grant formula such as that suggestedin the
text are estimatesof the ability of localitiesto raise revenue and not necessarilyestimatesof the ability of local
residentsto pay taxes. Both the wrichness'of the localityrelativeto othersand its taxing power are critical to such
calculations.
EFor preliminaryexplorationsalong these lines, see Bird and Wallich, 'FinancingLoal Governmentin Hungary.'
- 58 -
to officials in particular localitiesto spend as they see fit, thus vitiating the essentialdemocmtic (and efficiency)
principle of accountability.
198. In eality,of course, the existingRomaian systm offers almost no room for local incentiveeffects since
in fact it is still the old centralcontrolsystem in a slightlydifferentguise. That is, insteadof unconditionaltransfers
- set in some undeterminedbut apparently'gap-fllling way - in fact up to now local govermnrets have (regardless
of what the law appeas to say) acted as though their expenditure too wern fixed (e.g. with respect to sali,
number of employees,levels of services provided, and so on). If the level and compositionof local expenditure
is, de facto, deUmined centrally,it doss not matte that the level and compositionof local revenues is similady
dete_rmie there is no effectivelocal autonomy. As such autonomybegins to become real, however, and tbho
is evidence that this is now happeningin at least some are of Romania, the bad designof the present tras
system will becomeincreasinglyunsatisfactory.
199. For this reason, it is importantthat the future developmentof tansfers in Romania(and other trnsitional
economies)should takemore explicitlyinto account the basic principles of a good transfer system. For exaple,
j2jgk)
the size of the total transfer to local govemmentsfrom th centrl govemment(the Adistibutable mightbe
set as a fixed percentageof totd centraltaxes. This is desirableto provide both budgetarystability - to insulatethe
centralbudgetto some extentfrom local problemsand to providea *hard' budget constraintfor local governts -
as well a sufficient budgetaryflexibility over time to accommodatethe reasonable growth of local finances.
Moreover, experience suggeststhat it may be preferable to set the size of the *distributablepool' with refernc
to all taxes rather than as a percetage of a particular tax or taxes in order to avoid distorting central tax policy
decisions.
200. In addition,as suggeted above, the distibution fbEnulshould in principlebe basedon some feasible,and
acceptable,measuresof a aciynd necd and the transfer paymentsshouldbe made conditionalon performance.
Although in the end the extent to which any transfers are equalizing' in any country is inevitably essentially
political, it is critical to have an adequae quatitative base in order both to be able to design such equalition
features propedly(e.g. with respect to mathing grants) and especidly to assess their probable effect. A critical
research need in most transitional economiesis therefore to derive such measures of capacity and need and to
developa system for monitoringperformanc. Close attention must be paid to the supervision, acounting, and
monitoringof trnsfers, and associatedlocal expenditurs. This questionis also importantwith relationto both local
borrowing and any use of matchingfunds e.g. for investment purposes. As noted earlier, a transfer fomula
properly designed in terms of needs, capacity, and performancewill also encourage local fiscal efforL Finally,
although for political reasons, it may prove necessary to bas distribution in part on derivation or to make the
transfers payments partly or wholly unconditional,it is essential from the point of view of both efficiency (in
resource allocation)and accountability(in public sector decision-maling)that these elementsbe minimized.
VII. BORROWING
201. In most transition economies, a striking feature of subnationalfinances is the generous legal access of
subnationalgovernmentsto unlimitedborowing, sometmes even foreign borrowing. In Hungary, the Local Self-
Government Act penrits unresticted domestic borrowing (for current and capital expenditures) to aU local
governments. In the RussianFderation, the law on Rightsof Local Self-Govenmentpermits the same for oblast
and rayon level govemnments.In Bulgara, the municipalitiescan borrow from banks, and in addition, under the
existinglaw theyhave an unrestrictedright to issuebonds. In Poland, borrowingauthorityis more limited: by law,
localities are not allowed to bonrow more than 59%of current budgeted expenditures. In Romamia, local
governmentscan borrow in unrestrictedfashion, as an elementof their *autonomy",provided they can prove their
creditworthinessand amortizationcapacity. In China, provincial govemmentshave until recently had unrestricted
access to foreimiborrowingthroughprovincially-owneddevelopmentbanks,; this accesshas recently beenteduced
through a new law on foreignexchange managementand debt registration. Local govemments in China cannot
59 -
borrow domestically,however, at least directly, although they do so indirectlythrough enterprises they own and
through local financial intennediarieswhich float bonds designatedfor local governmentprojectsaL'.In the CSFR,
the two republicshave restrictedaccess to borrowingthroughT-bills issued on an agreed basis and amount by the
Federal Ministryof Finance. By and large, however, despite these liberal provisions, local governmentborrowing
has so far not been significant,as table 9 shows.
Eat Borrowing Practices.
202. In some countries, however, the heritageof debt from the previousregime has becomea problem. In the
pre-reform situation, in which the national credit plan was designedto finance investmentand other expenditures
called for in the Plan, to the extent that in the unified frameworkof the Plan, local governmentsrequired loans,
borrowingwas not only permitted,but indicated. Th loans weromade by one or another of the State Banks,whose
lendingwas wholly plan-directed.In Hungary, for example,my localitiesborrowed under the credit pla in the
pre-reform period, with each loan being guarnteed by a higher-level authority. In the Slovak Republic the
ind_btednessof the local (municipal)governmentto tho banking sector is espociallyhigh: some 30 municipaities
have beenable to pay neither principal nor intereston these loans. In Poland, it has been reported that almost 900
geminas (40% of all local governments)are in debt, as a result of borrowing, mainly from banks, under the pre-
reform regime.
203. The facilitiesfor borrowingprovided to local govenments in the post-reformlegislationthus have roots
in the planningperiod. In presentday circumstances,however, such generousand virtuallyuncircumscribedaccess
by local govemmentsto loan financeseems out of place. Such acces is restrictedeven in most westerncountries
for severalreasons. First, the centrl governmentusesdebt financeas a stabilizationtool, and it does not want local
govemmentsactingin such a way as to cAmntor its policies. Second, local borrowing may in some cirumstances
crowd out private sector borrowingwhichmay be more economicallybeneficialto the country. Third, to the extent
central govenumnts wish to avoid local govemmentsbecoming bankrupt they in effect implicitlyguaranteelocal
govenmuentdebt, so that local governmnt borrowingbecomes a potentiallyopen (and destabilizing)door to the
nationaltreasury.
Table 10
SubnationalGoverment Borrowing
Country Local BorrowingAuthority Borrowing % of total

