Information about Asia and the Pacific Asia y el Pacífico

16 Horizontal Equalization Grants

Ehtisham Ahmad, Vito Tanzi, and Qiang Gao
Published Date:
September 1995
  • ShareShare
Information about Asia and the Pacific Asia y el Pacífico
Show Summary Details
Jon Craig*

Chapter 16 provides a broadly based rationale for the creation of a system of grants. This chapter concentrates on just one aspect of such a grants system: the implementation of equalization grants designed to offset differences in the fiscal capacities of different regions.

Putting Horizontal Equity in Perspective

Within a country with multiple levels of government, it is usual to have both vertical and horizontal imbalances in fiscal arrangements. Vertical imbalances within a fiscal system derive from mismatches in the relative revenue and expenditure assignments or responsibilities established for the two levels of government. The conference sessions on expenditure and revenue assignments yielded insights into the forces molding the relative responsibilities of different levels of government in market economies. Each country must wrestle with conflicting pressures: the need to centralize, to improve macroeconomic stabilization policies and management of overall expenditure priorities, and the desire to decentralize, to improve the quality of service delivery and allow the population to enjoy the benefits flowing from diversity.

While the solutions found for each country will differ, a number of countries have found that there are advantages in centralizing policy control, administration, and collection of broad-based taxes, such as value-added taxes, excises duties, and personal and corporate taxes, at the national level. As a consequence, these countries will normally have a vertical balance favoring the central government. Table 1 shows the vertical balances in a number of major industrial countries.

Table 1.Vertical Current Balances1(Ratio of Own-Source Revenues to Own-Source Current Expenditures)
Level of Government
“Federal” countries
United States0.930.971.241.410.660.75
“Unitary” governments
United Kingdom1.231.260.450.46
Source: IMF, Government Finance Statistics Yearbook (Washington: IMF, 1993).

The data show average ratios over selected periods for each country. The periods chosen are Australia, 1987–91; Austria, 1987–91; Brazil, 1982–91; Canada, 1985–39 (excludes 1987forcapital balance); Denmark, 1987–91; France, 1988–92; Germany, 1983–91; India, 1985–92; Netherlands, 1988–92; Spain, 1987–90; Sweden, 1988–92; United States, 1987–92; and United Kingdom, 1985–92.

Source: IMF, Government Finance Statistics Yearbook (Washington: IMF, 1993).

The data show average ratios over selected periods for each country. The periods chosen are Australia, 1987–91; Austria, 1987–91; Brazil, 1982–91; Canada, 1985–39 (excludes 1987forcapital balance); Denmark, 1987–91; France, 1988–92; Germany, 1983–91; India, 1985–92; Netherlands, 1988–92; Spain, 1987–90; Sweden, 1988–92; United States, 1987–92; and United Kingdom, 1985–92.

By contrast, the horizontal balance within a multilevel system of government relates to the inequality of fiscal capacities between regional subnational governments. Those inequalities stem from differences in relative expenditure needs in different regions due to population compositions, scale, distance, and other factors, as well as differences in the relative revenue-raising capacity of regions from the tax base available to regional governments. The end result of such differences is that some regions find themselves unable to provide comparable levels of service to their citizens without imposing substantially different levels of taxation.

In principle, these two imbalances within a fiscal system could be resolved by separate policy measures: (1) vertical imbalances could be resolved by a variety of tax-sharing and grant mechanisms with the regional governments, constructed in such a way so as to leave any “desired” or “accepted” horizontal balance unchanged; and (2) horizontal imbalances could be resolved by payments from “richer” regions with higher fiscal capacity to poorer regions (a feature of the German intergovernmental relations).

In practice, the two issues are often intertwined. Measures taken to resolve vertical imbalances inevitably have effects on the horizontal balance, and pressures to resolve horizontal imbalance often call forth actions by the central, as well as regional, government, which affect the vertical balance.

