Draft for Discussion Only
Federal-Provincial-Local Fiscal Transfers in Pakistan
Anwar Shah, World Bank
December 6. 2003
Matching revenue means as closely as possible to expenditure needs of the various levels
of government serves to strengthen accountability in a federal system. Allowing subnational government greater access to own tax bases accompanied by stronger tax
collection performance and better attention to cost recovery policies will help in this task.
In Pakistan, in view of the very large vertical imbalances – gap between expenditure
needs and revenue means at provincial and local levels – and the associated implications
for weak accountability, tax decentralization options require serious consideration to
narrow these imbalances without eliminating them entirely. This is because, it is
desirable in federal systems for higher level governments to have access to more
revenues than those dictated by their direct program responsibilities alone. These
additional revenues can be used to support national and provincial economic objectives
such as setting national standards, securing economic union and ensuring inter-regional
and inter-local fiscal equity. The design of these transfers is however, critical to
achieving the objectives sought. The issues pertaining to the design of these transfers are
critical to achieving the objectives sought. This note reviews the design of existing fiscal
transfers in Pakistan for their consistency in achieving the objectives being pursued.
Vertical fiscal imbalance in a federation reflects the revenues raised by the federal
government that are made available to provinces to finance their expenditures. There are
two broad ways this transfer of funds can take place. One is by assigning a predetermined share of federal revenues to the provinces, while the other is by making
federal-provincial transfers whose magnitude is based on criteria other than federal
revenues. The first is revenue sharing while the second falls under the general rubric of
federal-provincial transfers. The distinction is a conceptual one only, since both amount
to a transfer of funds and both share some common characteristics. In Pakistan, revenue
sharing is the dominant form of federal-provincial fiscal relations for financing operating
expenditures. Conditional grant programs are practiced primarily for capital projects that
are typically designed and determined at the federal level and local governments
implement these projects on behalf of their provincial governments.
Current Practice of Federal-Provincial Revenue Sharing in Pakistan
The task of the distribution of revenues between the federation and the provinces have
been assigned by Pakistan Constitution (article 160) to an intergovernmental, interlegislative cum civil society body, the so-called National Finance Commission (NFC).
The NFC comprises of the Finance Minister of the Federal Government as chair and
Provincial Finance Ministers and other persons (from provincial legislatures, experts,
academia, think tanks) appointed by the President after consultations with the provincial
1
Governors. The constitution mandates convening of National Finance Commission every
five years to decide (make awards) on (a) the proportion of the constitutionally specified
federal revenues that should go to the provinces; and (b) the criteria for the distribution of
these revenues among provinces. Previous quinquennial awards stay in force in the event
a consensus is not reached on the new award. Due to a lack of consensus on the 2002
award, the 1997 NFC award remains in vogue at the present time but efforts are currently
underway to reach a new consensus. For this purpose, a newly constituted National
Finance Commission is starting deliberations in December 2003.
The 1997 NFC Award
The 1997 NFC award (renewed in 2002) consisted of three components : (a) revenue
sharing component – distribution of a pool of federal revenues to provinces by formula;
(b) returning to the province of origin resource royalties, charges and excises after
deducting a 2% federal collection fee; and (c) special lumpsum transfers to NWFP and
Balochistan provinces. These are described as follows:
(a) Revenue Sharing Component
Taxes on income, wealth, capital value taxes, sales taxes on goods, federal excises
(excluding natural gas) and the customs duties, i.e. nearly all of federal consolidated
revenues ( after retaining a 5% collection charge) excluding income taxes on federal
employees constituted the divisible pool for the 1997 award. The NFC mandated transfer
of 37.5% of the divisible pool to the provinces on the basis of provincial share of the
national population. In addition, to these revenues, provinces currently receive an
additional 2.5% of general sales tax (GST) revenues as a pass-through funds for local
governments in lieu of their elimination of octroi tax. For FY2002-2003 total divisible
pool comprised of 448.5 billion Pak Rupees and provinces received 150 billion rupees as
revenue sharing transfers (see Table 1 for details).
INSERT TABLE 1 AFTER THIS PAGE
(b) Revenues returned by origin
Royalty on crude oil, royalty, excise duty and surcharge on natural gas and general sales
tax on services are returned by origin to the provinces after deducting a 2% federal
collection charge. In FY2003-2004, provinces received 33.4 billion rupees on account of
these revenues/
(c) Special Grants to Fiscally Disadvantaged Provinces
The NFC mandated payments of special lumpsum grants to NWFP and Balochistan
provinces in recognition of their special needs. In FY2002-2003, Rs. 20 billion in federal
revenues were transferred to the two provinces.