revenuein recent years
Hungary Yes (unlimitedin law) some municipalitieshave
past debts
Poland Yes (limitedin law) 1.4%
Romania Yee" N.A.
Bulgaria N.A. N.A.
Czechoslovakia Yes significantpast debts
_______ ___ ___
___ ___ ___ ___ ___ ___ ___ OUtst nding
Russia Yes (unlimitedin law)f' '
China Nay 0
Vietiam N.A. N.A.
Notes: [/ Even though local govemmentsin Romaniaare pemitted to borrow in principle, in reality ther exists
no borrowing-just arrsea (non-paymentof bills); 21 This was suspendedin 1992; 3/ Sometimesborrowingtakes
place through enterprises (see text).
IL1See 'China: Revenue Mobilizationand Tax Policy', World Bank Country Report, WashingtonDC, 1990, and
-China External Trade and Capital, World Bank CountryReport, Washington,DC, 1988.
- 60 -
204. For these reasons in virtuallyevery developedcountry local governmentaccess to capital marketsis strictly
controlled. Among the methods used to control local borrowingare: (i) permitting borrowingonly for approved
capitalprojects; (ii) requiring prior approvalof local taxpayersfor borrowingabove a certainamount;(iii) requiring
prior approval of central authorities for borrowing;(iv) restricting the amount of debt to some percentageof local
revenues;and (v)permittingborrowingonly from a cental mnunicipal fund.' All such restrictionsobviouslyreduce
local autonomy. On the other hand, it is also commonto provide some capital assistance to local governmnta, in
the form of matchinggrants (as discussedearlier)or in the analyticalequivalentof explicitlyor implicitlysubsidized
borrowingconditions.
205. The precariousmacroeconomicsituationin many transitionaleconomiesmalkesthe case for limitinglocal

access to capital markets even stronger than usual. Additionalreasons to be concerned with problemsthat might
arise becausemost of the new municipalgovernmentshaveno experiencewith financingand are in many instancs
not yet capable of preparing and presentingcompleteand meaningfulprojects. Nor, for that matter, are most of
the financial institutionsin transitional economiescapable of evaluating long-term risks or hantling long-tem
financialinstruments.
206. Nevertheless,to the extent that benefitsfrom some projects are enjoyed in the future, it seemsboth fair
and efficient that future residents share in the cost of financingsuch projects. Borrowingfor local capital projects
may thus have a sound theoreticalbase. Moreover, borrowingmay someedmesbe the only practicalway to finance
large capital outlayswithout huge, and undesirable,variationsin local tax rates from year to year. There is thus,
in principle, a strong case for financing capital projectsat the local level through debt finance.
Sourcesof Borrowin .
207. In the short term, however, there should probably be imited or no local borrowing in most such
economies for macroeconomicreasons. In the longer run, however, some facilities for limited sub-nadonal
borrowingwill be needed. Long-term borrowingby local governmentsshould be restricted to the financingof
capital projects, perhaps with regional goverments and large cities having a little wider discretionin using debt
At all times, borrowingfrom the central bank shouldbe prohibited.
208. A common view in some of the transitionaleconomiesappears to be that local governmentsshould be

tapping the financial markets for capital funding, thus serving the dual aims of developingcapital markets and
financing long-gestationprojects. However, only the very largest such govemments are ever likely to be able to
access financial markets directly. M.'oreover,only in the U.S. are financial markets the major source of debt
financefor local governments.Commercialbanks are a more likelysource of funds in transitionaleconomies- and
theymaywell reman the main source, as in Germany.Bondissuesmightcome later, if ever, once marketsfor such
financialinstrumentsare developed,and sub-nationalgovernmentsestablish creditworthiness.
209. For most local govermments,however, the only way to access capital markets may be throughsome such
deviceas municipaldevelopmentbanks, especiallyif they are not set up by the central government,but spring up
as cooperativeefforts of, for instance, groups of cities. Such banks (or funds) could give technicalassistanceto
the individualgovermmentsin this new field of borrowing, would take up deposits from them and lend them the
necessaryamounts. Local govermmentsshould not, of course, receive subsidized credit through any such source
(except when the national governmentexplicitlywishes to subsidizeparticular types of municipalinvestment,for
reasons similar to those discussed earlier in conjunctionwith matchinggrants). In view of the tangled relations
betweengovernmentsand enterprises discussedelsewhere,it is equally importantthat such credit not be extended
indirectlythroughborrowing by goverrunententerprises.
210. The history of special MuniciRalDeveloomentFundsa in other countries is in some respects not very
encouraging(Columbiais an exception). Few of these funds have developed a capacity for sustained(continued)
-61-
assistanceto municipalgoverment and investmenton the scale needed.9 Most have been undercapitalized.As
a rule, they have tended to play a narrow and passive financingrole, applyinglittle technicalor financialappraisal
to the schemesthey have funded, and offering little positiveassistanceto municipalitiesother than capital finance.
Thosedesignedto provide technicalassistancehave tendedto concentratoskillsand resource on the direct execution
of capital works, which are handed over to municipalitiesfor operationand debt service. While this apprch hs
added to the stock of urban infiastructure,it does little to promote the capacity or commitmentof municipalitiesto
operate or expand that infatructuo effectively,or to recovercosts. Too many municipaldevelopmentfmdashave
tolerated substantil arars, with poor repaymnetrecords leadingto weak institutins and incentivesfor localities
not to repay. For al thm reasons, are must be exercisedin sottingup specid municipal financinginstitutionsin
transitioaldeconomies.
VIII. PRIVAT[ZATIONAND OWNERSHIP
211. A strikingfeature of the fiscallandscapein the transitionaleconomiesin the continuedsignificantowne_ip

