70097 Technical note on the methodology for the allocation of intergovernmental grants in the Republic of Belarus Andrey Timofeev and Jorge Martinez-Vazquez* June 2010 The World Bank *International Studies Program, Andrew Young School of Policy Studies, Georgia State University 1 Table of Contents EXECUTIVE SUMMARY .................................................................................................................................. 3 I. INTRODUCTION ..................................................................................................................................... 5 II. BACKGROUND AND SITUATION ANALYSIS............................................................................................ 5 III. FISCAL DISPARITIES AND EQUALIZATION FRAMEWORK ....................................................................... 9 IV. ASSESSMENT OF THE INDIVIDUAL FISCAL CAPACITIES OF LOCAL GOVERNMENT UNITS.................... 15 V. ALTERNATIVE APPROACHES TO MEASURING EXPENDITURE NEEDS ................................................... 19 VI. SIMULATIONS ....................................................................................................................................... 24 VII. CONCLUSIONS ...................................................................................................................................... 31 VIII. ANNEXES .............................................................................................................................................. 33 2 EXECUTIVE SUMMARY In this technical note we evaluate the methodology proposed by the Ministry of Finance for the allocation of transfers to subnational governments and suggest a number of alternative options for various grant design elements Overall, the framework laid out in the Budget Code and the implementation approach developed in the draft methodology conform to sound principles in fiscal decentralization and the best international practices. However, a number of elements of the methodology still need to be elaborated and some elements might need to be revised. Immediately below we summarize our comments and suggestions that could be helpful for finalization of the draft methodology. 1. The drafted mathematical and narrative presentation of the proposed allocation mechanism can be made more simple and transparent. Overall, the current mathematical formula in the form of the ratio of the revenue index over the expenditure index is not as transparent and intuitive as the classical form allocating grants proportional to the gap between the revenue capacity and the expenditure needs. 2. Under the proposed formula, inequalities in revenue capacity are completely equalized while the extent of equalization of expenditure needs is determined by the size of the transfer pool. Rather than being an inadvertent outcome of the formula design, the sensitivity of the grant amount to the disparities in expenditure needs and fiscal capacity could be specified as formula parameters reflecting explicit policy choices. 3. Besides the estimated capacity for the revenue derived locally, the revenue capacity should also include the VAT revenue allocated per capita and possibly other grants if the latter are used to finance expenditures that are taken into account when determining per-client norms and included in the side of expenditure needs in the equalization formula 4. The Ministry of Finance could perform statistical estimation of the impacts of various local characteristics on the revenue yield from a unit of the revenue base. The resulting estimates could be provided to the oblast governments as part of the budget circular along with other budget parameters serving as inputs to their budget planning exercise. Performing estimations on a larger sample would ensure the accuracy of the estimates. Furthermore, in the calculations of adjustment coefficients for their cities and rayons, oblast level officials might be more comfortable with applying the provided values of elasticities to the percentage differences in various factors rather than performing the actual estimation of elasticies. In the data on local characteristics made available to the study team we have not found statistical evidence of any major impact on the revenue yield from a unit of the revenue base, that would warrant 3 introducing adjustment coefficients to the Representative Tax System. However, analysis of other regional characteristics, especially when performed on samples larger than just localities of one oblast, can uncover relationships that would necessitate adjustment coefficients. 5. Given considerations of practicality and transparency, it might make sense to group the proposed number of 30+ expenditure norms into a smaller number of expenditure categories based on the target clientele and fiscal importance. 6. To preserve objectivity of the grant allocation, the Ministry of Finance could perform statistical estimations of elasticities of per client costs with respect to various cost factors on the sample pooling together cities and rayons from all oblasts. The resulting estimates could be provided to the oblast governments as part of the budget circular along with other budget parameters serving as inputs to their budget planning exercise. 7. The Ministry of Finance could also develop and make available to subnational government a simple spreadsheet tool that would facilitate the application of the adopted methodology (or facilitate the evaluation of the proposed methodology at the stage of soliciting feedback from the stakeholders). 4 I. PREFACE This report was prepared at the request of the Ministry of Finance of the Republic of Belarus as part of follow up technical assistance after completion of the first Public Expenditure and Financial Accountability Assessment for Belarus in 2009. The report was prepared by Andrey Timofeev and Jorge Martinez-Vazquez (Consultant). The technical assistance was co-managed by Sebastian Eckardt and Marina Bakonova under overall guidance by Pablo Saavedra. Larysa Hrebianchuk provided logistical support to the team. The team would like to express its gratitude to government officials of the Belorussian Ministry of Finance for their constructive collaboration. The team in particular would like to thank Maksim L. Ermolovich and Shabalina TAtiana Nikolaevna. II. INTRODUCTION Recently Belarus initiated a new wave of fiscal reform efforts in part spurred by the impact of the global financial crisis on its economy. The two important elements of the legislative framework for these reforms are the Budget Code, adopted in 2008, and the Tax Code, finalized in 2009. Among other things, these initiatives aim to reform the system of intergovernmental fiscal relations in the country. As subnational governments account for almost half of the public sector (excluding the social security fund), the success of reforming the intergovernmental fiscal relations will be an important element of raising the efficiency of the public sector. As part of the implementation of the new principles of intergovernmental fiscal relations introduced in the Budget Code, the Budget Policy Department of the Ministry of Finance drafted a Regulation on the Allocation of Transfers and Assessment of Fiscal Sufficiency. This methodology covers the allocation of grants from the national government to oblasts, from oblasts to the budgets of cities and rayons, and from rayon governments to sub-rayon budgets. In this technical note we evaluate the proposed methodology of transfer allocation along with a number of alternative options. As a starting point for the intergovernmental grant reform study, Section II outlines the salient features of Belarus' public sector and reviews the latest developments in the broader area of fiscal policy reform. Section III evaluates the proposed methodology of transfer allocation along with a number of alternative options. Sections IV and V discuss in more details the approaches to measuring fiscal capacity and expenditure needs respectively. Section VI provides simulations and discusses implications of various options. We conclude with a summary of our findings. III. BACKGROUND AND SITUATION ANALYSIS Intergovernmental fiscal transfers are not a stand-alone issue, and cannot be considered void of the context of the overall fiscal decentralization framework in Belarus. Any viable reform proposal will have to take into account the broader fiscal policy approach of the government and the fiscal and political environment in which the reforms would have to take place. Although a comprehensive assessment of the decentralization context in Belarus falls beyond the scope of this study, the design of any intergovernmental transfer system should be informed and take place in the context of a broader understanding of Belarus’s unique fiscal decentralization context. To provide such context, in this section we summarize the relevant findings of the 5 Policy Notes on Selected Issues in Public Finance produced by the World Bank in 20071 and briefly describe subsequent developments. This overview is structured along the main pillars of the decentralization system and provides an assessment of the main challenges and issues the Government of Belarus faces in each of these areas. The main areas covered include: vertical structure and scope of the government sector, expenditure assignments, revenue assignment, intergovernmental transfers, and borrowing. Vertical structure of government Belarus’ structure of government is comprised of four tiers:2 o Republican (national) government o 7 regional governments (6 oblasts and the City of Minsk) o 130 base level jurisdictions (118 rayons and 12 towns of oblasts' subordination) o 1,426 primary level jurisdictions (1,348 rural districts, 64 rural settlements and 14 towns of rayon's subordination) In 2009, subnational budgets accounted for 47.7% of general government expenditures excluding social security funds (down from 51.5% in 2005) and 34.2% of general government revenues excluding social security funds (down from 40.7% in 2005). Budgets of all levels of governments are executed by the treasury department of the national Ministry of Finance. However, only for the national government agencies are cash balances consolidated in the Single Treasury Account (STA) at the National Bank of the Republic of Belarus (NBRB). For subnational governments, cash balances are held at local branches of (state-owned commercial) Belarusbank and Agrobank. All transactions with these local budget accounts are subject to the same treasury clearing procedures and the MoF Treasury Department has full access to information about these accounts. Subnational budgets are drafted by the (centrally appointed) executives but adopted by locally- elected councils. Expenditure assignment Articles 44-47 of the new Budget Code (adopted on July 16, 2008) provide somewhat specific assignments of expenditure responsibilities among the levels of government. However these assignments are prescribed not in terms of functions but financial responsibility. Thus, even though the Budget Code provides some clarity about who is responsible for financing, murkiness 1 Taking Stock of Intergovernmental Fiscal Relations: Issues, and Challenges, 2007, World Bank. 2 Improving transparency, integrity, and accountability in water supply and sanitation : action, learning, experiences: Belarus - Public expenditure and financial accountability (PEFA): public financial management assessment. 