(d) Federal Transfers in lieu of Octroi and Zila Taxes (OZT)
2
Until 1998, taxes on inter-municipal trade represented major source of revenues for
district and tehsil governments in Pakistan. Octroi tax was imposed by urban
governments (current equivalent of Tehsil governments) goods and services imported
from other jurisdictions and the Zila tax was levied by rural governments (current
equivalent of district governments). In 1998 , these taxes were abolished at the direction
of federal government with a promise to compensate local governments for loss of
revenues through a federal transfer program that passed funds through provincial
governments to local governments. Currently 2.5% of revenues from the federal goods
and services tax are transferred to provincial governments for distribution to local
governments. In FY2002-2003, provinces received 14.5 billion rupees or 5.8% of total
transfers from the federal government as the OZT transfers.
Conditional Grant Programs
Federal government provides financing mostly for capital projects that are centrally
sponsored either using federal revenues or grants received from foreign governments and
agencies. In FY2002-2003, these transfers amounted to 15.6 billion rupees or 6.3% of
total transfers. Foreign grants comprise 75% of such financing (see Table 1). The
following grants programs were in vogue in FY2002-2003:
Conditional Capital Assistance Programs
Khushhal Pakistan Program (KPP): Under KPP, funds are made available to local
governments through provincial governments for small local capital projects. The federal
government specifies the criteria for project selection and a negative list for the use of
funds. The district governments decide on the project selection and implementation. In
FY 2002-2003, XX billion rupees were made available for this program.
Tameer-Watan Program (TWP).: Federal Government provides members of national
parliament (National Assembly) and the provincial governments entitle members of
provincial assembly, ten million rupees each to be spent on any project in their electoral
districts. These programs are implemented through district governments. In FY20022003, XX billion rupees were made available for this purpose.
Education Sector Reform Program (ESRP): This program makes available federal
financing primarily for rehabilitation of school facilities. The funds are made available to
the provinces as development grants who in turn allocate these to the district
governments. In FY2002-2003, X billion rupeeeswere transferred to the provinces on this
account.
The Extended Immunization Program (EIP): For the childhood immunization program,
the federal government provides in-kind contributions in the form of vaccines, vehicles
and other equipment. In FY2002-2003, xx million rupees were spent on this program by
the Federal Government.
Conditional Operating Assistance
3
Lady Health Worker Program (LHWP): Under the National Program for Family
Planning and Primary Health Care, thre federal government finances the salaries of lady
health workers jointly recruited by the provinces and the districts. In FY2002-2003, this
program cost the federal government xx million rupees.
The Pros and Cons of Pakistan’s System of Federal-Provincial Fiscal Relations
As noted earlier, Pakistan’s system of federal resource transfer to the provinces is
predominantly a revenue sharing system having important merits of a transparent fiscal
system as well as lacking important levers a well designed fiscal transfer system can
potentially offer. The smaller conditional grant programs employ less transparent
criteria. These arguments are discussed below:
Merits of the Existing Revenue Sharing System
Pakistan’s fiscal system is laudable on several counts.
The constitutionally mandated structure of the National Finance Commission represents
an excellent model of a federal-provincial participatory decision making body with all the
key stakeholders represented . The consensus decision rule further offers opportunity for
political compromise and accommodation. This model is superior to so-called
independent and largely academic grants commissions in India and Australia as the
decision makers are not represented on latter bodies. The practical merits of such an
institutional arrangements are borne out by the simplicity and transparency of the
commission awards.
The revenue sharing system practiced in Pakistan represents a simple, objective (formula
based) and transparent way of reasonably secure and growing amount of revenues to the
provinces. It provides transfers in a lumpsum, unconditional and predictable way to the
provinces. The provinces are left with full discretion and complete autonomy over how
to spend these funds. This facilitates the decentralization of fiscal responsibility and
contributes to the efficiency of the federal system. The provinces remain accountable to
their own constituents via the legislative process for the manner in which they provide
services.
The NFC award also facilitates the preservation of a fully harmonized tax system in
Pakistan. It has enabled the federal government to retain major revenue raising
responsibilities to achieve greater efficiency in tax collection and administration and
lower the compliance costs of the tax system. These arrangements have avoided tax
jungles that continue to block the reform of sales taxes in India.
The NFC awards equalize to an uncertain standard the fiscal needs of the provinces as
population represents a good proxy for fiscal needs.
4
The revenues returned by origin represents an acknowledgement by the federal
government that resource royalties are provincial own source revenues and simply
collected by the federal government.
The special grant program for NWFP and Balochistan provides additional unconditional
assistance to the two smaller provinces with lower than average fiscal capacities and
higher than average expenditure needs. This recognizes the differential needs arising
from a dispersal of population over a large area to be serviced in Balochistan and the
mountaineous terrain and influx of Afghan refugees in NWFP .