role of subnationalgovemmentsat a tirm when center/nationalpriorities support privatization. One reason is of
course the transfer of assets to them in the process of 'decentrlization of ownership described earlier. Another
reason arises in some degree fom the inadequate revenue base given localities and their continued financial
dependanceon enterprise revenues. If the privatizationprocess is to be completed, getting local govenments out
of business, and into the business of government' is crucial. In the course of this process, importantquetions
remain to be answered, relating to the use of the tmnsitory revenues from privatization, and how best to take
advantageof the asset stock held by local governments.
Asset Transfers
212. One of the most importantinitial featuresof fiscal denalizaton has boen the transfer of asset to loca
governments. Local govenments are increasinglybecomingownersof enterprises,housing, vacantuban land and
retail establishments,as these assets are transferred from the center. In Hungary, the 'Act on PropertyTrnsfer'
(1991)gave local govemnments ownership of the housing stock previously owned by the central governmmt. In
Romania, local governmentshave been given all fonnerly 'statew (centrally) owned properties within their
territories, rangingfrom public doman ptoperty (parks) to private domainpropetty such as enterprises. In Russi
enterprises have long been owned by each level of government, and housing and land has been transferred to
municipaland rayonlevel governmentsbeginningin 1991. In China, decralization of enterprise ownershipfrom
the state to provincialand local ownershipwas a key feature of the overall decentalization of economicdecision
making,and took place as earlyas the mid-1980's.Housinghas always been a municipalor enterpriseresponsibility
(See Table 10).
213. Suchasset transferscan be a mixed blessing.Whileconsistentwith decentraliation and rhetoricabout local

autonomy, such transfers also fit with the center's aim of reducing its involvementin the economy. Indeed, one
motivationfor transfersof enterprises, land and public housing, has likely been to avoid the fiscal burden (suchas
subsidies,maintenanceand repair) that ownershipresponsibilitiescarry. LocalgovernmentswiUthusbe fhcedwith
these major maintenanceand subsidy burdens, unless costs can be quickly recovered. Privatizationof the houwing
stock in particular is a priority (see discussionbelow on timingof sales) as is sale of locally-ownedenterprises.
See Ktn Davey, 1990;MunicipalDevelopmentFunds and Intermediaries";PPR Working Paper No 32; World
42/
Bank
- 62 -
Table 11
PrivatizationRevenues:Ownehip and Disposition
RecentTraosfers of Ownrshp
Country Housing Reail Units Domstic Industrial
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Enterprises
Hungary Lal Local Centrl, Local
Poland Local. LoAM Central, Lca
Romania _Al _ _ caW_ CntUrl, Local
Bulgaria Local I/ Locale Centratl,LoA
CSFR Local LOc Cental, Local
Russia Local,Enterprise LoLcal
China Local,Enterprise Local, Central Contl, Local
Vietaln Lcal,Enterprise Local, Central Centrl, Local
Notes: In Bulgara only Q-T1
t tothlc
of th s ot local authoriies.
l he r
85-90% is now privatey owned.; * almost exclusivelylocal.
Privatizationvs Entmreonoeudlism
214. Unforuately, in at leastsome countriesjust the opposite appa to be taling place. There is conideble
evidencethat this process of proprty transfer is giving broader scope to thoeenrpreneulal ambitionsin the local
governmentsas they use their now property rghts to activateoidle property. A womsomo patten in vitually
all transition economieshas been the involvementof governmentsat differentlevels in joint economicvenure i
purely private maket oriented activities. Romania'slaw puts it best, but is not exceptional:* Local gove m_U
are authorizedto rerae, contrbute equity and or take shares in commercialenterprises formedwith their assets
belongingto or transferred to their ownership, includingthe possibilityof formingjoint ventures with foteign or
domesticparlners.
215. Localitiesappear to be excessivelyoptimisticabout their abilities to enhance reveaue from this source.
One vehicle some hope to use is to establishjoint ventures with a domesticor foreign parbter, or nothr state
enterpise, using local assets as the locality's equity share. In those localitieswell endowedwith land, ths appeas
to be the preferred equity contribution,and loclities appear also to see potentialin developingand seiricing empty
land so as to enhance its value as equity. Of course, real estate developmentand financing is of course Oneof the
riskiestbusinessareas in maket economies,and it is not likely to be a better bet in the transitionaleconomie. In
Hungary, toursm lodges, recreationalfacilitiesand golf coures were cited as examplesof such venturesin-process,
and in localities which inheited important real properties similar potential was sen in developing thom as
contributionsto industral joint ventures such as bakeries, construction firms, and food processing. In Russa,
virtually all rayon governmentsvisited in the course of field work indicatedthe noed to developand exploit their
asset base through domesticjoint ventures, and looked forward to the revenue flow from the projects, one
completed.Similarly,in Poland, the provisionof private services as a meansof eaming additiond revenuesapper
to have strong support, with housingand real estate development,food and furinitu manufacure cited as exampls.
216. From the prspective of locd government,there are good reasons for governmentinvolvementin these
new market ventures. With availableland that can only with difficultybe privatized, the cost to govrment of
contributing this resouce to a joint venture may be perceived as very low. Also many local public officias
understandablywish to create buiness employmentopportunitiesin theirjurisdictions,whichare often oca'omically
-63 -
depressedand lackng privao e preneurship.0However,tho hope for profitsfromsuchventur my quickly

tumintodemnds on govemmet budgetsto coverlosses. Experienceeverywher showsthtutheperspectives and
goalsof publicofficibl tendto Makethempoorbusins managers.In piculr, it is veryhard for a govrnmt
to closea faling bum, especiay whe thelocal mploqynt iSaionis fa fromhalthy. Thereis alsotherisk
thatthelocalgovaunat ald theirlocallyowndbusinesseswll udcut privat competitors, possiblybyplayig
rogulatorypmes to protectlocd monopolies, withthereoult tet trole of govement in the econmy willnot
diminish.And,of course,if localgovernmenownbusines, thy mayfailprey to theridskof ownerhip. In
the indusil marketeconomies,te rats of smallbusine failu is high: only on in five of suchbusins
survive their first three yea. Thee is no ran to expect tt the tranitionl eonomies can beat theom
odda.
217. Moreimpotty, in additionto beingddy, rur activitybylocalgoveanmt is fundaimay

inconsistentwiththeprivatiuon drive,andreprets a bottteck to tnuodenrl o-that is, d aliton,
not fromthe stateto localgovenments,but fiomgovenmentto the privateoctor. Localetrepreneuria ty
is neitherprivadtiaonnor tru de ition.Suchb entrdeuaiaHsm sows down evenreverses,thoprooe
of privatization- thelong-termgoalof gettinggovenmt out of toe activitiesthatcanbe handledby theprivate
sector.
218. It is thuscriticalin reformingsocilist economiaeto provideadequa revenueinstrumentsand flexibility
at the subnationallevel. If the only flexibilityavailableto local govermnts in their stuggle to cope with
budgetarypressureis by usingeconomically undeirablesourcesof revenuesuch as profitsderivedfrm direct
publicownershipof local busnes, thy willdo so. Municipalities shouldnot be encouragedto developlocal
monopolyenterprise in order to securethe revf A theyneedto function,just as theyshouldnot be forcedto
establishsmall taxeson a wide nmreof proutsb andactivitieswhichae expeive to adminise, genate
widesrad publicressnc, and typicallyyieldlittlo.Somelocal aconssto a robustnationaltax base (trough
tansfers), the abilityto levylocal surcha on ctai nationaltax wheroapproprate,and especlly some
significantrvenue sourcs underIlocal controlare thusessential,if the strugglefor local revenueis not to delay
and perhaps dstroy prvatization reforms.
Timineof AssetSales:Salesvs Give-wava