2009. World bank Report No. 48239-BY. 6 remains on what level of government is responsible for regulating and delivering these services. For many functional categories, the assignments overlap either due to identical wording for different levels of government or tautological language stating that responsibility for funding public entities lies with the level of government that owns these entities. Thus, the Budget Code states that local governments are responsible for maintaining institutions of education, healthcare, social services and others that are owned by or put under authority of local governments. Thus, the delineation of responsibility is centered around facilities rather than functions provided with these facilities. As a result, the division of responsibilities for the core functions between the levels of government is determined with the distribution of social assets between the levels of government. Nevertheless, in a number of functional categories the Budget Code establishes exclusive responsibilities, mostly for the central government, in particular in the area of defense, law enforcement and national economy. For the oblast level there are only a few exclusive assignments: rescue diving, territorial defense, management of land resources, and inter-city transport. For the city/rayon level the exclusive assignments are: city transit, subsidization of fuel wood and coal, subsidization of utilities, public housing and saunas, ambulances, day centers and foster families. For the settlement (primary) level governments, most expenditure responsibilities overlap with those of the city/rayon level: maintenance of public areas, water supply, public saunas, and minor activities in roads, housing, culture, and environment protection. Revenue assignment The current national legislation does not allow any level of government to introduce taxes beyond those enumerated in the Tax Code (Part I adopted on December 19, 2002; amended December 29, 2009; Part II adopted on December 30, 2009). The list of permitted taxes is broken into two categories: “republican� revenue sources and “local� revenue sources. Classification of a tax into a particular category does not always determine the level of government that receives the proceeds from this tax. In fact, as can be seen from Figure 1, the most important sources of subnational revenue (PIT, VAT, and Profit Tax) are national taxes shared with oblast governments, who in turn share these revenues with constituent localities. VAT revenue is shared proportional to population while all other taxes are shared on the basis of the derivation principle. The Budget Code sets minimum rates of sharing in the national tax revenues both for oblast and local governments. These minimum sharing rates vary by tax and also by type of government (city versus rayon). The rates of the derivation-based sharing of tax revenues with oblast governments (except Minsk City) have been uniform across oblasts since 2006. 7 Figure 1. Composition of subnational revenue Source: Calculated by the authors based on Ministry of Finance data. Note: Apart from the sales tax (abolished in 2010) and the fee for development of territories (reported in earmarked charges), all other local taxes are reported in the “other� category. Since the abolishment of a number of subnational taxes, including the retail sales tax, the list of local taxes contain only five items (see Table 1). Even before that, local taxes contributed less than 10 percent of subnational government revenues (see Figure 1) Table 1. Local taxes Tax name Who can Who defines the base Who sets the rate Who receives introduce the revenues Tax on services Oblast National legislation (gross Oblast government Local government revenue) within national limit governments (5%) Dog tag tax Oblast National legislation (number National government Local government of dogs over 3 month old) sets adjustment for dog governments size (x0.5-1.5) Fee for Oblast National legislation (total Oblast government Local development of government profits for enterprises, within national limit governments territories income for individual (3%) entrepreneurs) Resort fees City/rayon National legislation (resort City/rayon governments Local governments treatment price) within national limit governments (3%) Fee for flora Oblast National legislation (harvest Oblast government Primary level harvesting government value based on the gate within national limit government price) (5%) Source: Tax Code (Sections I and VI) 8 Transfers The Budget Code (Chapter 12) envisions subsidies, subventions, and other forms of intergovernmental transfers. The allocation of subsidies is to be determined based on the gap between the estimated expenditures and revenues of the recipient government (Art 75). The estimation of revenues is to be based on the revenue capacity while the estimation of expenditures is to be driven by the norms of "fiscal sufficiency" and adjustment coefficients. The norms of "fiscal sufficiency" are to be applied to the number of local residents or the number of beneficiaries of public goods in a locality. Adjustment coefficients are to capture differences in the costs of providing public services arising from differences in population size, socio- economic, demographic, climatic, environmental, and other characteristics of a locality (Art. 76). Subventions are to be provided for specific functional categories of expenditures and should be refunded to the higher-level government in case of less than full utilization (Art 77). As can be seen from Figure 1, intergovernmental transfers, excluding redistribution of the VAT revenue, accounted for about 27% of subnational revenues in 2009. The bulk of the national grants (over 85 percent) was accounted for by subsidies (general purpose transfers). The rest were subventions earmarked for the mitigation of the effects from the Chernobyl accident, capital infrastructure, housing vouchers, and so on. Borrowing The Budget Code allows for three forms of subnational debt (Art. 65): intergovernmental loans, issuance of securities, and extension of credit guarantees to other parties. In any given year the outstanding volume of subnational debt cannot exceed 30 percent of pre-transfer revenue of the jurisdiction. In case of exceeding this limit, local governments can only attract intergovernmental loans to cover cash shortfalls during the fiscal year. Credit guarantees are to be extended for a fee. IV. FISCAL DISPARITIES AND EQUALIZATION FRAMEWORK The concept of fiscal disparities provides a useful framework to design and analyze a system of grants. Fiscal disparity can be defined, for any government unit, as the excess of its expenditure needs and/or its fiscal capacity relative to some benchmarks. For example, in 2005-2007 per capita revenues of the six oblasts before grants on average deviated by 30 percent from the mean, that is the coefficient of variation was around 0.3.3 At the same time, per capita expenditures of these six oblasts on average deviated by 11 percent from the mean, that is the coefficient of variation was around 0.11. However, as we explain below, rather actual revenues and expenditures, a sound system of grants should be based on the objective notions of expenditures needs and fiscal capacity. 3 Taking Stock of Intergovernmental Fiscal Relations: Issues, and Challenges, 2007, World Bank. 9 Traditionally, expenditure needs represent funding necessary to cover all expenditure responsibilities assigned to the government at a standard level of service provision. Fiscal capacity can be broadly defined as the ability of a government to raise revenues from available revenue sources, exerting a standard level of fiscal effort. In general, local governments with larger disparities on the revenue and/or expenditure sides require a larger amount of transfers in order to discharge their competencies at some standard level. The differences in fiscal disparities among units of governments are called fiscal imbalances, and represent the unequal conditions under which the government units are discharging their competences. Once the problem of fiscal imbalances has been identified, the next step is to measure the size of the fiscal imbalances. These measures should, theoretically, give a very straightforward indication of the direction and amount of the intergovernmental transfers that are necessary to balance the fiscal disparities across all government units. More specifically, the transfer program should be (co-) financed by government units with negative fiscal disparities,4 and should benefit those local governments with positive fiscal disparities in proportion to the size of the disparities. In practice, however, measuring fiscal imbalances is an extremely difficult and challenging matter, and most of the problems related with the design and performance of the existing transfer systems deal in one way or another with the measurement of fiscal disparities or the actual size of the fiscal imbalances. This is because fiscal capacity and expenditure needs are notional values that can rarely be estimated accurately. In a sense, both concepts retain a great deal of subjectivity, because the decisions about what can be considered a “standard level of service provision� (and the associated expenditure need) or “a standard level of fiscal effort� (to measure fiscal capacity) are subject to debate. Therefore, there is a great need for reaching consensus through a policy debate involving relevant stakeholders regarding acceptable ways of measuring fiscal capacity and expenditure needs that would be perceived as fair. In addition, usually there are restrictions imposed by data availability that limit the use of statistical techniques for estimating expenditure needs and fiscal capacity. As a consequence, often the measures of fiscal disparities actually used can be quite imperfect, and sometimes they become a source of perverse incentives for local governments. Thus, special care must be taken in the designing of a transfer program in order to achieve adequate treatment of fiscal disparities with the available information, and to avoid inducing undesirable behavior of local governments. In the following two sections we will discuss these and other dimensions of the design of intergovernmental transfer programs aiming to remedy the aforementioned fiscal disparities. Generally, intergovernmental equalization transfers are intended to ensure that all local governments have the means to provide a comparable level of local public services while imposing an average level of local taxes. This is because it could be considered unfair if 4 Conceptually, horizontal fiscal imbalances could also be reduced by transferring resources from those jurisdictions with negative fiscal disparities to those with positive fiscal disparities. This is sometimes called a “Robin Hood system,� practiced in such a diverse group of countries as Denmark, Germany, Latvia, Lithuania, Switzerland, and Ukraine among others. In theory, it would represent a good solution to the problem of horizontal imbalances, and it would also allow us to face the vertical imbalances between the central government and the aggregated local governments separately in isolation. However, this system is difficult to implement in practice, because it can easily become unpopular and could also generate undesirable incentives to both the grantor and recipient governments. 10 residents in two localities face the same rates of, say property, taxes but cannot enjoy the same level of services because of the differences in the value of taxable property per capita across the two municipalities. It is widely accepted that equalization transfers should consist of unconditional lump-sum payments to local governments or general purpose funding, such that the recipient government by itself can independently decide how to use the additional revenues considering the particular preferences of its constituents Even though there exists, implicitly or explicitly, a widespread agreement about defining the operational objective of equalization transfer programs as the reduction of the differences among fiscal disparities of local governments, there are some cases where the concept of fiscal disparity is applied only partially. Thus, there are some countries (e.g. Macedonia) that equalize exclusively according to expenditure needs, while some other countries, such as Canada, equalize only fiscal capacity. The loss of excluding the measures of fiscal capacity (or expenditure needs for that matter) from the computation of the transfers needed by each jurisdiction would be negligible only in the rather implausible case where the excluded variable can be assumed constant across all local governments.5 Therefore, based on the previous discussion, it can be said that the consideration of both concepts, expenditures needs and fiscal capacity, is superior to the use of only one of them. However, some countries, such as Sweden, equalize disparities in fiscal capacity to a lesser extent than disparities in expenditure needs, thus providing some incentives for municipalities to develop their revenue capacities. A common goal of many transfer systems is to reduce the differences in fiscal disparities (or fiscal gaps) across jurisdictions. Those localities with negative fiscal disparities do not require, in principle, funds from the equalization program. At the same time, those localities with larger (positive) fiscal disparities should receive greater per capita transfers than others with smaller fiscal disparities. These are widely accepted principles. However, how big a per capita fiscal disparity should be in order to define a locality as beneficiary and how much more equalization transfers should be given to a relatively “needy� jurisdiction are open questions which cannot receive definitive answers. The most common approach is apportioning the available transfer funds among local governments as a fixed proportion of their (positive) fiscal disparities. All local governments with positive fiscal disparities will receive a transfer, and the size of the transfer will depend on the magnitude of the disparity relative to some benchmark. 