Drawbacks of the Existing System
For all its simplicity and effectiveness in getting funds into the hands of the provinces
with a minimum of intrusion, revenue sharing has some drawbacks especially if relied on
excessively as is the case in Pakistan. The main issue concerns the fact that provinces
receive these funds like manna from heaven with no accountability to tax payers. They do
not experience the pains of justifying additional taxes to their constituents while enjoying
the pleasures of spending on their pet projects. Due to their unconditional nature, and
rightly so, there is also no oversight requirements by the federal government.
A related concern is that while provinces are left with considerable discretion in the use
of revenue sharing funds, they have virtually no discretion over the amount of funds they
receive. If revenue sharing is seen as an alternative to other forms of federal-provincial
transfers, this is no drawback. But revenue sharing in Pakistan leaves little revenue
raising responsibilities with the provincial governments. In these circumstances tax
decentralization options – such as tax base sharing might offer superior alternatives.
Enhancing the ability of provinces to raise their own revenues can increase the
accountability of provincial governments for their fiscal performance.
Another important drawback of revenue sharing system is that it does not permit fiscal
capacity equalization to a national average standard. Such an equalization system fosters
a sense of unity in the nation as citizens benefit from having reasonably comparable
levels of public services at reasonably comparable levels of taxation across the nation.
Please note that the proposals currently under discussion in Pakistan to include
backwardness, tax effort and other indicators in the NFC formula would make the system
complex without achieving fiscal equalization.
An additional important drawback of the revenue sharing system is that it deprives the
federal government from having the lever to influence provincial priorities in order to
achieve national objectives of say having national minimum standards in education and
health to foster internal economic union. Such a system also works as an impediment to
introducing an accountability for results and performance orientation culture in the public
sector at the provincial level.
Another potential drawback of the existing system is that the formula determining a
province’s revenue allocation bears little relation to province’s expenditure
5
responsibilities. A province’s revenues grow at the rate of growth of federal revenues.
This may differ considerably from the rate of growth of province’s expenditure
responsibilities. Provincial expenditures on education, health, and infrastructure are likely
to far outstrip growth in federal revenues.
Finally, the existing revenue sharing program exposes the provinces to risks associated
with changes in federal tax bases and collection performance. Both these risks, however,
may be quite small as almost all federal revenues are included in the base.
The program of returning resource royalties by origin seems also difficult to justify as the
federal government collects revenues at centrally determined rates and then returns them
to provinces on the basis of collection (see also Shah, 1996). While it can be argued that
resource royalties as opposed to resource profit or rent taxes could be reasonably
assigned to the provinces, but then provinces should be able to decide on the rates and
structure of such royalties while collection could be either provincial or federal.
The special grant program for NWFP and Balochistan is also flawed as the size of these
grants are determined in an ad hoc manner and both the size and duration is unrelated to
any provincial fiscal capacities. This program weakens accountability of provincial
governments to their taxpayers without dealing with regional equity concerns in a
satisfactory manner.
Comments on the Conditional Grant Program
The conditional grant programs have meritorious objectives – dealing mostly with access
to basic services. Federal financing of programs that aim at expanding education, health,
nutrition and water and sanitation services, is desirable. The ad hoc and discretionary
nature of allocation of funds, focus on inputs alone with almost complete disregard for
outputs make it less likely that these meritorious objectives will be advanced. Instead,
micro management by federal and provincial administrations create potential for
financing projects which may not suit local circumstances. The funds to MPA and MNA
create an executive role for these legislators to the neglect of their role in providing
oversight and accountability. These funds provide incentives for log rolling and political
patronage. The total amount of funding available for conditional capital grant programs is
so small that it is unlikely to deal with infrastructure deficiencies for basic services. A
long term planning view to upgrade provincial-local services to setting national minimum
standards is absent from the existing system.
Restructuring Options for Federal-Provincial Fiscal Transfers
If tax decentralization and tax base options as advocated from time to time (see Shah,
1996) are adopted, the need for a revenue sharing program to deal with vertical
imbalances would be less urgent. Tax decentralization would, however, accentuate
regional fiscal disparities. Therefore the case for a fiscal equalization program that
enables all residents regardless of their place of provincial residence to have access to
reasonably comparable levels of public services at reasonably comparable levels of
6
taxation , would take prominence. In addition, there would be a need for the federal
government to have the ability to influence provincial policies through the use of its
spending power (fiscal transfers to provinces). Keeping these two objectives in mind, the
following long term options for the reform of fiscal transfers would be worthy of
consideration.