The aim of the policymaker concerned with these imbalances in finances must be to develop measures that address both objectives simultaneously within an overall budgetary setting that allows fiscal policy to make some ongoing contribution to the stabilization of the economy.

Why Is Horizontal Balance Important?

In order to understand the case for horizontal fiscal equalization, it is useful to think of why various independent regions might agree to sacrifice some or all of that independence and choose to join together into a “nation” or some other common economic space. While there may be a number of historical and cultural ties behind such actions, the most likely economic factor driving such unions is a desire to create a national marketplace that generates a “surplus” over and above the real income generated by the component parts.1 If this were not so, there would be an economic incentive for regions to become independent with their own laws and policies. Such arrangements would preserve their regional diversity at the expense of unity. The residents of such separate and independent regions could have no expectation of equity with their neighbors.

Assuming that there are demonstrable benefits sufficient to underpin an economic union, three possible forms of intergovernment arrangements might be considered. The first would be to abolish all regional governments and form a unitary government to control the whole area. All citizens would then be subject to the same laws and policies, with common taxes and service provision of public services. The goals of unity and equity would be achieved, but the nation formed would lack decentralization and diversity in service provision and taxation.

The second approach would be to maintain regional governments but not to make any effort to equalize for horizontal inequities. The nation then has unity and some diversity but, to the extent that different fiscal capacities exist in different regions, no equity.

The third approach would involve the preservation of regional governments plus a set of horizontal fiscal capacity equalization arrangements. The unity of the nation is preserved. Regions maintain a degree of diversity yet the goal of interregional equity is preserved.

The economic “case” for equalization therefore rests mainly on the desire to build a national structure that can simultaneously attain the key goals of unity, diversity, and equity. The issue of “interregional equity” is not simply a question of perceived social justice, however, important as that issue may be in its own right, it is also necessary to maximize the efficient use of mobile resources in an economy. Failure to introduce fiscal systems that broadly treat “equals as equals” across regions can induce flows of population that lower the overall national productivity of the workforce.

To grasp the essence of this point, consider the case of two identical individuals living in different regions, each of which provides identical public services. Region A has access to natural resources and can fund its budget expenditure from mining royalties, while region B has to rely on local income taxes. The individual in region A receives the full benefit of public expenditures in addition to the benefits derived from retaining all his personal income, whereas the person in region B must pay through personal income taxes for the benefits derived from public expenditures. Clearly, a net fiscal benefit is conferred on the resident of region A. A horizontal inequity has been created because individuals who were equal in the absence of a public sector are treated differently after the injection of fiscal measures. If residents of region B respond to this horizontal inequity by migrating to region A, they may be leaving more productive activities for less productive activities to capture the higher net fiscal benefits. Such fiscally induced migration may be undesirable on the grounds of economic efficiency.2

The political aspect of equalization has also been important. Fiscal inequity as between regions, resulting from the absence of equalization payments, is likely to cause political disharmony, economic dislocation, and fiscally induced population and capital movements. It was such political instability in Australia, associated with strong secession movements in three of the six states that led to the development of a fiscal capacity equalization under the oversight of an independent Australian Grants Commission in 1933. The continuation of the German and Indian systems of intergovernmental arrangements may also owe something to the presence of arrangements containing some elements of fiscal equalization.

Capacity Equalization and Performance Equalization

In the discussion above, equalization is defined in terms of the relative fiscal capacities of the regions. Specifically, the aim of the equalization process is to provide each region with a share of a general purpose unconditional grant such as to enable it to have the capacity to provide a comparable level of services to other regions, provided it makes the same effort to raise revenue from its own tax bases and conducts its affairs with an average level of efficiency.

Capacity equalization must be clearly distinguished from fiscal performance equalization. This approach relies on providing each region with a specific-purpose (usually) conditional grant capable of providing a specified minimum standard of public service.