6 In order to complete the discussion about the structure of the equalization transfer program, two qualifications need to be made. First, since the available funds are generally scarce, maximizing the equalization power of the transfer program will require the exclusion of some local governments from its benefits. This is especially true in the case of perfect vertical balance, where the yield of tax revenues assigned to the local level generates funds sufficient to cover 5 Another possibility would be the case where the excluded variable is perfectly correlated (negatively) with the one that is being considered. Again, this would not be a reasonable assumption. 6 An alternative apportionment mechanism consists in assigning transfers to localities with highest fiscal disparities first, and then to continue transferring funds to those local governments with the highest fiscal disparity in order to progressively reduce the maximum fiscal disparity up to the point where the transfer fund is exhausted. 11 expenditure needs at least of the wealthiest municipalities.7 Given the estimated fiscal disparities for each local government, the natural criterion is to concentrate the benefits of the grants exclusively in those jurisdictions with a positive fiscal disparities (the “needy� jurisdictions), and further to distribute the transfer fund to each of them in proportion to the size of its disparity or similar apportionment criteria. Defining the equalization transfer fund as X,8 the amount of transfers to be given to jurisdiction i, Ti, can be computed as the relative size of the local fiscal disparity with respect to the sum of all relevant fiscal disparities nationwide: X Ti   FDi* ,  FDi* where FDi* (with an asterisk) corresponds to those fiscal disparities with a positive sign. This is so because the jurisdictions with negative fiscal disparities should not, in principle, be beneficiated by the transfer program. A similar formula can be used to calculate contributions to the equalization fund from localities that have negative fiscal disparities if the system allows negative transfers. The classical formula for fiscal disparities measures the difference between expenditure needs and revenue capacity; or arithmetically: FDi = ENi – RCi, where FDi represents the fiscal disparity of local government i, ENi stands for its expenditure needs and FCi stands for its frevenue capacity. Alternatively, some countries define fiscal disparities by combining with certain weights a shortfall in revenue capacity FC–RCi, and an excess of expenditure ENi –EN over some benchmarks FC and EN, such as national averages. Arithmetically, such generalized definition of a fiscal gap can be expressed in per capita terms as: fgi = α• [eni –en] + β• [en–rc] + γ• [rc, – rci ] where fgi represents the fiscal gap of local government i, eni –en stands for its per capita expenditure disparity relative to the benchmark en and rc, – rci stands for its per capita revenue shortfall relative to the benchmark rc. The gap between the per capita expenditure benchmark en and the per capita revenue benchmark rc is called the vertical imbalance, capturing the shortfall in the yield of tax handles given to all local governments together relative to their expenditure needs. By contrast, the other two parts of the fiscal gap, that is eni –en and rci, – rc, are referred to as horizontal fiscal imbalances. In summary, fiscal disparities arise because the expenditure needs associated with the assigned expenditures responsibilities typically do not match the capacity of the governments to collect 7 For more discussion see Richard M. Bird, 1993. "Threading the Fiscal Labyrinth: Some Issues in Fiscal Decentralization." National Tax Journal, 46:2, pp. 207-27. 8 The size of the total equalization fund is a political decision informed by the overall resource envelope. Usually, the higher are initial disparities in local conditions and the lower is the inequality in fiscal outcomes that the society is willing to tolerate, the more resources would need to be allocated for that purpose. 12 revenues from own sources. We have already explained how the estimated size of the fiscal disparity of a local government can be used to define both its eligibility as a beneficiary of the program and the amount of transfers that it would ultimately receive. For these reasons, the effectiveness of the equalization transfer program depends crucially on the quality of the estimations of expenditure needs and fiscal capacity. In the following discussion we provide a set of methodologies that could be used in to estimate both variables, as well as a critical analysis that stresses their pros and cons. Framework proposed by the Ministry of Finance In the draft regulation, grant allocation mechanism is presented using the following mathematical formula: Tei = (RC/ P) • (К – RIi/EIi) • EIi • Pi , (1) where Tei – the amount of the equalization transfer to region i; RC – aggregate revenue capacity of all constituent regions in the higher level region for the coming fiscal year; P and Pi – the number of resident population in the higher-level region and its constituent region i correspondingly; К – regions’ revenue sufficiency level set as a benchmark for equalization of fiscal capacity by the Government of the Republic of Belarus (Council of deputies of an administrative and territorial unit of the respective higher level region) for the coming fiscal year; RIi – revenue capacity index for region i; EIi – the index of budget expenditures of region i; After substituting into the transfer formula the proposed expression for the index of revenue capacity, it becomes clear that for every region the proposed allocation mechanism brings the per capita level of potential revenues to the same level equal to a fraction К of the region-average revenue capacity adjusted for expenditure needs (see Annex I): Ti = К• (RC/ P) • ( EIi • Pi) – RCi (2) where Ti = Tei – Tui is the net amount of transfer, which can be positive or negative. Thus the net transfer Ti =Tie –Tub closes the gap between the revenue capacity of a given region and the equalization benchmark defined as the fraction К of the national average of the revenue capacity, adjusted for population and relative expenditure needs. 13 Evaluation and alternative options The drafted mathematical and narrative presentation of the proposed allocation mechanism can be made more simple and transparent. In particular, from the current formula it is not immediately clear that the transfer mechanism allocates more resources to regions with larger expenditure needs and smaller revenue capacity. Overall, the current mathematical formula in the form of the ratio of the revenue index over the expenditure index is not as transparent and intuitive as the classical form proportional to the gap between the revenue capacity and the expenditure needs: Ti = λ• ( ENi – RCi) The definition of the fiscal gap implied by the proposed mechanism (2) determines the sensitivity of the transfer amount to the differences in expenditure needs and revenue capacities among recipient jurisdictions. Let’s rewrite formula (2) in an equivalent form by substituting (ENi/ Pi )/ (EN/ P) for EIi (see Annex I): Ti = К• (RC/ EN )• ENi – RCi (3) From (3) it is clear that the equalization parameter K only affects the sensitivity of the grant amount with respect to expenditure needs, while disparities in revenue capacities translate one- to-one into differences in the grant amount. To further illustrate this point, let us consider two regions 1 and 2. Then the proposed allocation mechanism would allocate the following amounts of grants to these two regions: T1 = К• (RC/ EN) • EN1 – RC1 T2 = К• (RC/ EN) • EN2 – RC2 So that the difference in the grant amount between the two regions would reflect the differences in their respective revenue capacities and expenditure needs as following: [T2 -T1]= К• (RC/ EN) • [EN2 –EN1] – [RC2 –RC1]. Thus under the proposed mechanism, differences in revenue capacity translate one-to-one into differences in the grant amount while equalization of differences in expenditure needs depends on the value of parameter K, which in turn depends on the size the grant pool.9 9 The difference between the proposed approach and the classical formula is further illustrated by means of simulations performed for Mogilev oblast in Section VI. 14 By contrast under the standard definition of the fiscal gap, [T2 -T1]= λ• [EN2 –EN1] – λ• [RC2 –RC1]. Thus, under the classical approach, transfers equally treat differences in revenue capacity and expenditure needs. It has to be pointed out that some countries treat disparities in revenue capacity differently than disparities in expenditure needs. For example, in Sweden differences in expenditure needs are fully equalized while differences in revenue capacity are narrowed by 95 percent: Ti /Pi = [ENi /Pi –EN /P] +0.95• [RCi /Pi – RC /P)]. However, those countries do it by choice rather than as an inadvertent result of the formula design. Thus, less than complete equalization of revenue capacity in Sweden is expected to provide localities with incentives to grow their revenue base. Complete equalization of potential revenue can create the usual disincentives for economic development. To address this, in Russia only the poorest regions are brought to the same floor level of potential revenue while for all other regions the gap relative to the benchmark revenue capacity is narrowed by a certain percentage, which is uniform for all recipients. Other options for Belarus to consider: a. Bringing to the fixed level of potential revenues only those regions below some threshold; b. Narrowing the gap by a certain percentage, at least for those regions above certain minimum level of revenue capacity. This effectively means closing the same percentage of the revenue gap and the expenditure gap c. Explicit policy choice of the three separate equalization parameters: the extent of equalization of the revenue gap rc, – rci , the extent of equalization of the expenditure gap eni – en, and the extent of closing the vertical gap en–fc. V. ASSESSMENT OF THE INDIVIDUAL FISCAL CAPACITIES OF LOCAL GOVERNMENT UNITS Approaches to measuring the revenue capacity Practical challenges commonly arise in estimation of fiscal capacity, which in the case of local governments may be defined as the potential revenues that can be obtained from the tax bases assigned to the local government if an average level of effort (by national standards) is applied to those tax bases. Ideally, tax capacity should be measured by the size of the tax bases available to local governments, or the revenue that these tax bases would yield under standard tax rates. Using the actual amount of revenue collections in a locality as a measure of fiscal capacity should be avoided if local authorities can control tax rates, tax bases, or the administrative enforcement effort for it can create perverse incentives. Using actual collections, even from the past, creates negative incentives, because sooner or later local governments will “learn� that higher collections translate into lower transfers. A better approach is using some objective and 15 widely available indicator as a proxy measure for revenue capacity. The examples of such proxy measures include the per capita level of personal income or the local equivalent of the national- level gross domestic product, which can be called gross regional product (GRP). The basic idea underlying the proper estimation of fiscal capacity is to calculate the amount of revenue that a locality would collect given the level of income or economic activity in its territory if it were to exert average fiscal effort. Some countries (for example, Canada, United