Intergovernmental Transfers in Developing and Tran
Intergovernmental Transfers in Developing and Tran
Intergovernmental Transfers in Developing and Tran
net/publication/253530818
CITATIONS READS
47 131
1 author:
Roy W. Bahl
Georgia State University
122 PUBLICATIONS 3,343 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Roy W. Bahl on 22 March 2014.
BACKGROUND
SERIES
Intergovernmental Transfers in
Developing and Transition Countries:
Principles and Practice
Roy Bahl
APRIL2000
TheWorld
Bank
INTRODUCTION
OBJECTIVESOF INTERGOVERNMENTAL
TRANSFERS
Vertical Balance
1
raising powers or by revenue transfers from the central governmentto the
subnationalgovernments.
LocalTaxes or Local Transfers?
2
The second issue to be faced by grant designers is that there is a
mismatchbetweenthe reliancedictated by verticalbalance considerationson the
one hand, and efficiency considerationson the other. For example, vertical
balance considerationsmightdictate that subnationalgovernmentsreceive$X in
transfers. But this amount may result in some services that should be
tax-financed being covered by grants. (See Box 1). This could lead to
overspending by the subnational governments because certain services that
should be financed with local taxes and user charges would be financed with
externalgrants.
Equalization
Externalities
3
The second issue has to do with three-levelfiscal federalism. If the grant
is made to the provincialor state government,it may not reach the government
that is responsiblefor the underspending.Relatedto this is the issue of how the
intermediate level govemment will allocate its resources to the local level
government.
AdministrativeJustifications
Transfers
Bad Justificationsfor Intergovernmental
4
may be put in place but underfundedat a later time when the central budget is
pressed.
DESIGNINGA TRANSFERSYSTEM
After the objectives are set, the next step in designing a system of
intergovernmentaltransfers is to structurethe horizontaland vertical dimensions
of the transfer. In fact, every intergovernmentaltransferhas two dimensions:the
first is the vertical dimension,the distributionof revenues between the central
and local government. The second is the horizontaldimension,the allocationof
transfers amongthe recipientunits.
5
Tax sharingis widely practicedamongdevelopingandtransitioncountries.
There seems to be no rhyme nor reasonto the choices made as to which tax
base to share, as is indicatedby the followingexamples:
* Russiaand China:VAT
* Colombia:The Tax on Beer
* India:Excise Duties
* Indonesiaand The DominicanRepublic:PropertyTax
* Peru: SalesTax
* Nigeria:NaturalResourcesTaxes
* Mexico:PayrollTax
* Brazil and Colombia:MotorFuel Taxes
The more relevant question is not the practice, but how the tax to be
shared and the sharing rate should be determined. This brings us back to the
question of objectives of the transfer system and how it fits into the general
decentralizationprogramof the country. A sharedtax can be used for no more
than some degree of vertical balance to offset the mismatch between local
expenditureresponsibilitiesand local revenues. In this case, countries have
assigned shares of a variety of different types of excises to the local
governments,or even property taxes in the case of Indonesia. This is a
centralizing approach, and it protects the central government from having
committed a significant share of the tax base to the subnationalgovernment
sector. However,the sharedtax pool may also be used as a seriousapproachto
decentralizingthe fiscal system, and strengtheningthe fiscal position of local
governments. In this case, a larger tax base with a more income-elasticgrowth
may be sought. Consumptionand incometaxes are shared between levels of
governmentin manycountries. But this is a big commitmentfor a developingor
transition country, and unless there is an intention to withhold distributionsin
times of budgettightness,it can seriouslycompromisethe fiscal flexibility at the
centrallevel.
Ad Hoc Transfers
6
1. It is not transparent,and quite subjectto politicalmanipulation.
7
In sum, the ad-hoc approachto determiningthe size of the distributable
pool is the most centralizingapproachto designingan intergovernmentaltransfer
system. Despite some very apparent flaws, It is widely used, even in some
countriesthat feature decentralizationas part of their developmentplan.
Cost Reimbursement
The transfer to cover these costs may be open ended, i.e., the
central govemment stands ready to cover the cost of all
expenditure incurred by the local government. More often, the
transfer is closed ended, i.e., the central governmentwill incur the
costs up to some maximum.
8
Cost reimbursementgrants are widely used as a method of determining
the total flow of funds to subnational governments. It gives the central
governmentcontrol over the amount of funds allocated to the local government
sector, and it gives the center some say in how the funds will be spent. It is a
centralizingapproachto intergovernmentaltransfers.