Capacity Equalization Model

The most generalized model of capacity equalization can be formulated as follows:


G = the per capita equalization payment to a region by the granting government;

E = the per capita expenditure of the regional government concerned;

R = per capita revenue collections from own sources of the regional government concerned;

i = this subscript denotes the regional government involved; and

s = a national “standard” against which E and R have been compared in order to eliminate differences that arise purely from policy decisions as opposed to underling regional differences in revenue-raising capacity or the cost of providing services.

The means used to set the national standard vary considerably between countries. In Canada, for example, the standard for nationwide revenue equalization is based on revenue capacity in 5 eastern provinces (out of 12 provinces). A similar approach was used by Australia in the initial phase of its grants system, but now the grants are calculated against a standardderived as a weighted average standard of actual per capita revenues and expenditures for the six states and two territories.

This general model effectively tells us that the equalizing grant to a region will be equal to its relative expenditure needs, Eis, on the one hand, and its relative revenue capacity, Ris, on the other—each being measured against a national (average) standard. But there are many variants to this basic model. Some models only involve equalization of expenditures, usually by way of some conditional grant mechanism. Others involve partial equalization just in respect of revenues and ignore the needs created by differences in the cost of providing comparable services.

Revenue capacities are assessed by investigating the potential tax bases available to each region for a number of standard taxes and then comparing estimates of the revenue that could be raised by each region if the region imposed taxes at a national standard rate3 to its own tax base.

Calculating relative expenditure needs involves examining the underlying demand and cost factors that determine the cost of providing a standard level of service in each functional area in each region. In the case of education, for example, it may examine regional differences in demand and cost arising out of such factors as

  • differences in the ratio of school-aged children to total population or in the proportions of students in different age classes;
  • economies or diseconomies of scale arising out of administration of different school sizes;
  • population dispersion resulting in different costs of providing services for scattered populations or populations in remote areas;
  • socioeconomic factors, such as differences in the ethnic or socially disadvantaged composition of students in some regions and climatic or physical environment differences in the cost of maintaining school buildings, transporting students, and so on.

Performance Equalization Model

Performance models provide equalization grants against a set of externally determined standards—typically minimum standards—that are seen as “warranted” for various expenditure areas. Although the models are typically restricted only to expenditure analysis, it is possible to incorporate allowances that further adjust the grants according to the relative revenue-raising performance of a region.

The expenditure performance model usually determines the grant payable to a region as being equal to the product of the number of units (e.g., students) the region is required to serve and the standardized unit cost per unit. The standardized unit cost is equal to the standard cost determined for all regions as a whole plus an adjustment for the differential cost to the particular region concerned arising out of its own special circumstances.

The following example provides a formula that might be used for an education grant.

Nt = number of students enrolled with designated enrollment criteria for school “i

ts = standard teacher-student ratio;

ws = standard average salary cost per student; and

ks = standard ratio of salary cost to total cost.

Appropriate disability cost loadings may then be applied to each of the factors influencing unit cost, as follows:

where vij = differential cost of element “j” per unit met by grant recipient “i” “in providing standard service’s.”

Comparing Performance and Capacity Equalization

The difference between the two approaches lies essentially in the method of establishing the standard against which regions are to be “equalized.” The performance approach sets the standard by reference to criteria that are exogenous to the actual revenue and expenditure operations of the regions as a whole. Typically, the standard will be set arbitrarily by the central government on the basis of judgments as to the desired minimum set of expenditure conditions within particular functional areas. One important corollary of performance models is that they are often associated with some conditionality on the grants provided. That follows because the granting governments want to make sure that funds are used to achieve the minimum objectives established. In that sense, they may be seen to be inconsistent with the concept of decentralized multilevel finances discussed earlier. A country employing conditional grants to its regions is behaving similarly to a unitary government; as such it may attain some degree of equity in its fiscal system, but it will sacrifice decentralized responsibility and diversity.