States, and Australia) have used a multidimensional measure of fiscal capacity known as the Representative Revenue System (RRS). This is done by collecting data on revenue collections and tax bases for each of the taxes under consideration for every locality. Using information on all tax bases for every jurisdiction as well as the national/regional average fiscal effort for each of the taxes, one can compute the amount of revenues that each jurisdiction would collect under the average fiscal effort. This amount is then considered to quantify the fiscal capacity of each jurisdiction. The main benefit of the RRS is that computations are made at a disaggregated level and based on detailed knowledge of (proxies for) the statutory tax bases. Fiscal capacity has been defined above as the potential revenue that a local government can raise from its tax base, exerting an average level of effort. Thus, in order to measure fiscal capacity, it would be natural to focus on those revenues sources over which local governments have a certain degree of control (i.e. the capacity to modify either the base, the rates applied, or the enforcement rigor). These are usually referred to as own revenues. Other revenues, such as per-capita shared collections of national taxes and other intergovernmental grants, of course, provide for local governments, but remain outside our focus because they cannot be directly affected by local governments and can be accounted for by the amounts directly received by local governments. The problem of estimating fiscal capacity is therefore reduced to the adequate estimation of (properly defined) locally-generated (own) revenues. Furthermore, since equalization transfers are not earmarked for a specific sector and therefore are meant to assist local governments to finance expenditure responsibilities in all sectors, for equalization purposes we can define fiscal capacity as the sum of estimated potential own revenues (EORi), shared revenues (Si, VAT in the case of Belarus), and all transfers received other than equalization transfers (hereinafter OTi). The fiscal capacity of a locality i can then be computed as: FCi = EORi + Si + OTi. Regardless of the methodology used to estimate potential locally-generated revenues, the overall fiscal capacity is obtained, as shown in the formula, by adding up the estimate of own revenues to the actual shared revenue retention and all transfers (except for those received for equalization purposes). Methodology for measuring revenue capacity proposed by the Ministry of Finance The proposed methodology utilizes the Representative Tax System (RTS) to assess potential revenue for the main six revenues sources (see Annex II for details). For all other revenues, the capacity is estimated jointly using the RTS estimate as a proxy. In other words it is assumed that 16 the capacity for all other revenue sources is proportional to the potential revenue estimated for the six main revenue sources. In the RTS, the estimate for tax j in locality i is computed by apportioning the forecast of the aggregate revenue collections from all localities in proportion to the share of a particular locality in the aggregate value of the indicator used as a proxy for the tax base. Furthermore, the share in the aggregated value of the tax base is calculated as a weighted average over the previous three years. Arithmetically, the computation can be represented with the following formula: RCij = rij*RFj * [0.25*St ij + 0.3*St-1 ij +0.45*St-2 ij] where RFj – the forecast of the aggregate revenue collections for revenue source j from all constituent regions in the higher-level region for the coming fiscal year; rij – revenue retention rate for tax j in region i; St ij = cij· Bt ij / Σi cij· Bt ij is the relative share of region i in the aggregate value of the revenue base proxy Bt ij totaled for all regions combined, where cij is an adjustment coefficient. Alternative options While overall the methodology conforms to the sound principles and best practice in estimating revenue capacity, there are a number minor modifications that can further improve the methodology: 1. The proposed measure of revenue capacity can be formulated in simpler terms, for example by dropping the past revenue collections from the calculations as they cancel out after appearing in both the numerator and denominator of the revenue capacity formula. 2. For revenue sources other than the core set estimated by means of the representative tax system (RTS), the revenue potential is assumed to be proportional to the RTS estimates. However, the RTS takes into account the retention rate for shared taxes. By contrast, other revenue sources are 100% retained at the local level and thus should not depend on the retention rates. To fix this, the RTS can be estimated for the potential revenue for the consolidated oblast- local budget, which can be used as a proxy for other local revenue sources. Then the consolidated RTS estimates can be modified by the retention rates to arrive at the potential local revenues. To take a careful account of tax revenue retention rates varying across localities, the estimation of revenue capacity should be made for the consolidated budget first and then local retention rates applied to these consolidated estimated. Thus the revenue capacity formula for revenue source j in locality i should look like this: 17 RCij= rij• RCCij, where RCCij is consolidated revenue capacity including both the share retained by a lower-level government and the share remitted to the higher-level government from revenue source j, estimated using the following formula: Then RCCij = RFj • [0.25•St ij + 0.3•St-1 ij +0.45•St-2 ij]. 3. Besides the estimated potential revenue derived locally, the revenue capacity should also include the VAT revenue allocated per capita and possibly other grants if the latter are used to finances expenditures that are included in the calculation of per-client norms. 4. In the proposed methodology, for each region and each revenue source, the revenue base proxy is adjusted by a “coefficient based on social and economic peculiarities of region i directly affecting tax (payment) amount j in the reported year t.� The draft regulation does not specify how these adjustment coefficients will be calculated. One way to preserve objectivity of the allocation mechanism is through statistical estimation of impacts of various regional characteristics (e.g., industrial composition as in Russia’s formula) on the revenue yield from a unit of the capacity proxy. Then, different regions will have different adjustment coefficients only to the extent that they have differences in their objectively measured characteristics. To be able to utilize the convenient notion of elasticity for the computation of adjustment coefficients, the formula for computing revenue capacity would have to be slightly modified. In the modified formula, the adjustment coefficients are applied to the relative share of the locality in the aggregate value of the revenue base: RCCij = RFj • [0.25• ct ij· St ij + 0.3• ct-1 ij· St-1 ij +0.45• ct-2 ij· St-2 ij] Then we can utilize the notion of elasticities for the computation of the adjustment coefficients in the following form: Ci= 1–a1 · (x1i – X1) / X1 + a2 · (x2i – X2) / X2)+ …+ aK · (xKi– XK)/ XK, where: o ak signifies the relative weight of adjustment factor k; and o xi / X represents the value of each factor that is recorded in locality i relative to the value of the same factor for all regions combined. As an illustration, let’s consider an example with just one factor: 18 RCCij = RFj • [0.25•(1–a · (xti – Xt) / Xt)· St ij + 0.3•(1–a1 · (xt-1i – Xt-1) / Xt-1)· St-1 ij +0.45•(1–a· (xt-2i – Xt-2) / Xt-2)· St-2 ij] (4) It can be interpreted as a weighted average of the three separate estimates for each of the three previous years respectively: RCC,tij = RFj • [(1–a · (xti – Xt) / Xt)· St ij]= RFj • [(1–a · (xti – Xt) / Xt)· Bt ij / Σi Bt ij]. In turn for the each of the three years, the formula can be rearranged as following: (RCC,tij / Bt ij) / (RFj / Σi Bt ij) = (1–a · (xti – Xt) / Xt) or [(RCC,tij / Bt ij ) -( RFj / Σi Bt ij ) ]/( RFj / Σi Bt ij ) = a · (xti – Xt) / Xt From this last expression it is clear that a translate percentage differences in the value of the given adjustment factor into percentage differences in the value of revenue capacity per unit of the proxy indicator (see Appendix III for more details on the estimation of elasticity). Furthermore, if we want the individual revenue capacities to add up to the aggregate revenue forecast for a particular tax, Xt should be computed as the weighted average with the weight for each locality proportional to the size of that locality revenue base Bt ij (see the Annex IV for derivations). VI. ALTERNATIVE APPROACHES TO MEASURING EXPENDITURE NEEDS The typical formula for expenditure needs of locality i, or Expi, on a per client basis adjusted for the relative cost can be written as10 Expi = ci * Norm* pi where Norm is the standard expenditure per client computed by the higher-level government for all constituent localities following some methodology, pi is the client population in locality i (as reported for example in the latest actualized census data) that benefits from the local service in question, and ci is the parameter measuring the higher (more than one) or lower (less than one) costs in the locality relative to the average cost of providing that service in the country. Determining a per client norm 10 For more details on international practice see Jun Ma (1997) “Intergovernmental Fiscal Transfer in Nine Countries: Lessons for Developing Countries,� World Bank Policy Research Working Paper 1822. 19 The Ministry of Finance is considering using explicit per-client financial norms adjusted for different costs of service delivery. There are two basic approaches to computing a per-client norm: bottom-up and top down. In its ideal form, the bottom-up approach departs from a specific service target in terms of service outcome and estimates per client amount of financial resources necessary to achieve the desired outcomes (see Box 1 for experience with the bottom-up approach to financing education in the USA). If experience of other countries is any indication of how daunting this task is, Russia’s Budget Code of 1998 envisioned a set of “minimum social standards� as the basis for budgeting, which still have not been developed after ten years having elapsed. Even when such service standards are put in place, translating them into financial amounts is a very demanding exercise. It is data intensive and requires technical calculations, with demands on the skills sets and time of public officials. For example, in the USA, rather than doing it in house most states commission to external consulting companies "costing-out studies" for funding education.11 The bottom-up approach has its downsides. The bottom-up approach can produce norms that are unrealistic or unaffordable from a budget viewpoint, so that the actual funding would need to be reduced to make it affordable, thus frustrating government officials and citizens. By contrast to the bottom up approach, the top-down approach ensures affordability because it starts from the feasible level of aggregate appropriations. When the top-down approach is used to develop affordable norms, the basic figures for the norms are not necessarily estimated from past expenditures. The data may be given by the budget authorities and derived from budget aggregates.12 For example, to arrive at the basic norm per student in primary education, the Ministry of Finance might decide while initiating budget preparation that education should represent a particular share of total aggregate local government expenditures, and that secondary education should represent a particular share of total education expenditures. From this fraction, the total funds to be allocated to the sector can be easily worked out. The resulting funds divided by the number of secondary school students provide the basic expenditure norm per student in secondary education. This basic norm can be adjusted up and down for cost differences, special needs, conditions, and so on. Box 1: Financing adequate education in the USA In the USA, 40 out of total 50 states finance school districts based on the minimum (“foundation�) per student norm. Sometimes the explicit norm is necessitated by court rulings concerning the promise of adequate education in the state constitution. Different states use different methodologies to determine the minimum norm, which can be generally classified into three broad categories: statistical analysis, expert opinion, and a successful school district. The statistical (“evidence based�) approach uses empirical evidence to link an educational program to desired outcomes and then costs (on per student basis) the identified programs to 11 Michael Griffith. A Survey of Finance Adequacy Studies, May 2007; National Access Network. Status of Education Cost Studies in the 50 States, May 2007. 