219. Potentially,sellingtheassets(housing,entorpdses,land)heldby localgovemments couldyieldsyigficat
evenues.There is, however,considerablecontroversyover how this procossshouldtakeplace, and evenover
whetherassetsshouldbe sold or givenaway. The metods beingemployedin diferent countriesrangefom
auctionsto 'spontaeous privatztion' to vouchersystm In mxstof thecountris understudy,at the locallevel
assets are being sold, not givenay: In Hunguy, such are th responibilityof the StatoPrperty
ManagementCompany(for largeenteprises).For smal entepns as wel u locallyownedhouing, in both
HungaryandPolsnd,localgovernments haveresponsibility.
In Russi, thehousingstockandlocally-owned
facties
are a subject to sba. CSFR's privatizationprognm combinesgive-aways(throughvouchersfor aUcitizens)with
revenue-generating sales. Mostenterpis will be soldby the two republics;the role of the local govemmonts
withinthesorepublicsis not clear. Bulgada,andRomuniahaveyet to deviseconcreteprivatization progams, but
theyare likelyto includebothgive-awaysthroughvouchs and sales.
220. Maximining the yieldfromassetsalesisa majorchalleng. Shouldtheybe soldnowor later? Selling

assetsnow,whenmarketsfor ownershipandprperty rightsare underdeveloped, maymeanthat theydo not bnog
a goodpricefor the localgovernm t - a process,as oamanalysisput it, akinto usingthebeatfurnitur to uWel
!WAnoteof cautionis also neededwith repoct to the apparentdesireof manylocalgovemmentsto use fiscal
incentivesto bolsterprivatebusinessin theirjudsdictios. he recordof such'local development policies is not
veryencouragingin any country,andeffortsin thisdirectionare all too likelyto degenerateintoa zero-sumgame
for local governmentsunlessgreat care is exerted. On the whole, this is yet anotherares in whicha national
'fnmework law- - a binding cael arrangement- seemsneeded.
- 64 -
the boiler on a ship running out of fuel'. But at the same time, governmentneeds revenuesnow. SellingaM
later may improve yields, but such yields will come at a time when other local financing sources will also be
available. Complicatingthisdecisionis the fact that manylocally-ownedpropertiesare presentlycostingmuchmore
to operate and maintain than they are realizing in rents an, hence constitutoa major drain on subnationalpublic
finances. Gettingrid of this burden is a high priority. Unfortuately, in almost all countries,the pocem of saeling
locally-ownedenterprises, and especiallyhousing, is slow, impeu by the absenceof credit (especiallymort
finance),and the absnce of a legal framework.
Revenuesfrom Privatization
221. What should local goverments do with the proceeds from the privatization of public enteprise and tdo
sale of other assets? First, it is importnt to remembertha hes rvenues are 'non-recurring' and do not rpret
permanent sourcesof financing for any local govemment Efforts thus have to be made to developadequae tx
sources in additionto user charges for the future, when the revenues from asset sales disappear.That said, should
such asset revenuesbe capitalizedin special fundsor investedon capital projects, or shouldthey be used to finance
current expenditures?
222. In practice, there appear to have beenno special stringsattached to the use of revenuesfrom privatization.
Te unsurprisingresult is that the most commonpracticeappeassto be the use of these fundsas an additionalsource
of revenuein the recurrentbudget. This is not all bad: indeed, it may be both distributily attractiveas well
economicallyefficient(less dead-weightloss) to raise revenuesfrom privatization than from taxes. But, as noted
earlier, one cannotlive foreverby selling the furniture. Sometimessuch revenueshave beenkept in exrabudgetay
accounts. in Hungary, for instance, rents from locally-ownedhousingaccrue to off-budgetaccounts, while sles
proceeds from locaUy-ownedhousing accrue to the general budget, but are not earmarked for any partculwr
purpose. In the case of privatizationof locaUy-ownedenterprises, the proceeds also accue to the budget,with no
restctions on use. In Poland and Russia, revenuesaccrue to the budget, with no restrictionson their use.
223. A more conservativeprincipleof public financewould cal for the investmentof these proceedsin long-
term assets to alow local govenmerts to draw income from assets over a long period of time. Altenatively,
revenuescould be used to repay debt.44 In Hungary, revenuesfrom privatization involvingforeignjoint ventr
and generatingforeignexchange are channeledthrough the State Property ManagementAgency (SPMA)to repay
Hungary's external debt. In the Russian Federation, where most privatization has been domestic, it has been
recommendedthat the stock of domesticgovemmentTreasurydebt shouldbe reducedthis way. In Poland, revenu
from privadzationare being used to finance current operations, and to compensatefor inadequate tax yields. In
China, no commitmentto privadzadonis as yet apparent.
IX. CONCLUSIONS
224. Subnationalfinanceand local governmentpoliciesare emerging as crucialelementsof the ongoingrefors

in transitionaleconomies. Subnationaigovernmentsplay a large role in overall economicactivity, account for a
large share of public sector outlays, and own a significantshare of public enterprises and the public housingstok
Recent attemptsto strmlne governmentbudgetshave led to increased local governmentresponsibilityfor such
crucial matters as health, education and safety net expenditures, including unemploymentand entepis.
X'SeeNewbery,D. 'Reform in Hungary: Sequencingand Privatization,' E EconomicReview.No 35, 1991,