Long term Options
(i)
Fiscal capacity equalization program. This program will upgrade provincial
fiscal capacities to a national average standard. A representative tax system
approach to fiscal equalization is practiced in a small but growing number of
countries. Under such an approach the fiscal capacity of a province equals the
revenues that could be raised by a province if it applied national average tax
effort to all its tax bases. A federal program attempts to bring all provinces to
a national average per capita fiscal capacity (see Shah, 1994, 1996 for details
on the application of this approach). Among developing and transition
countries, Poland, Russia and Indonesia have adopted variants of this
approach.
(ii)
Operating transfers to set national minimum standards: These transfers
would be based upon simple demographic factors for distribution among
provinces and there would not be any conditions on the use of funds but there
would be expectations on the results (service delivery performance).
Compliance failure would invite public censure but persistent failure to
achieve agreed upon results could result in discontinuation of grant funds. An
example illustrating such a transfer program is given in the annex.
(iii)
Capital transfers and credit access for upgrading infrastructure to national
minimum standards. A planning view is required to determine minimum
needs for infrastructure for various jurisdictions and a menu of capital finance
(grants and loans) and public-private partnership choices to meet those needs.
Short Run options
While a serious attempt to develop a consensus on the long term fiscal system is made, it
would be desirable not to tinker with the current revenue sharing system with stop gap
measures, as currently under discussion by the new NFC, as such measures could
potentially block movement to a more desirable system. The capital grant system, on the
other hand, even in the short run, could be revamped to bring a planning perspective on
capital grant allocation and access to credit finance and a results or service delivery
perspective in the design of conditional operating assistance.
Provincial-Local Fiscal Relations in Pakistan
Local (district and tehsil) governments in Pakistan historically were run on selfsustaining basis with little (about 5% of total expenditures) assistance from the
7
provincial governments. Two recent events brought about major changes. First in 1998,
local governments were directed to by the federal government to abolish octroi taxes
against a federal IOU of receiving equivalent (2.5% of GST) revenues as grants. Second,
the local devolution in 2001 brought a significant decentralization of expenditure
responsibilities especially for social services. Again, this was to be accompanied by a
system of unconditional fiscal transfers from the provinces determined by provincial
finance commissions. These commission are typically chaired by the provincial finance
minister and have as members: additional chief secretary, planning and development,
secretary finance, accountant general, and two or more private citizens as members. The
Commission has a limited tenure with a specific mandate but the commission secretariat
is a permanent body usually housed in the finance department. The mandate of the
commission is usually much broader than fiscal transfers only and covers most areas of
provincial-local fiscal relations, All provinces have now established provincial finance
commissions and these commissions have made awards for the distribution of provincial
transfers to local governments. For both operating and capital transfers to district
governments, the provincial governments determine the divisible pool and the
commission recommends criteria for distribution of these funds to local governments. For
operating transfers to tehsil governments, total pool is simply the federal transfers
received in lieu of octroi and zila taxes and the province determines the criteria for
distribution of funds among tehsil governments. The following paragraphs provide
further elaboration of these programs.
Provincial Operating Transfers to District Governments
These are the largest unconditional formula based grant programs from the Provincial to
local governments. Nearly 40% of provincial revenues are transferred to districts
through these programs. The definitions of the divisible pools vary slightly but allocation
criteria vary a great deal across provinces as discussed below.
Provincial Divisible Pool
What constitutes a divisible pool for allocation of local share varies somewhat across
provinces with Punjab including all provincial revenues in the pool but other provinces
making some exceptions as noted below:
Provincial divisible pool by province:
Punjab: Provincial revenues from all sources minus statutory of “charged” expenditures
(Governor’s House, Provincial Assembly and the High Court) and debt charges,
pensions and subsidies.
Sindh: Provincial receipts from federal revenue sharing; federal revenues retuned by
province of collection; provincial tax revenues. Provincial non-tax revenues are excluded.
NWFP: Same as in Punjab. Provincial revenues from all sources minus provincial
obligatory expenditures. The latter include: interest payments, debt repayments, pensions,
subsidies, contributions to General Provident and Pension Funds and charged
expenditures on the Governor’s House, Provincial Assembly and High Court.
Balochistan: Provincial revenues from all sources except revenues from 2.5% of GST.
8
Distribution Criteria
Interestingly enough all provinces have adopted a formula based revenue sharing
approach for the transfer of funds from the divisible pool to the district level
governments. (see Table 2). None of the provinces as of this date has considered an
objective based fiscal transfer system.