12 The top-down approach to determining the expenditure norm exhibits more predictability and transparency when the budget aggregate is set on a multi-year basis in the framework of medium-term projections of aggregate service demand and costs. 20 deliver desired outcomes. It has been used only by a few states to validate estimates obtained by other methodologies. Under the expert opinion approach, a panel of experts creates a model school that they believe would achieve the desired outcomes and then costs out this model school (again on per student basis). The panel also determines weights that should be given to characteristics of a school district (special needs students, economies of scale, isolation) that would affect costs relative to the model institution. The successful school district methodology identifies schools that achieve the desired outcome (e.g., test scores) in the average environment (in terms of demographics and resource base) and then determines the average per-student costs from actual expenses in those schools. Regardless of the methodology, the study results are presented to the state legislatures, who set the new financial norm somewhere between the current amount and the one recommended by the study. If standards are ambition targets rather than the currently prevailing level, the costs cannot be immediately legislated into financial norms but rather serve as a medium-term target. Source: Augenblick, J.G., Myers, J.L., Anderson, A.B., 1997. Equity and adequacy in school funding. Future of Children 7, 63–78. As discussed above, the estimation of expenditure needs under the per client financial expenditure norm methodology requires the calculation of the number of clients for the relevant categories of expenditures. There are several choices that must be considered in this exercise. One choice is the range of expenditure functions over which to measure the needs. The computation of expenditures needs can be done separately by category of expenditures or for the composite of several local services. Under either approach, all public services should in principle be included in order to avoid bias against those local governments where an excluded service might be especially important. However, it would be impractical and even misleading to try to define a per client norm for every single expenditure program undertaken at the local level. A large number of expenditure standards would reduce transparency in the system and enhance the likelihood of complex discussions about the proper client bases. In general, it makes sense to group under “general public services� some functions that are unimportant in budgetary terms, and use the local population to estimate the number of clients for this composite category. The international experiences with the number of separate categories of expenditures having separate norms vary among countries. Some countries use one composite norm covering all local expenditures as in Switzerland, Denmark, and Macedonia. Other countries have several norms for separate groups of expenditures: six in Japan (police, public works, education, welfare and labor, industry and economy, public administration), seven in the United Kingdom (education, social services, highway maintenance, police, fire protection, capital expenditures, and the residual services), eleven in Australia (welfare, culture and recreation, community development, 21 general public services, services to industry, education, health, law and order, transport, economic affairs, trading enterprises).13 Methodology for expenditure needs proposed by the Ministry of Finance The proposed methodology utilizes a per client basis adjusted for the relative cost to assess expenditure needs for 31functional sub-categories Belarusian budget classification for the oblast level expenditures (29 for the city/rayon level and 3 for the sub-rayon level; see Annex V for details). Alternative options 1. Belarus’ budget classification identifies 10 functional categories and 43 sub-categories. However, given the stated above considerations of practicality and transparency, it makes sense to group them into a smaller number of categories based on the target clientele and fiscal importance. For the oblast-level expenditures, only the following four categories exceed ten percent of the total expenditures: 1) Social policy 2) Healthcare 3) Agriculture 4) Housing and utilities. For the sub-oblast levels, the significant expenditure categories are 1) Housing and utilities 2) Healthcare 3) Secondary education 4) Preschool education. 2. In the proposed methodology, the clientele for the public administration category is defined as the number of public servants. This can create negative incentives to hoard staff. Alternatively, the clientele for the public administration category can be defined as the total population. 3. For education services, the clientele is defined as the actual enrollment. This can penalize localities with more challenging education environment (more poor households, etc) as it is costlier to maintain enrollment. Alternatively, the clientele for education services category can be defined as the school-age population. 4. For each expenditure responsibility, the expenditure need is estimated as the clientele size times a per client norm. However, for each region and expenditure responsibility, the per client norm is adjusted by a “coefficient of expenditure j for region i.� The draft regulation does not specify how these adjustment coefficients will be calculated. One way to preserve objectivity of the allocation mechanism is through statistical estimation of impacts of various regional characteristics (e.g., land area to control for economies of scale) on the per client costs of providing public services. Cost adjustment coefficient for jurisdiction i can be generally computed as Ci= (1– a1 – a2–…– aK)+ a1 · (x1i / X1) + a2 · (x2i / X2) + …+ aK · (xKi/ XK), where: o ak signifies the relative weight of each cost factor; and o xi / X represents the value of each factor that is recorded in locality i relative to the value of the same factor for all localities combined. 13 Ma (1997). Idem. 22 Rearrangement of terms in the expression above yields an equivalent expression, which now clearly shows that the adjustment coefficient is larger than one whenever a locality has a higher value of the cost factors than the average across all localities: Ci= 1+a1 · (x1i – X1) / X1 + a2 · (x2i – X2) / X2)+ …+ aK · (xKi– XK)/ XK, In general, the choice of exact variables or cost factors to be used depends on the cost structure of the specific category of local functions at hand. Even though varying with specific local functions, the utilized factors should:14  Accurately reflect the targeted characteristics. The number of children with special needs captures better the costs for education than the actual enrollment in schools/classes for children with special needs.  Be regularly updated in the future (every year or every two years). Data originating from one-of-a-time study should not be used.  Come from an independent source respected by all stakeholders. A national bureau of statistics often has more credibility than a line ministry.  Be drawn from a source that cannot be manipulated by one or more local governments. Self-reporting of data by local governments cannot be trusted if local governments are well aware of the link between what they report and the amount of resources they receive in the future.  Reflect objective drivers of costs (for example, the number of clients) rather than the chosen mode of service provision. Population density is more objective in capturing cost drivers than the actual class size. Given that variety of functions assigned to local governments in different countries, there is a great variety of factors included in their formulae.15 These factors can be classified into two broad categories: 1. Indicators of service need: illiteracy rate, poverty rate, sickness rate, and infant mortality; 2. Cost indicators: price index, land area, average temperature, and mountainous areas. One of the most important practical issues is how to determine factor weights that relatively capture the contribution of each cost factor to the relative disparities in service costs across municipalities. Several approaches can be used to arrive at a particular set of factor weights in a more or less objective manner. A more scientific approach is to utilize local budget data to establish how the cost of delivering standard services varies across local jurisdictions, and, in particular, how these costs are responsive to variations in socio-economic characteristics of 14 Bahl, Roy, Jamie Boex and Jorge Martinez-Vazquez., 2001. The Design and Implementation of Intergovernmental Fiscal Transfers, Atlanta: Andrew Young School of Policy Studies, Georgia State University. 15 Boex, Jamie and Jorge Martinez-Vazquez. 2007. "Designing Intergovernmental Equalization Transfers with Imperfect Data: Concepts, Practices, and Lessons" In: Challenges in the Design of Fiscal Equalization and Intergovernmental Transfers. Jorge Martinez-Vazquez and Robert Searle (ed.), New York: Springer. 23 localities (see Annex III).16 Alternatively, the factor cost weights can emerge from consensus building consultations where all parties agree that it would be fair to allocate that much more resources per capita to municipalities with, for example, that much larger land area per capita. The third alternative to determining factor weight is through examining shares of various expense items in the aggregate sectoral expenditures. If, for example, 70 percent of costs are accounted for by personnel, which increases with population, while 30 percent accounted by transport, which increases with land area, then we might assign weight 0.7 to population and weight 0.3 to land area. VII. SIMULATIONS As a key analytical step in the development of a new system of intergovernmental grants, in this section we present a number of policy simulations. Such simulations show the amounts of transfer funds that local governments would receive under the proposed transfer mechanisms. The simulations intent to clearly reflect the impact of introducing the proposed transfer mechanism on local finances, lay down a framework for transparency, demonstrate the would-be winners and losers, and whether establishing a “phasing-in� process for the new transfer methodology would be necessary. Because at the time of writing this report only data on the cities and rayons of Mogilev oblast were made available to the study team, our simulations are limited to the allocation of oblast grants in cities and rayons of this one oblast. Based on the available data, the simulations are performed for the year 2009 based on the previous three years (2006, 2007, and 2008). 16 Because this kind of analyses is based on statistical averaging, it assumes that the identified patterns reflect objective impact of socio-economic characteristics on unit costs. 