pp 571-580. Newbery argues against a rapid, 'fire-sale' or giveawayapproach becas many of the assots may
have been financedwith debt (as noted earlier, many local govemmentsin transitionaleconomieshave substantial
liabilities),in which case the proceeds shouldbe used to pay off this debt, rather than burdeningpreset and future
local taxpayerswith debt service.
- 65-
Underfimdedlocal govermmentsarn seleing coping mechanisms,not all of them constructive, to meet these new
challenges.
225. This paper has descibed the risks to privatizaon, to macroeconomicstability, and to an adequatesocial
safety net that present policies towards local governmnt may imply. Our themes are that the subnationalsectr
needsto be m realisticallyfactoredinto nationalplans, and thatsubnationalexpendituresbe more clearlyaigned
and revenue needs more realisticaly aessed. Such _ are likely to acknowledge* larger sphero for
subnationalgovermments and the need for access to more robust revenuesourcs. Giving local governmenta dsre
in the personal income tax is one possibleand perhaps desirableapproach to meeting these revenue noeds.
226. Carefulattention also needs to be paid to the desip and implementationof toe intergovenmental fisa
transfersthat are likely to remainprominentfeaturesof the intergovermental landscapefor years to come. Caution
is also needed with respect to borrowing by subnational govemment Consolidation and integration of
extrabudgetaryfunds at the subnational(and national)levels is crucid to enhanced budgetary transparencyand
macro stability.
227. Finally, proper funding of the subnationalsector is a necessary(albeit perhaps not sufficient)condition for
the successfulprivaization of enterprisesnow held by local governments. In its absence, local goveanmentswill
continue, as at present, to seek gains from keepingthese enterprises,and to add to their revenuesby settingvp new
commercialenterprisesand impedingprivatization.
228. In short, local govemmentfinance is not a 'local matter', but is crucial to asmost all of the key reform
goals of stabitization,safety net and private sector development
- 66 -
APPENDIX TABLES
GOVERNMENTREVENUES
ROMANIA 1991
Capital Revenue 0%
Nontax Revenue 2, Ta Revenue 27%
Capital Revenue 1%
Nontax Revenue 2%
Tax Revenue 97%
Subsidles 8 Grants 71%
Local Government Revenues Central Government Revenues

LOCALGOVERNMENTEXPENDITURESBY TYPE
ROMANIA 1989 and 1990.
Education 35%
Educatlon 5% Rec. Cult. Religlous 4%
Health 11% General Pub. Service 6%
Soc. Securlty & Wel. 1 re. Cult. Religious 3
Housing & Commun. 11% Soc. Security & Wel. 2 % General Pub. Service
Other Expenditure 1%
Other Eco. Affairs 3%
Ag, Foreaty. Fish 9% ther Expenditure 37% CY%
Mlning, Manu & Con. 1\ Transport. 8 Com. 10SO
Transport. & Com. 8% Housing & Commun. 27% T Cn
Other Eco. Affalrs 9%
Expenditure by Type 1989 Fuel and Energy 3X g oresty. Fish 11%

Expenditure by Type 1990
Public order & safety, Health and

Mining, Manufc and Construction had
zero expenditure in 1990.
-- - - ------ ---- ------ ------ --- - --
i [U
a
t
~~
~-
~ ~ ~ e0Z
A-
-e >--
-
<-
_N
- g
e
Is
tas
- -e
a
-- p
ADw~
el
; -" - a - ^--
;A - - - - --
4. ?IS F pU AD
ai . ~
~ ~ ~ ~ o4 ~ ~
------------------------------------ w0
; ;P ;A DI 8X ow h * --
o- -i ea e w *U w A
_ ~ ~ ~ 9 s a !I v a a w _4 oS.
'S at o a iI
me ~ ~
8v
8'a
fU
a _ s
0 U
;P > yro p* $
~~~~~~~~~~.,
~~, . ~ ~ ~ AD * II~~
o
ok j j.
^. g e 8 j gu a8 a 8Y
8 8. 8
a N a a !4 - a N 04~~~~~~~~~~~ '
W S .8 8----
- --
_ p p~~ Pp - P p0 P
.e p F P P - P ~
0P a*
!--: aj~- a
0 0 0 5 8P P P
- is g v R8 a S
P 8
P
---------------------------------- 6-_-
g°
P P P P 8P P P P P *-s8
-69-
= == == =O =~ =~ qI *
to OA
iw~ S~ = ~rjf a:.{1
~ 5pi N
6 6~~~~S d
w w C w4 - - - -
NM N W d CD #A co a o*
W: 'oa W _ _ -
j , 0 0 0 .0 ._ 0
.* R 8 88% a ! I
- - - - 00 Wi
o~ g o X . 8a o
--------------------
8 o- -
71 -
* tC. ' q -: - t -
EU a. .. 4 e..
a 4
- a nC ae a - o -t
2 ee q ! o e e ee* ~
e t 1 e~
v 3i ;. 0 f Po *o -C
-- -e n e: ^ w:::-- - -
£-~~~~ _____________________
a.~~~~~~~~~~~~~~~~~~~~~~a
E ,I,tSR . fa
Il2 .l I :
! q
-~~~~I
- -0. at . t N
3 ______________________ --
__ - - -__ -U- - -; - -
-___ ___
-- -*~~~~~~~
e e e. e wsa
I 2 :. q ui u
I;~~~~~
- -…--- - - - - - - -C-- -
a4 r:si4 ete 12 a - ;;a

^ i a
e;
a s.
i $ *
s
ii
"h
~~~~~~~AA-
AX < <:
-- -
".
-O- -
£1t
..... .. . _. e U
============================= - _
Xt
i j66
- -
d-Xa~~~~~~~
- - - -
a~~~~~a - - --- -
__~ ~ ___d_________ ___
I f..Ij # §|
6d
R 42.^I4E,a jI UiA ttt .55 R
________________
__E ~ j u I~I _
__ I! I Sis R
BUDGETARYEXP. AT THE SUBNATIONALLEVEL
Russian Federation, 1990
Communal Service 1%
Social Welfare 4%
Housing & Utilities 18%
Public Health 43%
Social & Cultufal 57%

Other Expenditure 12%
Education 53%
Pub. Authority & Ad. 3%
Variou's Complexe 8%
Total Expenditures Social and Cultural Activities
Source: MOF and Bank Staff Calculations

REVENUESTRUCTUREAND REVENUESHARING
FOR Q11992 (estimates) RUSSIA
Value Added Tax 28%

Company income Tax 31%
Company Income Tax 13% Prsonal Income Tax 10%
ExcIses 4% Peraonal Income Tax 4%
All Other 9%
Value Added Tax 29% Exciaes 12%