It is importantto note that this is a transfer and not a local tax, because
the local government has no control over the tax rate or the tax base. The
amount received by the subnational governmentis determinedfully by central
legislation. An alternative,tax base sharing, where the local governmentmay
piggybackon to a centraltax, is in fact a localtax and not a transfer. But among
the developingand transitioncountries,piggybackingis not yet common.
Some taxes are more suitable for derivation based sharing than others.
Much controversycentersaroundthe VAT as a choice for a sharedtax. A strong
argumentcan be madethat in most developingandtransition countries,the VAT
is not a suitablechoice. There are a numberof reasonsfor this. The first is that
the administrativeintegrity of the credit-invoiceVAT requires uniformity in the
definition of the tax base, and uniformity in the administrationof the tax across
the country. This cannot be guaranteedunder derivation based-sharingwhere
subnationalgovernmentsoften attemptto redefinethe tax base to betterfit local
conditions,or may administerthe tax differentlyin one provinceversus another.
9
A second reason has to do with protection-likebehavior of subnational
governments that might be induced by a derivation-sharedVAT. Provincial
governments, under pressure to create jobs and undertake development
enhancing projects, are tempted to institute polices to force producersto buy
from local suppliers. This increasesrevenue,increasesjobs, and shifts some of
the tax burden on to residentsof other provinces. It is not in the interest of
efficientnationalgrowth,but it does fit the objectivesof local politicians.
First, derivation based shared taxes are not equalizing. The richer local
jurisdictionshave the strongertax base and probablythe strongestadministrative
machineryfor collection. The result is that the disparities in taxable capacity
betweenrich and poor regionswill be widened.
The other side of the coin, however, is that derivation based sharing
leaves the central governmentwith less flexibility to make ad-hoc changes. If
local governmentscan identifytheir entitlementswith some certainty,then it may
be difficultfor the central governmentto make regular changesin the derivation
formulae.
FormulaGrants
What are the objectivesthat might drive the design of a formula grant?
The most commonreasonwhy governmentsmoveto formulaebased distribution
is to gain transparencyand certaintyin the distributionof grants. This createsa
sense of fairness in that all know the exact criteria by which distributionsare
made, and there is flexibility in that distributionsmay change as the needs for
public expenditureschange. In short, formulasare meantto removejudgement.
13
Some school aid in the U.S. is defined by a formula that links
minimum expenditure requirementswith property tax revenues
raised if a specifiedlevelof propertytax effort is exerted.
14
* Finally, there is the issue of the reliability of the data itself. The
accuracy of data is often questioned,even if gathered by official
bodies.
15
Evaluation. As noted above, one cannot evaluateformula grant systems
without consideringboth dimensionsof the grant: the method of determiningthe
total grant pool, and the formula used in the distribution. As may be seen from
Table 2, there are two formula grant systemsto be considered.A shared tax
distributedamong local governmentsaccordingto a formula is a Type B grant,
and an ad hoc pool distrustedby formula is a Type F grant.
A conditionalgrant will not give the same degree of local autonomyas will
an unrestrictedgrant. Cost reimbursementgrants usually carry restrictions on
the use of the funds such as standardsto which publicfacilities will be built, the
salary rates of public employees,etc. A cost reimbursementgrant based on a
sharedtax distributablepool will give more autonomythan either an ad hoc or a
cost reimbursementdetermineddistributablepool.
19
THE SPECIALQUESTIONOF EQUALIZATION
21
Table I
AlternativeFormsof IntergovernmentalGrant Programs
Table 2
TaxationChoices
*VAT
* CompanyIncomeTax
* IndividualIncomeTax
* ExciseTaxes
. InternationalTrade Taxes
* Retail Sales Tax
. PropertyTax
. Motor VehicleTaxes
* UserCharges
. CommercialVentures
22
REFERENCES
Bahl, Roy and Johannes F. Linn. 1992. Urban Public Finance in Developing
Countries.(NewYork: OxfordUniversityPress),pp. 385-427.
Bird, Richard M. and Duan-jie Chen. 1996. "Federal Finance and Fiscal
Federalism:The Two Worlds of Canadian Public Finance," Essays on
Fiscal Federalismand FederalFinancein Canada(Toronto:Universityof
Toronto).
23
Inter-AmericanDevelopmentBank. 1997. Latin America After A Decade of
Reforms: Economic and Social Progress. (Distributed by the Johns
HopkinsUniversityPressfor the Inter-AmericanDevelopmentBank).
24
Rao, M. Govinda. "IntergovernmentalFiscal Relations in a Planned Economy:
The Indian Case," Australia South Asia Research Centre, Research
School of PacificandAsian Studies,AustralianNationalUniversity,p. 17.
26
The World Bank
Urban Development
1818 H Street, NW
Washington, DC 20433
USA
Internet: www.worldbank.org/urban