24 Figure 2. Simulation of grant allocation: baseline formula (MoF proposal) Source: Calculated by the authors based on Ministry of Finance data. 25 Local revenue capacity in Mogilev Oblast The simulations of the revenue capacity make only a few deviations from the draft methodology. First, the Representative Tax System (RTS) is estimated for the consolidated oblast-local budget. Then the consolidated RTS estimates are modified by the retention rates to arrive at the potential local revenues. Second, because we did not have data on proxies for earmarked targets, land tax, property tax, and the special tax regime revenues, we could only estimates the RTS for the PIT, Profit Tax, and Retail Sales Tax. Other taxes, accounting for 24 percent of the pre-transfer revenues, are assumed to be proportional to the RTS estimates (see Table 2). Finally, as explained above, the measure of fiscal capacity is completed by including those revenues received by the local government from VAT sharing. Table 2. The computation of revenue capacity for cities and rayons of Mogilev oblast (2009 values, in thou. BYR) Proxy for the tax Aggregate % of base forecast total Revenue category (actual revenue revenues 2009) VAT Actual receipts 455,580,675 35% Property tax Proportional to the 71,475,899 6% RTS Profit tax Profits 106,856,032 8% PIT Payroll 374,233,572 29% Retail Sales Tax Retail sales, including 50,666,202 4% catering Others Proportional to the 227,339,744 18% RTS Source: Authors’ own calculations As can be seen from the bottom bars on Figure 2, the per capita values of the revenues capacity range by a factor of 1.6 from 490 thousand roubles in Krasnopolskiy rayon to 807 thousand roubles in Krichevskiy rayon. However, the overall disparities in per capita revenue capacity are rather modest, on average within seven percent of the mean, that is the coefficient of variation equal to 0.07. As an alternative, we re-estimate the revenue capacity using adjustment coefficients. For each revenue source we attempted to estimate elasticities of the revenue yield with respect to various potential adjustment factors, but they turned out statistically significant only for the profit tax with respect to the rate of urbanization. It can be expected that from the same amount of profits, 26 in the rural areas there is less tax yield due to various tax preferences often provided to the agricultural sector and the fact this is one of the “hard to tax� sectors Annex VI provides a sample of the procedure followed to estimate adjustment coefficients for the profit tax. Unlike in the methodology proposed by the Ministry of Finance, here the factor proportions are calculated without applying the adjustment coefficients, which are instead applied to the computed factor proportions. Columns (5-7) report percentage differentials from the oblast weighted average of one adjustment factor (proportion of urban population) over the previous three years. The percentage differentials in urbanization rate are multiplied by the estimated elasticity for the profit tax revenue (0.043) to arrive at the resulting adjustments coefficients. These resulting coefficients are reported in columns (8-10) of appendix table VI. The magnitude of the adjustments coefficients ranges from 0.97 to 1.02, reflecting the weak elasticity of profit tax revenues with respect to the urbanization rate. This suggests that the additional complexity from adjustment calculations might not be warranted by the extent of produced adjustments. However, this simulation should serve as an illustration how adjustment coefficients can be calculated if some local characteristic necessitates significant adjustments to the calculation of revenue capacity. Estimation of expenditure needs for Mogilev oblast using per client expenditure norms Due to data availability, we had to make a few deviations from the draft methodology in the computation of expenditure needs. Instead of separate expenditure norms for 29 categories of expenditures, we group them into five categories based on the target clientele and fiscal importance (see Table 3). Table 3. The computation of expenditure norms (2009 values, in thou. BYR) Aggregate Clients Estimated Expenditure Expenditure expenditures needs number norm per category (total exps. 2008) of clients client (2008) Housing and utilities 358,703,963 Total population 1,126,382 318 Healthcare 339,335,023 Total population 1,126,382 301 Secondary education 345,643,204 Actual enrollment 127,600 2,709 Preschool education 119,957,092 Actual enrollment 40,500 2,962 Others 459,152,773 Total population 1,126,382 408 Source: Authors’ own calculations Note: 27 As was explained above, the expenditure need for each function and locality is computed by multiplying the per client expenditure norm by the number of clients in the locality. The results are finally added up for all functions in order to estimate the local expenditure need. As can be seen from the height of the stacked bars on Figure 2, the per capita values of the expenditure needs range by a factor of 1.16 from 1,352 thousand roubles in Mogilevskiy rayon to 1,569 thousand roubles in Kostyukovichskiy rayon. However, the overall disparities in per capita expenditure needs are rather modest, on average within four percent of the mean, that is the coefficient of variation equal to 0.04. Our assessment of data availability reveals a limited set of indicators reported by Belstat for the level of cities and rayons. The only relevant data that are available for the local level include demographic structure, unemployment, average wages, and social infrastructure capacity and actual enrollment in schools. For each of the five expenditure categories we attempted to estimate elasticities of the per-client costs with respect to various potential cost factors. The largest magnitude of elasticity (-4.35%) was found for the housing and utilities costs with respect to the size of population. The negative elasticity is consistent with the expectation of lower cost in larger localities due to economies of scale. Annex VII provides a sample of the procedure followed to estimate adjustment coefficients for housing and utilities costs. Column (3) reports percentage differentials from the oblast weighted average of one adjustment factor: population size. To arrive at the resulting adjustments coefficients, the percentage differentials in urbanization rate are multiplied by the estimated elasticity for the profit tax revenue with respect to urbanization (0.0435). These resulting coefficients are reported in column (4) of annex table VII. The magnitude of the adjustments coefficients ranges from 0.95 to 1.04, reflecting the rather weak elasticity of housing and utilities costs with respect to population size. The few other factors that turned our statistically significant had such a small magnitudes of elasticities (less than one percent) that they made meaningless calculation of adjustment coefficients. Other potential adjustment factors, such as population density, for which data were not available to the study team, could have a larger impact on the per-client costs. Furthermore, performing estimations elasticities of per client costs with respect to various cost factors on the sample pooling together cities and rayons from all oblasts could yield statistically significant estimates of elasticities due to increased accuracy. Simulation of grant allocation for Mogilev oblast Annex VIII provides a sample of the procedure proposed by the Ministry of Finance to allocate grants from the oblast government to cities and rayons. Table VIII-1 presents the calculations of the fiscal sufficiency ratio defined as the ratio of the revenue capacity index over the expenditure need index. Table VIII-2 presents the calculations of the grant amount necessary to bring the 28 fiscal sufficiency ratio to the maxim-level possible given the size of the transfer pool of 618,578,405 thousand roubles (the actual amount in 2009). The given amount of funds allows raising the average level of fiscal sufficiency by a factor of 1.82 (See Annex I for details). That is the equalization parameter K in the MoF formula has the value of 1.82 in this simulation. As can be seen from the middle bars on Figure 2, the per capita value of subsidies range by a factor of 2.32, from a value of 364 thousand roubles for Mogilevskiy rayon to 843 thousand for Krasnopolskiy rayon. As an alternative, we simulate an allocation of grants following the classical approach closing a fixed proportion of the gap between revenue capacity and expenditure needs. For these simulations we use the same baseline estimates of revenues and expenditures as reported in Annex VIII. The given amount of funds allows covering 71% of the gap between expenditure needs and revenue capacities of cities and rayons. The middle bars on Figure 3, display the per capita value of subsidies computed according to this alternative method. These per capita amounts range by a factor of 1.92 from 409 thousand roubles for Mogilevskiy rayon to 773 thousand roubles for Krasnopolskiy rayon. A comparison of figures 2 and 3 reveals that, compared to the MoF formula, the alternative approach tends to allocate more resources to the five wealthiest localities (Krichevskiy, Mogilevskiy, Kostyukovichskiy, Osipovichskiy rayons and Mogilev City) and less to each of the less affluent localities. This is because the alternative approach bridges only 71% of the disparities in revenue capacity while the MoF formula completely eliminates the differences in revenue capacity. The implications from this fact for grant design depend on the extent of disincentives that a complete equalization could pose for local government effort to increase the revenue base. The purpose of this simulation is to point out this fact so that an informed grant design choice could be made by policy-makers. 29 Figure 3. Simulation of grant allocation: alternative formula (proportional closing of the fiscal gap) Source: Calculated by the authors based on Ministry of Finance data. 30 VIII. CONCLUSIONS In this technical note we evaluated the methodology proposed by the Ministry of Finance for the allocation of transfers to subnational governments and suggested a number of alternative options for various grant design elements Overall, the framework laid out in the Budget Code and the implementation approach developed in the draft methodology conform to the sound principles and best practices. However, a number of elements of the methodology still need to be elaborated and some elements might need to be revised. Immediately below we summarize our comments and suggestions that could be helpful for finalization of the draft methodology. 1. The drafted mathematical and narrative presentation of the proposed allocation mechanism can be made more simple and transparent. Overall, the current mathematical formula in the form of the ratio of the revenue index over the expenditure index is not as transparent and intuitive as the classical form allocating grants proportional to the gap between the revenue capacity and the expenditure needs. 2. Under the proposed formula, inequalities in revenue capacity are completely equalized while the extent of equalization of expenditure needs is determined by the size of the transfer pool. Rather than being an inadvertent outcome of the formula design, the sensitivity of the grant amount to the disparities in expenditure needs and fiscal capacity could be specified as formula parameters reflecting explicit policy choices. 3. Besides the estimated capacity for the revenue derived locally, the revenue capacity should also include the VAT revenue allocated per capita and possibly other grants if the latter are used to finances expenditures that are taken into account when determining per-client norms. 4. The Ministry of Finance could perform statistical estimation of the impacts of various local characteristics on the revenue yield from a unit of the capacity proxy. The resulting estimates could be provided to the oblast governments as part of the budget circular along with other budget parameters serving as inputs to their budget planning exercise. Performing estimations on a larger sample would ensure the accuracy of the estimates. Furthermore, in the calculations of adjustment coefficients for their cities and rayons, oblast level officials might be more comfortable with applying the provided values of elasticities to the percentage differences in various factors rather than performing the actual estimation of elasticies. In the available to the study team data on local characteristics we have not found statistical evidence of any major impact on the revenue yield from a unit of the capacity proxy, that would warrant introducing adjustment coefficients to the Representative Tax System. However, 31 analysis of other regional characteristics, especially when performed on sample larger than just localities of one oblast, can uncover relationships that would necessitate adjustment coefficients. 5. Given considerations of practicality and transparency, it might makes sense to group the proposed number of 30+ expenditure norms into a smaller number of expenditure categories based on the target clientele and fiscal importance. 