Tax on Foreign Trade 39%
Central Government Subnationals
Different values were reported for each

month of Jan. Feb and March 1992. Thus
estimates should be vlewed with caution.
* AGGREGATE
LOCALREVENUESIN BULGARIA
1991and 1992
shared lothS tEx _1|
shared Income tax 122

Own floa-tan revenue SS
OWmsource loee IS
own property lox Is
Own nonl-tax revenuo IS
shared Income Sax 54,
shored turnover tax OS
shared tErnovet tau Planned
1991 Executed 1992Planned

LOCAL GOVERNMENTEXPENDITURES- BULGARIA
1991and 1992
social security 105
Medlodamnts 4% Food TO -S
ti | _ ~~~~ivostments 10% _ 1
I_|ISt
Oflilo * economic 15s

atti_, ea Economtc 1211 _
General Repaia 3% orer 1i RaGnerepair
ts
1S Other
1991Executed 1992 Planned

LOCAL GOVERNMENTEXPENDITUREBY TYPE
HUNGARY 1990
Education 42%
General Pub. Service St
Health 6% I!!!l!!rITIIII!fhI
_______0_ Other Eco. Attairs OZ
Soc. Security & Wel. 7?
v W Transport & Comm. 8%
Housing & Commun. 11% Ag. Forest'y, Fish 3%

Rec, Cult & Religlou 8%
There was no exp. on Pub. Order and

Salety. Fuel and Energy, MIning.
Menu & Construotlon.
LOCALGOVERNMENTREVENUES
HUNGARY1990
=ax Revenue 30%
Nontaex Revenues 18%
Copitol Revenue 56
1lnonolng
a.._
Orants 44%
Tax Revenue Is only comprised ot

tax on Income, protit and capital
gains
- 78 -
s e8
4p w
8S ei
is! -- ---- ----- - - - - - -0 -
a D a o s o
-- - - - - - - - -- - - - - - - - - --
- 6;
C - .E . a; *i a
0 0 0 0 0 0 0c 0i 0 a a 0
. g8 R ; R e k * E C;
e~~
. _ o _ e _
r, _ o o 4 4a
- ~ ~ ~- .- - ---
-- - - - --- -- -- -- -- --
'a8 S a
-~ ~ ~ ~- - - - ----- - - - - ---- --- --- --
S - ~-| - -- 8- A
- - --- 'C - S
8 ------------------
0. 0S 0 ' N
_} i eg o * w s > o o rn oa e °a
- - - - - - - - - - - - - - - - - - - - - --
" | sa s. ' 3S Sa 3
8
3!
a to]S. i :g So .i
.~ __________________
ci - :i sz _3
I__
- -- - -- - - - C-0 l - - - -9'- -. - .
C i e, o. e, q q -9 q o
'.o N q o
C Cq *SC 0 - -- - - - - - C __
S°~~ A e-
jIt
J J I . i B
779_
.g _2 8 8 n s8 a s 8 - t N_ 8
a--_ 0 o o -' o4 l ° .:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - -
0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~D
i~~~ cm -cm ° N
10 - CY cm°°
00 0 M - M%~~~0
- - - - - - --- - - - ~~~~~~~~~~m-
- - - - -
- - - - - - - - r -
MM-fi C 'O - N f 1o O Nm .0 o U
-__ ___N ___

___ ________________________________-_
I D e XS
-=
o R g
40 w
2 4 - g- 4 - '- 0-- ----
_
U__~~~~~~ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _
-a____
______________
- 80 -
-. - - - - - -- - - -
1 R I I
14 - -
___ =======I======= ItI === .'
-- ______-- _------I'__ __ __ __ _
-- ------------- ---- - - - --- N
_S1 Ce 12c
- - - - - - - - -. - a
2
_ _l N j R N I
..
___ ________
_4 t ________________
N-_----- -- -I>
_____
.0.
:1 -I I I 8 '
ei I I U . t5
====~~~~~~~~~~ _==============
==
FP PI~~~~~~~~~
;t
{~~~~~ a a a ag
P t
e -3 S t w U * a 4 US _ e
~~* i ; .. L. 'U - * le..p

_ --
_* ~~ ~ ~~~~~~
0
__ _ _ _ __ _ _ _ __ _ _ _ __ _ _ _
~
_ ~ ~ w
P D - U 54 ei a a 4a 8 *
n o r
-- - - - -- - - - - - - - - - - - - - - - - - -
-- - - - - -8 ;- - 8-e- - - B
S _ _ w p 8 ° !k S~ I- 1
01
W CP ~ ~ ~ ~ ~ ~ ~ ~~~~~~
S a o 8 9g
gs -^-X-- - -us - 2S
- - - - - o- - - 8 - - 8- -
-a - - - - -a -
~ -
et
- - t- - - - - - -
- - - -
o-
- - -
- - -
a
SS S ~%
X
p p * p P P
____________________________________
P P _ S
a
___
_ ____
________________
__ __ __ __ __ ____ _
____________ __.____ __-
- 19 -
- 82 -
i- ___________________________________
atffi *. Oa 8 s c a, 8ey
.- ------------ - --- ------ - 8
km 'S _ 8NO a _ X - - -N
_ ~~~
__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ N_
_ _ _ _ _ _ N
__ ---_ __ _ _ __ _ _ _ __ _ _ _ __ _ _ _
- ~b '0 0 ^ - _~ N N
^, -*
I
~~~~~~~~
~~~~~~~~Wu
-
* ~~ t a
R8 i g t g g g Me S ~~ l 'l
R X B; iu _ t S W~~~~~~~~~~~~~~~~~a
0=~~ . s s o
E tSi 2_
t - < lP | ^3 s X is 3
- 83 -
o__ __ __ __ __ ______________ ____
O N .r 2 0 o. eN n n _ 'V o
__ __0: _ _ _ _ _ _ __ _ _ _ _ _ _ _
- PIV
_'
-- -- 4
NO
-- N
O1 g o. _
- _I e _
=_=- - - - - --- - - N - N-. N
__ _ _ _ _ _ _ _ _ _ ____ _ __
- - - - - -
a~~~~~~~~~ - - N N
-----
- 00 V 0
_ _I = _ N 1 _=_= N N4 N 'V 4 nV o0 =. ===