By contrast, the capacity models determine the standard endoge-nously. That is, the standard is usually some average or weighted average of the actual fiscal behavior by the regional governments concerned. In Canada, for example, the standard for nationwide revenue equalization is based on revenue capacity in 5 “wealthier” provinces (out of 12 provinces). A similar approach was used by Australia in the initial phase of its grants system, but now grants are calculated against a weighted standard.

In contrast to performance models, capacity utilization grants usually do not involve conditionality. Regions are given grants that allow them to achieve certain standards, but they can choose to do more (by taxing at above standard rates and spending those funds on above standard expenditures) or less.

In practice, the two approaches may not yield widely dissimilar results. Broadly similar information and judgments may be required for either approach, although the degree of precision required for capacity equalization is normally greater because of the necessity of constructing accurate measures about relative capacities of regions. Of course, that concentration on precision is itself directed at reducing the degree of judgment used.

Constructing Intergovernmental Relationships

In constructing intergovernmental relationships to deal with vertical and horizontal imbalances, it is important to develop an overall framework for analysis of these twin balances and, within that framework, to establish some priorities between the conflicting objectives, as it may not be possible to resolve both the vertical and horizontal balance issues satisfactorily within a shorter-term planning horizon. Seven steps can be defined.

First, the vertical balance must be quantified against proposed revenue and expenditure assignments. That quantification must be adjusted for any cyclical impact on revenue and expenditures to ensure that the resulting fiscal balance is sustainable over time.

Second, measures must be set down to resolve the vertical imbalance. Which revenues will be shared? What system of grants can be used? If some revenues are to be shared, will the items be shared regionally according to the source of derivation (e.g., a percentage share of all enterprise profit taxes collected within a region)? Or will some regional reallocation mechanism be imposed (e.g., revenue share “caps” or other differential sharing or grants formula) be introduced that leads to differences between collections of revenue and amounts retained within each region? Or will a mechanism be adopted that allows regions to “piggyback” on the back of national tax (e.g., to introduce separate personal income tax levies set by regional governments)? Similarly, will grants be on a simple per capita proportional basis or involve some reallocation between regions?

In performing this task, the implications of the measures chosen to resolve the vertical balance for the horizontal balance must be considered. For example, if regions are allowed to keep a percentage share of all collections made within their borders for particular taxes (the so-called derivation approach to tax sharing) or to impose a piggyback levy on a national tax base, it may benefit richer regions with higher-tax capacities relative to other regions.

Third, an objective statistical framework must be devised to measure the existing horizontal balance between regions. It is necessary to establish a reliable comparison base—often referred to as a representative or standard budget that allows valid interregional comparisons of fiscal capacities. Decisions must be made as to whether the comparison base should cover all revenues and expenditure functions or be restricted to certain items. And within each of these revenue and expenditure categories, will the coverage be restricted still further to include only items (e.g., education, health) that are seen as “basic public sector responsibilities” as opposed to items such as cultural activities or state public enterprise activities that might be performed by the private sector? Finally, whatever functional expenditure coverage is chosen, there is the decision on whether the comparisons should be restricted to recurrent items (which can be more easily measured and compared) or extended to cover capital items also.

Fourth, with the coverage of the comparison base determined, it is then necessary to reduce the regional aggregates to a per capita basis by dividing by an accurate and up-to-date measure of regional population to allow standardized comparisons of regional fiscal capacities.

Fifth, once a sound per capita comparison base for actual regional revenues and expenditures has been established, attention must turn to investigating whether differences in these budget aggregates can be justified by underlying differences in relative revenue capacities and relative expenditure needs and costs.

In practice, the analysis of relative revenue capacities and expenditure needs can proceed in two ways. One approach is to identify the impact of location-specific disabilities on relative revenue and expenditure needs. On the taxation side, this may involve definition of a standard base for each revenue category and determination of a “standard national tax rate”—often calculated as an average by comparing actual revenue collected to a measure of the national tax base. On the expenditure side, it may involve identifying and collecting of statistical data of demand and cost factors affecting relative needs in each region. Examples of such factors include variations due to differences in demand arising out of the relative size, age, sex, or socioeconomic characteristics of relevant user population, cost differences arising out of the scale of delivery or administration in different regions, differences in degree of urbanization or industrial structure, and variations in the dispersion of the user populations.