6. To preserve objectivity of the grant allocation, the Ministry of Finance could perform statistical estimations of elasticities of per client costs with respect to various cost factors on the sample pooling together cities and rayons from all oblasts. The resulting estimates could be provided to the oblast governments as part of the budget circular along with other budget parameters serving as inputs to their budget planning exercise. 7. The Ministry of Finance could also develop and make available to subnational government a simple spreadsheet tool that would facilitate the application of the adopted methodology (or facilitate the evaluation of the proposed methodology at the stage of soliciting feedback from the stakeholders). 32 IX. ANNEXES Annex I: Framework proposed by the Ministry of Finance The essence of the proposed allocation mechanism becomes clear when we substitute into the transfer formula the expression for the revenue capacity index (1): RIi = [(RCi – Tui) / Pi ]/ (RC/ P) (I1) where RC and RCi are revenue capacities of the higher-level region and its constituent region i correspondingly; Tui is an upward transfer from region i to the higher-level region. Substituting (I1) into (1) yields Tei = (RC/ P) • (К – {[(RCi – Tui) / Pi ]/ (RC/ P)}/EIi) • EIi • Pi =(RC/ P) • К• EIi • •Pi - (RCi – Tui) Rearranging the last expression gives (RCi – Tui +Tei )/( EIi • Pi ) = К• (RC/ P) (I2) From this latter expression it becomes clear that the proposed allocation mechanism brings the per capita level of potential revenues adjusted for expenditure needs to the same level equal to the fraction К of the region-average revenue capacity. To show the difference the proposed mechanism makes compared to the standard definition of the fiscal gap, ENi * – RCi, let’s rewrite equation (I2) in an equivalent form by substituting (ENi/ Pi )/ (EN/ P) for EIi and rearranging terms: Ti = Tei – Tui = К• (RC/ P)• ( EIi • Pi ) – RCi= =К• (RC/ P)• ( [(ENi/ Pi )/ (EN/ P)] • Pi ) – RCi= =К• (RC/ EN )• ENi – RCi (I3) From (I3) it is clear that in the proposed formula the availability of funds only affects the sensitivity of the grant amount with respect to expenditure needs, while disparities in revenue capacities translate one-to-one into the differences in the grant amount. To further illustrate this 33 point, let us consider a given pool of funds (Pool) and determine the extent of equalization under the proposed and the standard definitions of the fiscal gap. Under the proposed approach, Pool = Σi Ti = Σi К• (RC/ EN )• (ENi) – Σi RCi = К• (RC/ EN )• (Σi ENi) – Σi RCi = RC • К – RC. Therefore, К= (Pool +RC)/RC. Substituting the latter into (I3) yields, Ti = [(Pool +RC)/RC]• (RC/ EN )• (ENi) – RCi =[(Pool +RC)/EN]• ENi – RCi It is clear that a larger pool makes the transfer size more sensitive to differences in the expenditure needs but does not affect the sensitivity of the grant amount with respect to disparities in revenue capacity. The only case when this mechanism would treat revenue and expenditures equally is when (Pool +RC)/EN = 1, that is when Pool = EN– RC. By contrast under the standard definition of the fiscal gap, transfer Ti closes fraction λ of the fiscal gap EIi – RCi, so that Pool = Σi Ti = Σi λ (ENi – RCi)= λ (EN – RC). Therefore, λ = Pool /(EN – RC) and Ti = [Pool /(EN – RC)]• (ENi – RCi) It is clear that a larger pool makes the transfer size equally more sensitive to differences in both the expenditure needs and revenue capacity. 34 Annex II: Representative Tax System proposed by the Ministry of Finance for measuring revenue capacity Table II-1. A core set of taxes for calculation of the index of taxable capacity in total for an oblast Tax designation Indicator defining a tax base according to the data of Subsection the National Statistics Committee of the Republic of Section Belarus and (or) other sources PIT 01 00 Total gross wages, including small businesses Profit tax 02 00 Before-tax profit of organizations; Sales proceeds from goods, products, work and services of small businesses Fee for area development 04 00 After-tax profit (net profit); Net profit of small businesses Land tax 08 00 Taxable area of land including cadastral valuation (calculated based on average weighted number) Tax on real estate 09 00 Value of buildings and constructions of the organization; Value of buildings and constructions of individuals Other taxes and fees from goods 12 00 Sales proceeds from goods, products, work, (work, services), sales proceeds – services, including sales proceeds from goods, simplified tax, unified tax paid by products, work and services of small businesses and individual entrepreneurs and other profits (proceeds) from individual entrepreneurs’ individuals, unified tax paid by business activities agricultural producers 35 Table II-2. A core set of taxes for calculation of the index of taxable capacity of an administrative and territorial unit of the basic level Indicator defining a tax base according to the data of Subsection the National Statistics Committee of the Republic of Section Belarus and (or) other sources PIT 01 00 Total gross wages, including small businesses Profit tax 02 00 Before-tax profit of organizations; Sales proceeds from goods, products, work and services of small businesses Fee for area development 04 00 After-tax profit (net profit); Net profit of small businesses Land tax, except for the tax on the 08 00 Taxable area of land including cadastral valuation land located on the territory of the (calculated based on average weighted number) local council, urban settlement, town under regional jurisdiction, paid by individuals Real estate tax assessed on the sites 09 00 Value of buildings and constructions of the on the territory of the town under organization; oblast jurisdiction; Value of buildings and constructions of individuals real estate tax assessed on the sites on the territory of the town under regional jurisdiction which is the center of an administrative and territorial unit and with no budget of its own; real estate tax assessed on the sites in ownership of the organizations located on the territory of the region, except for real estate tax assessed on unfinished building sites; Other taxes and fees from goods 12 00 Sales proceeds from goods, products, work, (work, services), sales proceeds – services, including sales proceeds from goods, simplified tax, unified tax paid by products, work and services of small businesses and individual entrepreneurs and other profits (proceeds) from individual entrepreneurs’ individuals, unified tax paid by business activities agricultural producers 36 Table II-3. A core set of taxes for calculation of the index of taxable capacity of administrative and territorial unit of primary level Indicator defining a tax base according to the data of Subsection the National Statistics Committee of the Republic of Section Belarus and (or) other sources Real estate tax assessed on 07 00 Cost of buildings and constructions of above-level unfinished building sites located on unfinished construction of organizations; the territory of the local council, Cost of buildings and constructions of individuals urban settlement, town under calculated on the basis of unfinished construction regional jurisdiction sites The tax on the land located on the 08 02 Taxable area of land including cadastral valuation territory of the local council, urban (calculated based on average weighted number) settlement, town under regional jurisdiction, paid by individuals Real estate tax assessed on the sites 09 02 Value of buildings and constructions of individuals in ownership of individuals located on the territory of the local council, urban settlement, town under regional jurisdiction Fee for provision of services in 19 01 Data of tax authorities at the location of individuals agroecotourism providing services in agroecotourism State duty imposed in accordance 30 00 The number of resident population ; with the legislative acts The number of provided services subject to state duty Fee from procurers of farm products, 16 03 Volume of procurements (purchase) determined on medicinal herbs, mushrooms, wild procurement (purchasing) prices fruits and berries paid by the payer at the location of taxable activities and assets 37 Annex III Computation of elasticities The elasticity coefficient is a parameter that indicates the percentage change that will occur in one variable (y) when another variable (x) changes one percent. Mathematically, β is the elasticity of variable (y) with respect to variable (x) if Δy/y = β· Δx/x. As was discussed in the main text, elasticities of local government expenditures and revenues with respect to some factors can be used as weights for these factors in a grant distribution formula. Therefore, in this appendix we explain how these elasticity coefficients can be estimated in practice using data on actual expenditures and revenues for different local governments. The estimation strategy is based on two main steps. First, we use the fact that the elasticity for two variables approximates a linear relation between logarithms of those two variables. Second, we take advantage of the availability of a standard statistical method for estimating a linear relationship between two variables, called the linear regression. First, let us see that logarithms capture elasticity. By definition of elasticity, Δy/y = β· Δx/x. At the same time from calculus, it is known that ln(x+Δx)≈ ln(x)+Δx/x. Therefore, ln(y+Δy)–ln(y) ≈ β·[ ln(x+Δx)–ln(x)]. In other words, the elasticity coefficient roughly captures the linear relationship between ln(y) and ln(x). Indeed, if ln(y) is a linear function of ln(x), so that ln(y) =a+b·ln(x), then ln(y+Δy)–ln(y)= [ a+b·ln(x+Δx)]– [a+b·ln(x)]=b· [ ln(x+Δx)– ln(x)]. Ordinary least squares (OLS) estimation (or regression) is a statistical technique that is appropriate to use when the values of one variable—in our case ln(y)— are systematically determined by the values of other variables —in our case ln(x). It is obtained by minimizing the sum of squared residuals, where the residual is a difference between the actual values of ln (yi) and their linear approximations a+b·ln (xi). It can be derived mathematically that the sum of squared residuals is minimized when    lnx   ln y   n lnx   ln y  1 i i i i b i i i     lnx   lnx   n lnx   lnx  1 i i i i i i i  In practice, one does not have to use this formula for the calculations, because most statistical/spreadsheet applications have it pre-programmed as a standard command. Thus, in Microsoft Excel the least square regression can be performed using the worksheet function LINEST or the Regression Analysis Tool from the Analysis ToolPak.17 17 To access tools from the Analysis ToolPak, click Data Analysis in the Analysis group on the Data tab. If the Data Analysis command is not available, the Analysis ToolPak add-in program needs to be loaded. 