i - -- - - - -- - - - - -- - - - -
CVIPt V! 0.1 VI I 4 - V V
a - Q 16 ~ ~ - II
~~~~~ ------- ---- .. - - - --- 'VW N -_N - I
b_
ta S Z a I 0 8
4n, _I o
----o
5 44 -I
e.
e, °, °.--- NI 0
e-
f0 'N
N I N, o
-- - - - - --
Ix
- - - - - - -°- - - --.................
g~~~~~~~~
NO OI
-84-
REFERENCES
Bird, Richard, FederalFinance in Comparaive Persgecive (Toronto: Canadian Tax Foundation, 1986).
Bird, Richard, and Christine Wallich, 'Financing Local Govemment in Hungary,' (Washington:World Bank
WorkingPaper No. WPS 869, March 1992).
Bahl, Roy W., and Christine Wallich, -l'tergovarumentalFiscal Relations in China,' (Washington:Wodd Bank
Working Paper No. WPS 863, Febmary 1992).
Davey, Ken, 'MunicipalDevelopmentFunds and Internediaries,' (Washington:World Bank PPR Wo,rkingPaper

No 32, 1990).
Go, Delfin, 'Revenue Uncertainty in Transition Ec9nomies,' (Washington: World Bank RPO No. 67718,
forthcoming).
Jarvis, Sarahand John Micldewright,'The Targeti4gof FamilyAllowancein Hungary,' (Washington:World Bank

Conferenceon Public Expendituresand the Poor Incidenceand Targeting, June 17, 18 and 19, 1992).
Kornai, Janos, The SocialistSN:tem (Princeton:PrincetonUniversityPress, 1992)
k4ilanovic,Branko, 'Distributional Incidence of Cash and In-Kind Transfers in Eastern Europe and Russia,'
(Washington:World BankConferenceon PublicExpendituresand The Poor: Incidenceand Targeting,June
17, 18 and 19, 1992).
Newbery, David, 'Reformin Hungary: Sequencingand Privatization,' EuropeanEconomic ReviewNo 35, 1991,
pp 571-580.
Shah, Anwar, Pe.rspectiveson the Design of lntergovernmentalFiscal Relations,' (Washington:World Bank

WorkingPapor No. WPS 726, July 1991).
Szapary, and Mario Blejer, 'The Evolving Role of Fiscal Policy in Centrally Planned Economiesunder Reform
The Case of China. (Washington:IMF WorkingPapor No. 407, 1989).
Tanzi, Vito, ed., Fiscal Policies in Economiesin Transition(Washington:IMF, 1991).
Wallich, Christine, 'Fiscal Decentralization:Intergovermmental

Relations in Russia,' (Washington:World Bank
Studies in Economiesin Transition, No. 7, forthcoming).
Wong, Christine, 'Central-LocalRelations in an Era of Fiscal Decline: The Paradox of Fiscal Decentralizationin
Post-MaoChina,' China Ouarterly (December, 1991).
Wong, Christine, 'Fiscal Reform and Local Industrialization,'Moder China (April, 1992).
World Bank, China: Revenue Mobilizationand Tax Policy (Washington:World Bank Country Study, 1990).
World Bank, 'China: Extenal Trade and Capital,' (Washington:World Bank Country Report, 1988).
- 85 -
Notes for Table 3
1. Datd for total expenditureas a percent of GDP was obtainedfrom the Recent Economic Developmentsof the
IMF.
Hungary: SM/92152,March 12, 1992
Poland: SM/9l/7O, April 11, 1991
Romani: SM/92/96, May 8, 1992
Bulgaia: SM/91/52, Marh 7, 1991
CSFR SM/92/65, March 23, 1992
Russia . SM/92/26, Febnuy 6, 1992
China SM/91/17,January 28, 1991
Vietnam: SM/91/234, December5, 1991
2. Expendituiedata are for a consolidationof the budget of the cra govemet and the financialoperations
of provinces,counties, municipalitiesand townsips. Extabudgetay financialopaonS of the various lovelsof
governmentare not included. Data is based on budgetedvalues, not actuals.
3. Data for 1987.

4. Data for 1988.
5. Data for 1991.
6. For 1986.
7. For 1992.
8. For 1985.
9. For 1989.
10. 1988data for the USSR.
I1. Data for the first cuarter of 1992, the RussianFederation.
12. Data for 1990.
13. As % of USSR expenditure;USSR GDP.
* Provincial(not local) expenditures.
Pollcy Research Working Paper Series
Contact
Title Author Date for paper
WPSI102 Didthe DebtCrisisor Declining

Oil AndrewM. Warner February1993 M. Divino
PricesCauseMexico'sInvestment 33739
Collapse?
WPS1103 CaphalMobilityInDeveloping PeterJ. Montlel February1993 R. Vo

Countries:
SomeMeasurement Issues 31047
andEmpircalEstimates
WPS1104 TradePolicyReformin LatinAmerica AsadAlam February1993 J. Troncoso

andtheCaribbean in the 1980s SarathRajapatirana 37826
WPSI105 Estimating Deficisin PhilippeLeHouerou

Quasi-Fiscal February1993 N. Velasco
Framework:
a Consistency TheCase HectorSierra 34346
of Madagascar
WPS1106 Improving
Women'sAccessto Ha1lDundar February1993 S. David
A Reviewof World JenniferHaworth
HigherEducation: 33752
BankProjectExperience
WPS1107FinancialReformLessonsand GerardCaprio,Jr. February

1993 W. Pitayatonakarn
Strategies kzakAtiyas 37664
JamesHanson
WPS1108PublicOutputandPrivateDecisions:ThanosCatsambas February1993 A. Correa

ConceptualIssuesin the Evaluation 38549
of Government ActivitiesandTheir
Implications
for FiscalPolicy
WPS1109 RiskManagement andStable AndrewSheng March1993 M.Raggambi

FinancialStructures YoonJe Cho 37664
WPS1110 WhatWouldHappen HAll DevelopingWillMartin March1993 D. Gustafson

TheirManufactured
CountriesExpanded 33714
Exports?
WPS1111ForeignInvestment
LawInCentral CherylW.Gray March1993 M. Berg
andEastemEurope WilliamJarosz 31450
WPS1112Privatization,
Concentration,
and YingQian March1993 S. Lipscomb
Pressurefor Protection:
A Steel RonaldC.Duncan 33718
SectorStudy
WPS1113TheLuckyFewAmidstEoonomic Christiaan Grootaert March1993 E. Vitanov

Decline:Distributional
Changein RaviKanbur 38400
Coted'lvoireAsS(-enThroughPanel
DataSets,1985-88
WPS1114DoesPriceUncertaintyReally Annta
George March1993 D. Bievenour
ReducePrivateInvestment? JacquesMorisset 37899
A SmallModelAppliedto Chile
Policy Research Working Paper Series
Contact
Title Author Date for paper
WPS1115Lookingat the Facts:WhatWeKnow RossLevine March1993 D. Evans
aboutPolicyandGrowthfromCross- SaraZervos 38526
CountryAnalysis
WPS1116 Implications
of Agricultural
Trade AntonioSalazarBrandao March1993 D. Gustafson