Alternatively, an attempt can be made to identify policy differences between regions that could explain all or part of the differences in the actual budget figures. Once the impact of these policy factors is assessed, they can be deducted from the actual data to derive a cross sectional “needs” series that will reflect underlying revenue or cost disabilities. In concept, both approaches should yield similar results. Experience suggests that identification of location-specific disabilities may be a more fruitful course. Calculation from each approach, however, would provide a double check on what may sometimes have to be a broad judgment on the differences in regional fiscal capacities.

Of course, in some expenditure or revenue categories, it will be found that actual expenditures reflect actual need quite well—in which case, there may be no need to adjust the actual per capita regional budget figures in the needs calculation. In other budget categories, it may be found that needs are broadly similar on a per capita basis in each region—in which case, needs for each region can be assessed by including an equal per capita needs calculation for the items concerned in each region.

A sixth task involves decisions on the treatment to be accorded to any existing nonequalization grants to regions from the central government for specific purposes (e.g., education). Should these grants be continued once an equalization framework is established, and, if so, will they be taken into account in the overall calculations of revenue availability used in determining any equalization grant?

The last task relates to the choice of an “equalization model” to be used to formalize the measurement of horizontal balance. The formula for a general capacity equalization model is shown in the appendix, but a number of variations would be possible to take account of specific considerations. For example, it would be possible to take account of relative tax “efforts”—in addition to relative underlying revenue capacities—by each region and the specific allowance can be made for specific purpose nonequalizing grants. Both cases are discussed in the appendix.

Who Should Conduct the Equalization Studies?

No generalization is possible. In Australia, an independent Grants Commission has been appointed to advise on the distribution of grants. This system has worked successfully. The Commission has made its deliberation very open to scrutiny by the state and territory governments affected by its judgments, as well as the public generally. Public hearings are conducted and comprehensive reports published. The central government retains responsibility, however, for the final decisions on grant allocations. Normally, the Commission’s recommendations are accepted, but the government retains the right to vary them or to make other compensating adjustments in the total payments structure to the states and territories.

By contrast, the Canadian Ministry of Finance conducts the equalization grant calculations; however, there is again considerable emphasis on consistent methodologies and public scrutiny of the results obtained.

Who Should Pay the Grant?

One question that may need to be addressed is who should pay the equalization grant. In Canada and Australia, the “equalizing grant” is paid by the central government from the shares of total taxes available to it. This approach reflects a vertical balance favoring the national government to pay grants to lower levels of government. But the grants could equally be made from regions with above-average fiscal capacities to those with below-average fiscal capacities. This practice is followed in Germany where the vertical balance is small, leaving the central government with a limited role in equalization grants.

Timeframe of Equalization

It is important that equalization grants be calculated on a three-to-five-year average basis so that the effects of particular random events (e.g., a natural disaster, such as fire, flood, drought) or disparate economic trends (e.g., energy price fluctuations) are evened out. Experience in Canada and Australia suggests that regular updating of studies of regional capacities is essential, and therefore a process of rolling averages may need to be used.

The inevitable delays in designing and collecting the vast amount of data required for objective measurement of horizontal imbalances in fiscal capacities, however, usually means that any system of equalization grants must be based on historical data—possibly up to five years old. In other words, the implementation of equalization is usually a retrospective rather than forward-looking exercise. This could be a problem in a country such as China, which is undergoing rapid structural change.

Phasing-ln Equalization Arrangements

Given the large horizontal imbalances that already exist in fiscal capacities between regions in countries such as China, it may not be realistic to seek to obtain full equalization in the near term. Indeed, it seems probable that a goal of full equalization may make little sense given the distortions in public and resource allocation that are still being eliminated as China makes the transition to a market economy.