38 Thus, the application of these regression analysis tools to the Belorussian data finds the following relationships: ln (hui//popi)=2.725 – 0.044·ln (popi) where hui stands for expenditures on housing and utilities in locality i and popi stands for the population of locality i. Thus, according to this estimation, per capita expenditures on housing and utilities have a negative elasticity of about 0.044 with respect to the population size. 39 Annex IV Here we show that, if we want the individual revenue capacities to add up to the aggregate revenue forecast for a particular tax, the value of the adjustment factor for all regions combined (Xt ) should be computed as the weighted average with the weight for each locality proportional to the size of that locality revenue base Bt ij. Indeed, from equation (4) we have RCC j =ΣiRFj • (1–a · (x1i – X1) / X1)· [0.25•(1–a · (xti – Xt) / Xt)· St ij + +0.3•(1–a · (xt-1i – Xt-1) / Xt-1)· St-1 ij +0.45•(1–a · (xt-2i – Xt-2) / Xt-2)· St-2 ij] = RFj•(1–a) Σi [0.25•St ij + 0.3•St-1 ij +0.45•St-2 ij]+ +Σi a · [0.25•St ij · xti / Xt+ 0.3•St-1 ij · xt-1i / Xt-1+0.45•St-2 ij · x t-2 i / Xt-2] Let us define Xt as the weighted national average with the weight for each locality proportional to the size of that locality revenue base Bt ij, that is Xt= Σi St ij · xti Then, substituting this into the expression for RCC j above yields RCC j= RFj •(1–a1)+ RFj • a1 = RFj. 40 Annex V: Framework proposed by the Ministry of Finance for measuring expenditure needs Table V-1. A list of indicators for calculation of spending rights for a region on the whole Designation of sections, subsections Indicator for calculation of expenditures per capita Subsection Section and types Type 01 General public activities Total for subsections 01 01 04 Local government and self-government Number of employees of local government and bodies activities self-government body 01 01 07 State archives Number of resident population 01 10 Other general public activities Ditto 02 National defense Ditto 03 Judicial system, law enforcement and Ditto state security 04 National economy Total for subsections 04 01 General economic affairs Number of resident population 04 02 Agriculture and fishing Ditto 04 04 Industry, construction and architecture Ditto 04 05 Transport Ditto 04 06 Roads Ditto 04 08 Fuel and energy Ditto 04 10 Other national economy activities Ditto 05 Environmental protection Ditto 06 Housing and utilities and housing Total for subsections construction 06 01 Housing construction Number of resident population 06 02 Housing and utilities Ditto 06 03 Municipal improvements Ditto 06 05 Other housing and utilities activities Ditto 07 Healthcare Ditto 08 Physical culture, sports, culture and Ditto mass media 09 Education Total for subsections 09 01 Pre-school education Number of children attending establishments providing pre-school education 09 02 General secondary education Number of students attending establishments providing general secondary education 09 03 Vocational education Number of students attending establishments providing vocational education 09 04 Specialized secondary education Number of students attending establishments providing specialized secondary education 09 06 Retraining and improving qualification Number of resident population of personnel 09 07 Out-of-school education and training Number of students attending establishments providing general secondary education 41 09 09 Other education activities Number of resident population 10 Social policy Total for subsections 10 01 Social protection Number of resident population 10 03 Allowances to families raising children Number of resident population 10 04 Youth policy Number of population from 15 to 29 years of age 10 05 Assistance to employment Number of unemployed citizens registered at labor, employment and social protection authorities 10 06 Assistance in housing provision Number of citizens listed as citizens in need of improved housing conditions 10 08 Other social policy activities Number of resident population 42 Table V-2. A list of indicators to calculate expenditure needs for an administrative and territorial unit of the basic level Subsectio Designation of sections, subsections Indicator for calculation of expenditures per capita Section Type and types n 01 General public activities Total for subsections 01 01 04 Local government and self-government Number of employees of local government and bodies activities self-government body 01 01 07 State archives Number of resident population 01 10 Other general public activities Ditto 03 Judicial system, law enforcement and Ditto state security 04 National economy Total for subsections 04 01 General economic affairs Number of resident population 04 02 Agriculture and fishing Ditto 04 04 Industry, construction and architecture Ditto 04 05 Transport Ditto 04 08 Fuel and energy Ditto 04 10 Other national economy activities Ditto 05 Environmental protection Ditto 06 Housing and utilities and housing Total for subsections construction 06 01 Housing construction Number of resident population 06 02 Housing and utilities Ditto 06 03 Municipal improvements Ditto 06 05 Other housing and utilities activities Ditto 07 Healthcare Ditto 08 Physical culture, sports, culture and Ditto mass media 09 Education Total for subsections 09 01 Pre-school education Number of children attending establishments providing pre-school education 09 02 General secondary education Number of students attending establishments providing general secondary education 09 03 Vocational education Number of students attending establishments providing vocational education 09 04 Specialized secondary education Number of students attending establishments providing specialized secondary education 09 06 Retraining and improving qualification Number of resident population of personnel 09 07 Out-of-school education and training Number of students attending establishments providing general secondary education 09 09 Other education activities Number of resident population 10 Social policy Total for subsections 10 01 Social protection Number of resident population 10 03 Allowances to families raising children Number of resident population 43 10 04 Youth policy Number of population from 15 to 29 years of age 10 06 Assistance in housing provision Number of citizens listed as citizens in need of improved housing conditions 10 08 Other social policy activities Number of resident population 44 Table V-3. A list of indicators to calculate expenditure needs for an administrative and territorial unit of the primary level Subsectio Designation of sections, subsections Indicator for calculation of expenditures per capita Section Type and types n 01 General public activities Total for subsections 01 01 04 Local government and self-government Number of employees of local government and bodies activities self-government body 01 10 Other general public activities Number of resident population 06 03 Settlement improvements Number of resident population 45 Annex VI: Calculations of adjustment coefficients for the RTS estimations of potential profit tax revenues No Region Urbanization Urbanization Urbanization % diff % diff % diff Adjustment Adjustment Adjustment rate (2006) rate (2007) rate (2008) (2006) (2007) (2008) coefficient coefficient coefficient (2006) (2007) (2008) 1 2 3 4 5 6 7 8 9 10 1 Belynichskiy 0.4604 0.4723 0.4810 -42.90% -38.54% -32.38% 0.9815 0.9834 0.9861 2 Bobruiskiy 0.0658 0.0670 0.0680 -91.84% -91.29% -90.44% 0.9605 0.9607 0.9611 3 Bykhovskiy 0.4430 0.4494 0.4552 -45.07% -41.52% -36.00% 0.9806 0.9821 0.9845 4 Glusskiy 0.4136 0.4196 0.4228 -48.70% -45.40% -40.56% 0.9790 0.9805 0.9825 5 Goretskiy 0.6639 0.6704 0.6731 -17.67% -12.76% -5.38% 0.9924 0.9945 0.9977 6 Dribinskiy 0.2215 0.2240 0.2262 -72.53% -70.85% -68.20% 0.9688 0.9695 0.9707 7 Kirov 0.3539 0.3586 0.3638 -56.12% -53.34% -48.85% 0.9759 0.9770 0.9790 8 Klimovichskiy 0.5187 0.5200 0.5211 -35.67% -32.33% -26.75% 0.9846 0.9861 0.9885 9 Klichevskiy 0.4050 0.4090 0.4104 -49.78% -46.77% -42.31% 0.9786 0.9799 0.9818 10 Kostyukovichskiy 0.5692 0.5754 0.5822 -29.42% -25.13% -18.15% 0.9873 0.9892 0.9922 11 Krasnopolskiy 0.5048 0.5095 0.5162 -37.40% -33.69% -27.43% 0.9839 0.9855 0.9882 12 Krichevskiy 0.7501 0.7522 0.7537 -6.98% -2.12% 5.96% 0.9970 0.9991 1.0026 13 Kruglyanskiy 0.4380 0.4427 0.4477 -45.68% -42.39% -37.07% 0.9803 0.9818 0.9840 14 Mogilev 0.0000 0.0000 0.0000 -100.00% -100.00% -100.00% 0.9570 0.9570 0.9570 15 Mstislavskiy 0.4145 0.4189 0.4226 -48.59% -45.48% -40.58% 0.9791 0.9804 0.9825 16 Osipovichskiy 0.7144 0.7192 0.7237 -11.40% -6.41% 1.74% 0.9951 0.9972 1.0007 17 Slavgorodskiy 0.5217 0.5327 0.5380 -35.31% -30.68% -24.36% 0.9848 0.9868 0.9895 18 Khotimskiy 0.4820 0.0000 0.4892 -40.23% -100.00% -31.22% 0.9827 0.9570 0.9866 19 Chausskiy 0.4851 0.4910 0.4965 -39.84% -36.11% -30.20% 0.9829 0.9845 0.9870 20 Cherikovskiy 0.5205 0.5249 0.5285 -35.46% -31.69% -25.70% 0.9847 0.9864 0.9889 21 Shklovskiy 0.4674 0.4761 0.5170 -42.04% -38.04% -27.32% 0.9819 0.9836 0.9882 22 Bobruysk city 1.0000 1.0000 1.0000 24.01% 30.13% 40.59% 1.0103 1.0130 1.0175 23 Mogilev city 1.0000 1.0000 1.0000 24.01% 30.13% 40.59% 1.0103 1.0130 1.0175 24 AVERAGE 0.8064 0.7684 0.7113 0% 0% 0% 1.0000 1.0000 1.0000 46 Annex VII: Calculations of adjustment coefficients for the expenditure needs in housing and utilities No Region Population % diff Adjustment Expenditure Adjusted (2008) coefficient needs expenditure without needs adjustment 1 2 3 4 5 6 1 Belynichskiy 20,099.000 -88.81% 1.0386 6,400,662 6,647,923 2 Bobruiskiy 20,942.000 -88.34% 1.0384 6,669,121 6,925,391 3 Bykhovskiy 35,559.000 -80.20% 1.0349 11,324,004 11,719,041 4 Glusskiy 17,904.000 -90.03% 1.0392 5,701,650 5,924,940 5 Goretskiy 51,013.000 -71.59% 1.0311 16,245,435 16,751,331 6 Dribinskiy 13,011.000 -92.75% 1.0403 4,143,441 4,310,619 7 Kirov 23,818.000 -86.73% 1.0377 7,585,003 7,871,182 8 Klimovichskiy 28,043.000 -84.38% 1.0367 8,930,483 9,258,284 9 Klichevskiy 17,501.000 -90.25% 1.0393 5,573,312 5,792,120 10 Kostyukovichskiy 25,959.000 -85.54% 1.0372 8,266,819 8,574,434 11 Krasnopolskiy 11,767.000 -93.45% 1.0406 3,747,281 3,899,604 12 Krichevskiy 35,696.000 -80.12% 1.0349 11,367,633 11,763,815 13 Kruglyanskiy 16,142.000 -91.01% 1.0396 5,140,529 5,344,038 14 Mogilev 41,761.000 -76.74% 1.0334 13,299,073 13,743,027 15 Mstislavskiy 27,347.000 -84.77% 1.0369 8,708,837 9,029,971 16 Osipovichskiy 51,871.000 -71.11% 1.0309 16,518,671 17,029,642 17 Slavgorodskiy 15,360.000 -91.45% 1.0398 4,891,496 5,086,073 18 Khotimskiy 13,386.000 -92.54% 1.0403 4,262,862 4,434,472 19 Chausskiy 20,966.000 -88.32% 1.0384 6,676,764 6,933,289 20 Cherikovskiy 15,985.000 -91.10% 1.0396 5,090,531 5,292,255 21 Shklovskiy 32,239.000 -82.04% 1.0357 10,266,728 10,633,140 22 Bobruysk city 218,704.000 21.81% 0.9905 69,647,767 68,987,060 23 Mogilev city 371,309.000 106.80% 0.9535 118,245,861 112,752,310 24 AVEARGE 179,548 0% 1.0000 358,703,963 358,703,963 Note: The average is calculated as the weighted average with the weights proportional to population. 47 Annex VIII: Sample procedures proposed by the Ministry of Finance for grant allocation Table VIII-1. Calculation of the index of budget expenditures and fiscal sufficiency for the planned period (2009) No. Region Number Expenditure Index of Index of Budget of resident needs budget taxable sufficiency population expenditures capacity ratio 1 2 3 4 5 6 1 Belynichskiy 20,099 30,554,005 1.0552 0.9833 0.9319 2 Bobruiskiy 20,942 28,644,340 0.9494 0.7646 0.8054 3 Bykhovskiy 35,559 52,349,073 1.0218 0.9355 0.9155 4 Glusskiy 17,904 25,761,142 0.9987 0.8150 0.8161 5 Goretskiy 51,013 70,359,355 0.9573 0.8491 0.8870 6 Dribinskiy 13,011 20,161,743 1.0756 0.8233 0.7654 7 Kirov 23,818 34,142,768 0.9950 0.8713 0.8757 8 Klimovichskiy 28,043 42,418,521 1.0499 0.9265 0.8825 9 Klichevskiy 17,501 25,638,828 1.0169 0.9235 0.9082 10 Kostyukovichskiy 25,959 41,044,661 1.0975 1.1657 1.0622 11 Krasnopolskiy 11,767 18,597,981 1.0970 0.7364 0.6713 12 Krichevskiy 35,696 52,837,775 1.0274 1.2102 1.1779 13 Kruglyanskiy 16,142 24,227,357 1.0418 1.0056 0.9652 14 Mogilev 41,761 56,924,091 0.9461 1.1947 1.2628 15 Mstislavskiy 27,347 39,699,414 1.0076 0.8276 0.8214 16 Osipovichskiy 51,871 77,136,865 1.0322 1.1484 1.1126 17 Slavgorodskiy 15,360 23,440,344 1.0592 0.8885 0.8388 18 Khotimskiy 13,386 19,984,361 1.0362 0.7673 0.7404 19 Chausskiy 20,966 30,912,223 1.0234 0.8804 0.8603 20 Cherikovskiy 15,985 24,089,586 1.0460 0.7884 0.7537 21 Shklovskiy 32,239 45,904,643 0.9883 0.9088 0.9196 22 Bobruysk city 218,704 316,185,003 1.0035 0.9858 0.9824 23 Mogilev city 371,309 521,777,975 0.9754 1.0651 1.0920 24 TOTAL 1,126,382 1,622,792,054 1.0000 1.0000 1.0000 48 Table VIII-2. Calculation of transfers for the planned period (2009) No Region Number Revenue Index of Fiscal Transfer to of resident capacity budget sufficiency augment fiscal population expenditures ratio suffiency 1 2 3 4 5 6 1 Belynichskiy 20,099 13,166,934 1.0552 0.9319 12,608,254 2 Bobruiskiy 20,942 10,668,010 0.9494 0.8054 13,496,195 3 Bykhovskiy 35,559 22,161,974 1.0218 0.9155 21,999,411 4 Glusskiy 17,904 9,721,272 0.9987 0.8161 12,010,683 5 Goretskiy 51,013 28,857,932 0.9573 0.8870 30,496,826 6 Dribinskiy 13,011 7,135,961 1.0756 0.7654 9,872,373 49 7 Kirov 23,818 13,826,018 0.9950 0.8757 14,976,630 8 Klimovichskiy 28,043 17,309,895 1.0499 0.8825 18,474,132 9 Klichevskiy 17,501 10,766,992 1.0169 0.9082 10,861,779 10 Kostyukovichskiy 25,959 20,160,224 1.0622 14,464,822 1.0975 11 Krasnopolskiy 11,767 5,773,020 1.0970 0.6713 9,916,132 12 Krichevskiy 35,696 28,780,683 1.0274 1.1779 15,792,969 13 Kruglyanskiy 16,142 10,813,705 1.0418 0.9652 9,624,358 14 Mogilev 41,761 33,239,336 0.9461 1.2628 14,781,508 15 Mstislavskiy 27,347 15,078,365 1.0076 0.8214 18,411,839 16 Osipovichskiy 51,871 39,685,697 1.0322 1.1126 25,386,531 17 Slavgorodskiy 15,360 9,091,581 1.0592 0.8388 10,682,562 18 Khotimskiy 13,386 6,842,340 1.0362 0.7404 10,016,356 19 Chausskiy 20,966 12,297,508 1.0234 0.8603 13,779,870 20 Cherikovskiy 15,985 8,395,537 1.0460 0.7537 11,926,303 21 Shklovskiy 32,239 19,519,497 0.9883 0.9196 19,205,404 22 Bobruysk city 218,704 143,631,329 1.0035 0.9824 123,100,571 23 Mogilev city 371,309 263,476,086 0.9754 1.0920 176,692,896 24 TOTAL 1,126,382 750,399,897 1.0000 1.0000 618,578,405 50