Table 2: Provincial Operating Transfers to Districts (Provincial Finance Awards )
- FY 2002-2003
Total pool and distribution criteria
Punjab Sindh NWFP Balochistan
Local share of the Provincial Divisible Pool
39.8% 40%
40%
31%
Formula factors with weights:
100%
100% 100%
100%
Population
75%
50%
50%
50%
Backwardness
10%
17.5% 25%
Tax effort
5%
7.5%
Fiscal austerity
5%
Area
50%
Development incentive/ infrastructure
5%
25%
deficiency
Hold harmeless/ transitional assistance
25%
Source: Author’s stylized view of PFC awards
Population is the most important indicator used in all provincial awards. Backwardness
index is used by three of the four provinces and tax effort provisions are incorporated by
the two largest provinces. Balochistan gives 50% weight to district area and NWFP
assigns 25% weight to infrastructure deficiency. In all cases grant funds vary directly
with fiscal need but fiscal capacity has no influence on grant entitlements.
Provincial Operating Transfers to Tehsil and Union Council Governments
Provincial operating transfers to tehsils and union councils are simply based on the
revenues received by the province from federal grants in lieu of octroi and zila taxes.
Various provinces have adopted different allocation criteria for allocation of these funds
to tehsil and union council governments
Punjab: 88.5% of OZT transfers of 6,962 million rupees were allocated to tehsils based
upon historical shares. 4.3% each were allocated to union councils and “poor” tehsils and
the remaining 2.9% as special grants to selected local governments.
Sindh: 65% of OZT transfers of 10,000 million were distributed to tehsils using data on
historical shares of collection of OZT. 28% distributed as a replacement for the KPP
program and of these 70% allocated to tehsils and 30% to Talukas (Union Councils) by
giving 70% weight to population and 30% to backwardness.
9
NWFP: 90% of the OZT revenues (10% of OZT given to districts using the PFC formula
for operating grants) were transferred to tehsils (using historical shares) and union
councils (equal amounts).
Balochistan: OZT revenues were distributed to district , tehsil and union councils on the
basis of 40:25:35 percent of total respectively. The PFC award formula was used for
distribution to governments within each category.
Provincial Capital Transfers to Local Governments
Punjab: Total pool for these transfers is determined arbitrarily based upon available
funds from the provincial ADP (annual development plan) allocations. 75% of total pool
is for distribution to districts and the remaining 25% to tehsils. Two thirds of these funds
are distributed by population and the remaining by “backwardness”.
Sindh: 70% of available ADP funding is distributed to districts using the PFC award
formula for inter-district allocation.
NWFP: 60% of provincial ADP funds are transferred to districts. Of these 90% are
distributed among districts using PFC award criteria for capital transfers ( Population
:50%, Backwardness:30%, backlog for ongoing schemes 10%; equal share 10%). The
remaining 10% of the funds are given as special grants for the completion of on-going
projects.
Balochistan: Contrary to the practice in other provinces, Balochistan does not earmark a
share of ADP as development transfers to local governments. Allocation of any available
monies on account of capital transfers to local governments is based upon population
(50%) and service area (50%).
Provincial-District-Tehsil-Union Council Fiscal Transfers: A Qualitative Evaluation
The revenue sharing system at the provincial level has similar potentials as well as perils
as those highlighted for the federal system earlier. In the following, salient merits and
demerits of the provincial revenue sharing system are highlighted.
Merits of the system of Provincial-Local Revenue Sharing System in Pakistan
The system of revenue sharing adopted by the provinces is commendable for several
reasons. The new institutional arrangements for determining the divisible pool and the
distribution criteria are transparent and a large pool of resources are being transferred to
local governments using reasonably objective methods. The formula factors primarily
emphasize various indicators of local expenditure needs. The emphasis on these factors
had a positive impact on the provincial information base of local governments and their
fiscal needs. Local government enjoy full autonomy in the use of available transfers.
10
Demerits of the Provincial-Local Revenue Sharing System
The new revenue sharing system has important limitations.
1. The one size fits all approach used by the provincial finance commissions of all
provinces is seriously flawed. Under this approach all local jurisdictions are
treated alike creating significant inter-jurisdictional fiscal inequities. Large city
districts with a larger menu of local services and very different needs are lumped
together with smaller largely rural districts. OZT transfers to tehsils are primarily
based upon historical shares and no clear recognition of current responsibilities
for municipal services.
2. The awards completely ignore the differential fiscal capacities of various local
governments. This neglect makes the distribution of revenues solely based upon
expenditure needs inequitable and indefensible.
3. The awards do not fully match the transferred expenditure responsibilities at the
local level leaving relatively higher revenues with the provinces than required by
their newer responsibilities. Thus vertical imbalances persist even after the
transfer of significant resources to local levels. Such imbalances impede the
realization of improvements in service delivery associated with decentralized
management.