Over time, however, these concerns should diminish, and it may still be possible to set an intermediate goal such as assisting each region to achieve a fiscal capacity equal to, say, at least 70 percent of a visible national standard. That standard itself may not be a national average (which may be skewed by large rich provinces) but rather a subset of, say, larger eastern regions. Grants could be phased into achieve that goal, with poorer provinces achieving gradually greater shares of the total and richer provinces gradually reducing shares.


A Generalized Capacity Equalization Model

A formal equalization model is:


Gi = equalization grant to region i;

Pi = population of region i;

Rs/Ps = a national per capita revenue collection standard; qi = the differential revenue-raising capacity of region i;

qi = the differential revenue-raising capacity of region i;

Es/Ps = a national per capita expenditure standard; and

Vi = the differential cost of providing standard services in region i;

The per capita revenue term in equation (1) can be rewritten as


Rs/Ys = the national standard for regional government tax rates;

Ys/Ps = a national standard tax base; and

Yi/pi = the tax base of the region.

This term tells us that the revenue-raising capacity of a region is derived by applying the national standard regional tax rate to the difference between the tax base of the region and that of the nation as a whole. The calculation is normally carried out in per capita terms and may be applied to each of a number of individual taxes raised at the local government level and then summed to derive a total for each region.

The expenditure per capita term on the right-hand side of equation (1) can also be rewritten to explain relative expenditure needs for each region as a function of a number of independent factors, such as


ui = differential coverage of population eligible for services relative to the total populations’

si = differential costs arising out of scale factors;

di = differential costs arising out of concentration or dispersion factors; and

ei = differences in cost arising out of social, physical and economic factors.

Once a per capita measure is derived for each region, it is multiplied by the population of the region, Pit to derive the contribution of relative revenue-raising capacity or expenditure need calculation to the total grant payable for each region.

Separate calculations would be required for each own revenue source and expenditure, the effective weighing being determined by a standard budget that brings together the various revenues and expenditures judged to be relevant to the equalization exercise.

Including a Fiscal Effort Adjustment

Although the fiscal capacity models set out above do not impose performance conditions on recipients, it is possible to insert a fiscal effort adjustment factor to the grant entitlement calculation if that were judged to be appropriate. There will be three components of the grant instead of two in equation (6).

A variant of this model, which would have the objective of imposing a penalty for below-standard, revenue-raising effort but not of rewarding above-standard effort, would restrict the fiscal effort adjustment to negative amounts so that the equation above would be subject to the constraint that

Treatment of Nonequalization Grants and Tax-Sharing Arrangements

If the central government wishes to continue to make nonequalizing specific-purpose grants, the calculation of equalization grants would need to take account of the impact of these transfers on the initial preequalization position of each region. Equation (1) must be reformulated as follows:


OsPsσi= the differential specific-purpose grant.

This adjustment acts to nullify any relative advantage or disadvantages differential of existing nonequalizing specific purpose grants. Because the equalization calculations are undertaken with a lag, the adjustment is retrospective. Of course, if it were judged appropriate for regions to retain some or all of the advantages conferred by existing specific-purpose grants, it would also be possible to adjust the formula accordingly.


International Monetary Fund.


The “surplus” generated may flow from a number of factors. For example, the economies of scale from production and distribution within a national market place; common legal and other standards (working standards, education standards, product standards, and so on); common infrastructure (e.g., transport systems); and a predictable central fiscal structure and monetary system, including some harmonization of tax bases and rates for major taxes and common access to national public services (e.g., defense, foreign affairs).


See Robin W. Boadway and P.A. Hobson, Intergovernmental Fiscal Relations In Canada, Canadian Tax Paper, No.96 (Toronto: Canadian Tax Foundation, 1993).


Usually calculated as an average effective rate derived by comparing potential total tax collected with the total national tax base for the tax concerned.

    Other Resources Citing This Publication