Liberalization
forthe Developing WiUMartin 33714
Countries
WPS1117 PortfolioInvestment
Flowsto Sudarshan
Gooptu ttarch 1903 R.Vo
EmergingMarkets 31047
WPS1118Trendsin Retirement
Systemsand OliviaS. Mitchell March1993 ESP
Lessonsfor Reform 33680
WPS1119TheNorthAmericanFreeTrade RaedSafadl March1993 J. Jacobson

Agreement:
its Effecton SouthAsia Alexander
Yeats 33710
WPS1120 PoliciesforCopingwithPrice DonaldF. Larson March1993 D. Gustafson

Uncertainty for MexicanMaize 33714
WPS1121 MeasuringCapitalFlight:A Case HaraldEggerstedt March1993 H.Abbey
Studyof Mexico RebeccaBrideauHall 80512
SwedervanWijnbergen
WPS1122 FiscalDecentralization
InTransitionalRichardBird March1993 B. Pacheco
Economies:Towarda Systemic ChristineWallich 37033
Analysis

World Bank

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

World Bank

Uploaded by

Copyright:

Available Formats

Public Disclosure Authorized

Towarda SystemicFrameworkof Analysis

Designinga well-functioningintergovernmentalfiscal systemis

This paper - a joint productof the PublicEconomicsDivision,PolicyResearch Department,and the

The PolicyResearchWorkingPaperSeriesdissemirsates the fmdingsof workunderway in theBank. Anobjecii vof theseries

Producedby the Policy ResearchDisseninationCenter

TOWARD A SYSTEMIC FRAMEWORK OF ANALYSIS

1. INTRODUCTIONAND SUMMARY . ......................................... 1

Why 8Local1Finance Matters 1...................

IN PROCESS:TRENDSAND THEMES ........................

A Time of Transformation . ............................................ 9

II. WHO DOES WHAT? THE ASSIGNMENTOF EXPENDrrURES ....................... 29

The EfficiencyCase for DecentralizedLoca GovermenL .......................... 29

IV. FINANCINGLOCAL PUBUC SERVICESTHROUGHUSER CHARGES ................. 40

V. THE CHOICEOF SUBNATIONALTAXES..................................... 42

Characteristicsof a Good Local Tax ....................................... 46

Closing the Fiscal Gap ................................................ 50

Determiningthe Size of the Grant . . . . . . . . . . . . . . . . . . . ... .. ........ . ....... 56

VII. BORROWING ...................................................... 58

Past Borrowing Pr actic e s

VIII. PRIVATIZATIONAND OWNERSHIP ....................................... 61

Table I Trends in Decentralization:GovernmentOrganization. 16

We wouldlike to thankthose who commentedon thispaper, or whootherwisesharedtheirinsightson

Why 'Local" Fimnce Matters.

-JJanosKomai, T'heSocialistSystem; (PrincetonUniversityPress, 1992), p.577.

7. From an atlocativeprpctdve, the system of intove n fiscal rebltionswill also dtermine in a

9. Even gvimhi2n is complicatedby the current environment.Followingthe transfer in many countries

13. Second, th traditionalapproachignors th shrining role of govenment. in most t itional conomies,

17. The new subnaional governmentswiUlalso needsubstantialhelp and guidancein developingadequatelqol

18. It is critical that subnationalrevenueneeds and possibilitiesbe realisticallyfactoredinto nationalplans for

25. Finally, in somncountries,(Russia,China, andothers) price reformand other structural shiftswiUcontinue

Oreanizationof the Paper

Althoughtheprincipalfocusof Ae disussionis on the cental ad at Europon countria (especiallyHungary,

| ern~kagmeotal Fiwa Reltion in Trmtansn Ecnm2os

Political/OverallStructura Reform Context

* Intergovermnentalfinancesreforms are taking place in contextof overalUsbinking role of

* Intergovernmentalfinancesrefonns are taking place in context of a changingrole of government.

* Intergovernmentalfinancesreforms are being undertaken simultaneouslywith overU national In

Box ... (contnud)

|* AMakimi ond overall mco concernsdomiate the agenda so dt intergovernmentali finacial

* Use ot 'norms' to define how a serviceshould be provided ;imitsexpenditureautonomy.

* Cental 'm4a' on wages, and other expenditureslimits expenditureautonomyand continueto

Tax Assi e and Iafm

* Tax assignmentstypicallyunder-fimdlocal sector: low-yieldingrevenue sources are assignedor

* Generd absence of 'formula"- basedtransfers; they remain negotiatedin aggregatevolumeand by

Generalabsence of transparencyin allocationof investmentgrants/capitalfinancing.

Box .... (continued)

|* Major continuedrole of subnationalgovenments in ownershipventures at a timo when

* In most tansition and socialisteconomies,there is generous legal access of subnational

* Extabudgetay funds continueto play a major role, complicatingbudgetarytansparncy and

* In may transition economies,the instit>tionalarrngements needed for diaogue on

H. DECENTRALIZATIONIN PROCESS:TRENDSAND THEMES

Poland - L Self-Governmet Act, 1990

The End of Central Plannina: Filling The Abhorred Vacuum

42. Thesepoliticalreforms are paralleledby dcentaliton of fiscalresponsibilities.Someformof subnational

WDeconcentration refers to dispersionof responsibilitieswithinthe central governmentstructure from the center to

46. To makelocal autonomymenjrtef, local govements need adequatolocally-controiledrevenues. Not

2. Subnationalgovernmentsshould as a rule be accountableto central government (that is, to

51. Such accountabilityclearly requir thst subeationalgovernmentsshould, whenoverpossible, choa for

L' Theseare, strictly spealdng,decentralizedbranchesof the centralgovenmmentadministration,not an indpendent

country UnitayFdal # of Tiern Em N of Tiers (preet)

Ronania Unitary 3: central/regional// Unitary 2: central/local

Bulgaria Unitary 3: central/regional/ Unitary 3: central/oblass

China Unitary 3: centrl/province/ Unitary 3: central/