4. The awards have the potential to weaken local government accountability to
resident taxpayers. Such bottom-up accountability is considered critical for the
success of any decentralization program. With devolution, local government
dependency on provincial and federal transfers has become overwhelming. This
creates strong incentives for local elected officials (like their provincial
counterparts in the case of NFC) to be much more focused on influencing
provincial distribution criteria at the neglect of local revenue raising efforts.
5. Finally and most importantly, the provincial finance commissions have failed to
recognize the role of fiscal transfers in enhancing efficiency and equity of the
federal system by providing the right incentives for responsive, responsible and
accountable governance (see Annex 1). In fact, the finance commissions are
focused primarily on dividing the pie have given no consideration to creating an
enabling framework for improved local service delivery performance and better
local governance.
Restructuring Options for the Reform of Provincial-Local Fiscal Transfers
The system of provincial-local fiscal relations requires a systemic examination. The
impressionist examination presented in this note offers the following conclusions.
a. Pakistan needs a careful examination of tax decentralization options at the local
levels. Provincial Local Governments Acts, 2001 have advanced this agenda but
much more work remains to be done.
11
b. To achieve minimum standards of local services across various municipalities and
to introduce competition and innovation in service delivery and accountability for
results at the local levels, the need for introducing local fiscal equalization (see
Annex 1 for a discussion of fiscal capacity equalization transfers) and
performance oriented fiscal transfers for education, health and infrastructure (see
Box 1 for an example of a performance oriented transfer) can hardly be
overstated. These transfers must be separately constituted for each type of
districts: city districts, large, medium and small districts and tehsils of various
population sizes.
c. Capital transfers also need to take planning view about the requirements to
establish minimum standards across jurisdictions and the potential sources of
capital finance including grants and provincial assistance for credit market access
where feasible.
Box 1. An example of a performance oriented grant: education grant to set
minimum standards, while encouraging competition and innovation
Allocation basis among local governments: school age population.
Distribution to providers: equal per pupil to both government and private schools.
Conditions: universal access to primary and secondary education regardless of parents’ income;
improvements in achievement scores; no condition on the use of grant funds.
Penalties for non-compliance with standards: public censure, reduction of grant funds.
Incentives for cost efficiency: retention of savings.
Source: Shah (2002)
12
Annex 1
Designing Fiscal Transfers: Dividing the spoils or creating an enabling
framework for innovative and competitive service delivery1
Intergovernmental transfers are the dominant source of revenues for subnational governments in most developing and transition economies (DTEs). The
design of these transfers is of critical importance for efficiency and equity of local
service provision and fiscal health of sub-national governments (see Appendix
table 4 for general principles and better practices in grant design). For enhancing
accountability it is desirable to match revenue means (the ability to raise
revenues from own sources) as closely as possible with expenditure needs at all
levels of government. However, higher level governments must be allowed
greater access to revenues than needed to fulfill own direct service
responsibilities so that they are able to use their spending power through fiscal
transfers to fulfill national and regional efficiency and equity objectives. We can
identify six broad objectives for national fiscal transfers each of which may apply
to varying degrees in different countries and each of which calls for a specific
design of fiscal transfers.
i. To bridge fiscal gap:
A fiscal gap is defined as an imbalance between the revenue-raising ability of
regional and local governments and their expenditure responsibilities. To deal with
fiscal gap, it is important to deal with the sources of this gap through reassignment
of responsibilities, tax decentralization or tax abatement by the center and tax base
sharing (by allowing sub-national governments to levy supplementary rates on a
national tax base). Only as a last resort unconditional formula based transfers or
revenue sharing based on the origin/derivation (point of collection) principle may
be considered as options to deal with this gap. This is because such transfers
weaken accountability to local taxpayers. In DTEs, general revenue sharing are
typically used to deal with fiscal gap. A number of countries including China, India,
Malaysia, Pakistan, South Africa have in the past tried deficit grants to fill fiscal
gaps at sub-national levels with unwelcome results in terms of mushrooming of
sub-national deficits. These grants are still in vogue in China, Hungary and South
Africa.
ii. To correct fiscal inequities and fiscal inefficiencies arising from differentials in
regional fiscal capacities
Decentralized decision making results in differential net fiscal benefits
(imputed benefits from public spending minus tax burden) being realized by
citizens depending upon the fiscal capacities of their place of residence. This leads
both to fiscal inequity and fiscal inefficiency in resource allocation. Fiscal inequity
arises as citizens with identical incomes are treated differently depending on their
place of residence. Fiscal inefficiency in resource allocation results from people in
1
This is based upon Anwar Shah (2002). Fiscal Decentralization in Developing and Transition
Economies. In Raoul Blindenbacher and Arnold Koller, Editors (2002). Federalism in a Changing World
– Learning from Each Other. McGilli-Queen’s University Press, Montreal and Kingston, Canada.
13
their relocation decisions comparing gross income (private income plus net public
sector benefits minus cost of moving) at new locations whereas economic
efficiency considerations warrant comparing private income minus moving costs.
Thus a nation which values horizontal equity (i.e., the equal treatment of all
citizens nationwide) and fiscal efficiency will need to correct the fiscal inequity and
fiscal inefficiency which naturally arise in a decentralized government..
Central-state grants can eliminate these differences in NFBs if the transfers to
each region depend upon the tax capacity of the region relative to others and upon
the relative need for and cost of providing regional public services. The more
decentralized the tax system is, the greater the need for equalizing transfers.
Most transition economies have equalization components in their grant
programs to sub-national governments. Latvia, Lithuania, Poland, Romania,
Russia and Ukraine have adopted transfer formulae that explicitly incorporate
either fiscal capacity and/or expenditure need equalization concerns.
In
developing countries, programs using an explicit standard of equalization are
untried, although equalization objectives are implicitly attempted in the general
revenue sharing mechanisms used in Argentina, Brazil, Colombia, India, Nigeria,
Mexico, Pakistan and South Africa. These mechanisms typically combine diverse
and conflicting objectives into the same formula and fall significantly short on
individual objectives. Because the formulae lack explicit equalization standards,
they fail to address regional equity objectives satisfactorily.
iii. To compensate for benefit spillovers.
This is the traditional argument for matching conditional grants. Regional
and local governments will not have the proper incentive to provide the correct
levels of services which yield spillover benefits to residents of other jurisdictions. A
system of open-ended matching grants based on the expenditures giving rise to
the spillovers will provide the incentive to increase expenditures. Typically, the
extent of the spillover will be difficult to measure so the correct matching rate to
use will be somewhat arbitrary.
Although benefit-cost spill-out is a serious factor in a number of countries
such transfers have not been implemented in developing countries with the single
exception of South Africa. South Africa provides a closed-ended matching grant to
teaching hospitals based upon an estimate of benefit spillovers associated with
enrollment of non-local students and use of hospital facilities by non-residents.
iv. Setting national minimum standards to preserve internal common market and
attain national equity objectives.
Setting national minimum standards in regional-local services may be
important for two reasons. The first is that there is an advantage to the nation as a
whole from such standards as these will contribute to the free flow of goods and
services, labor and capital and reduce wasteful inter-jurisdictional expenditure
competition, and will therefore improve the gains from trade from the internal
common market. Second, these standards serve national equity objectives. Many
public services provided at the sub-national level such as education, health and
social welfare are redistributive in their intent, providing in-kind redistribution to
residents. In a federal system, lower level provision of such services – while
14
desirable for efficiency, preference matching, and accountability – create difficulty
in fulfilling federal equity objectives. Factor mobility and tax competition create
strong incentives for lower level governments to under-provide such services and
to restrict access to those most in need , such as the poor and the old. This is
justified by their greater susceptibility to disease and potentially greater risks for
cost curtailment. Such perverse incentives can be alleviated by conditional nonmatching grants where the conditions reflect national efficiency and equity
concerns, and where there is a financial penalty associated with failure to comply
with any of the conditions. Thus conditions will not be on the specific use of grant
funds but attainment of standards in quality, access and level of services. Such
grants do not affect local government incentives for cost efficiency but do
encourage compliance with nationally specified standards for access and level of
services. Properly designed conditional non-matching transfers can create
incentives for innovative and competitive approaches to improved service delivery
(see Box 1 in the main text for an example of such a grant).
Conditional non-matching transfers to ensure national minimum standard are
rarely used in DTEs. Central government transfers to provincial and local
governments in Indonesia, central per capita transfers for education in Colombia
and South Africa, and the capitation grant to Malaysian states come close to the
concept of such a transfer.
v. Influencing local priorities in areas of high national but low local priority
In a federation, there is always some degree of conflict among priorities
established by various levels of government. One way to induce lower level
governments to follow priorities established by the higher level government is for
the higher level government to use its powers of the purse, the so-called spending
power, by using matching transfers. Open-ended matching transfers with matching
rate (percent of expenditures financed from own sources by the recipient) to vary
inversely with fiscal capacity would be consistent with this use. The use of ad hoc
grants or open-ended matching transfers for local tax effort would be inadvisable.
The former is unlikely to have behavioral responses consistent with grantor’s
objectives and the open-ended nature of the latter may create budgetary difficulties
for the grantor. India, Malaysia, and Pakistan use conditional closed ended
matching programs. Pakistan in late 1990s got into serious difficulty by offering
open-ended matching transfers for provincial tax effort. Central government had to
abandon this program in midstream as it could not meet its obligations under the
program.
vi. To create macroeconomic stability in depressed regions
Fiscal transfers can be used to serve central government objectives in regional
stabilization. For this purpose capital grants would be appropriate provided funds
for future upkeep of facilities were available. Experience with capital grants shows
that such grants often create facilities which are later not maintained by subnational governments as they either remain unconvinced of the utility of such
facilities or do not have the means to provide a regular upkeep.
Capital grants are pervasive in DTEs and most countries have complex processes
for initiation and approval of submissions for financing capital projects. These
processes are greatly susceptible to lobbying, political pressures and
15
grantmanship and favor projects that give the central government greater visibility.
The projects typically lack citizen and stakeholder participation and often fail due to
proper local ownership, interest and oversight. In view of this difficulty, it may be
best to limit the use of capital grants by requiring matching funds from recipients
and by encouraging private sector participation in infrastructure by providing
political and policy risk guarantees.
Special issues in state/province-local transfers
General purpose transfers to local governments requires special
considerations as local governments vary in population, size, area served and
the type of services offered e.g. urban vs rural. In view of this, it would be
advisable to classify local governments by population size, municipality type, and
urban/rural distinction and have a separate formula for each class of
municipalities. Some common useful components in these formulae are: equal
per municipality component, equal per capita component, service area
component and fiscal capacity component. The grant funds should vary directly
with service area but inversely with fiscal capacity.
16
Annex Text Table 1. Criteria for the design of intergovernmental fiscal transfers
Autonomy. Subnational governments should have complete independence and flexibility in
setting priorities, and should not be constrained by the categorical structure of programs and
uncertainty associated with decisionmaking at the center. Tax base sharing – allowing
subnational governments to introduce their own tax rates on central bases, formula-based
revenue sharing, or block grants – is consistent with this objective.
Revenue Adequacy. Subnational governments should have adequate revenues to discharge
designated responsibilities.
Equity. Allocated funds should vary directly with fiscal need factors and inversely with the taxable
capacity of each jurisdiction.
Predictability. The grant mechanism should ensure predictability of subnational governments’
shares by publishing five-year projections of funding availablility.
Efficiency. The grant design should be neutral with respect to subnational government choices of
resource allocatin to different sectors or different types of activity.
Simplicity. Grant allocation should be based on objective factors over which individual units
have liitle control. The formula should be easy to comprehend so that grantsmanship is not
rewarded.
Incentive. The design should provide incentives for sound fiscal manAgement and discourage
inefficient practices. There should be no specific transfers to finance subnational government
deficits.
Safeguard of garntor’s objectives. This is best done by having grant conditions specify results
to be achieved and giving the recipient flexibility in the use of funds.
Singular focus. Each grant should be focused on a single objective.
The various criteria specified above could be in conflict with each other and therefore a grantor
may have to assign priorities to various factors in comparing design alternatives.
Source: Anwar Shah (1994). The Reform of Intergovernmental Fiscal Relations in
Developing and Emerging Market Economies. Washington, DC: World Bank
17
Annex Text Table 2: Principles and better practices in grant design
Grant Objective
Grant Design
Better Practices
Practices to avoid
To bridge fiscal gap

Reassign responsibilities

Tax abatement
Tax abatement and tax base
sharing in Canada
Deficit grants
Tax by tax sharing

Tax base sharing
General revenue sharing with
multiple factors
To reduce regional
fiscal disparities
General Non-matching Fiscal capacity
equalization transfers
Fiscal equalization programs of
Australia, Canada and Germany
To compensate for
benefit spillovers
Open-ended matching transfers with
matching rate consistent with spill-out
of benefits
Republic of South Africa grant
for teaching hospitals
Setting national
minimum standards
Conditional non-matching block
transfers with conditions on standards of
service and access
Indonesia pre-2000 roads and
primary education grants
Colombia and Chile education
transfers
Conditional transfers with
conditions on spending alone
ad hoc grants
Influencing local
priorities in areas of
high national but
low local priority
Open-ended matching transfers (with
preferably matching rate to vary
inversely with fiscal capacity)
Matching transfers for social
assistance as in Canada
ad hoc grants
Stabilization
Capital grants provided maintenance
possible
Limit use of capital grants and
encourage private sector
participation by providing
political and policy risk
guarantee
Stabilization grants with no
future upkeep requirements
Source: Shah (1994, 1998).
18