ALLOCATION OF COMMUNITY DEVELOPMENT FUNDS TO GMINAS BY FORMULA: ILLUSTRATIVE ANALYSIS FOR THE COMMUNAL CREDIT AND DEVELOPMENT PROGRAM

Transcription

1 ALLOCATION OF COMMUNITY DEVELOPMENT FUNDS TO GMINAS BY FORMULA: ILLUSTRATIVE ANALYSIS FOR THE COMMUNAL CREDIT AND DEVELOPMENT PROGRAM Prepared for Prepared by Raymond J. Struyk The Urban Institute Support for Economic Growth and Institutional Reform CLIN 0002: Legal and Institutional Reform U.S. Agency for International Development Contract No. PCE-I Under Subcontract to Price Waterhouse Coopers LLP THE URBAN INSTITUTE 2100 M Street, NW Washington, DC (202) March 2000 UI Project

2 TABLE OF CONTENTS EXECUTIVE SUMMARY... i INTRODUCTION...1 Formula Allocation in the CCDP...3 Framework for Analysis...4 Exploring indicators possibly associated with community investment needs...5 Basic patterns...7 Regression Analysis...15 Formula Allocation of Funds...17 Direct Procedure...18 Allocation formula...18 Illustrative allocations...19 Formula derivation...22 Step 1: calculate a score for each gmina using infrastructure deficit proxies...23 Step 2: normalize the scores...23 Step 3: compute a trial allocation of the infrastructure funds to a gmina...23 Step 4: compute amount of the infrastructure grant...23 Step 5: add housing funds...24 Illustrative funds allocation...24 Rural gminas...28 CONCLUSION...29

3 LIST OF CHARTS CHART EX 1... ii CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART CHART

4

5 EXECUTIVE SUMMARY 1 In Poland the national government supports local government investments in housing, communal services and economic development through a number of grant and loan programs, each targeted on a specific type of activity. Investments supported include extending and improving water and wastewater services, expanding infrastructure needed for new housing construction, housing rehabilitation and thermal modernization, and small area revitalization, among others. Because program funds are limited, transfers are typically distributed through an application process. Nevertheless, some programs, such as housing allowances and education grants, are allocating funds by formula. The application approach has the advantage of allowing the national government to channel funds toward those investments it believes important. But it has several disadvantages, as well. The main problem is that those gminas receiving funds through the application process are usually the better organized and better managed gminas, not necessarily those with the greatest investment needs. In particular, small gminas receive comparatively little funding. This paper explores an alternative approach for allocating communal development funds to gminas: allocation is made by using a formula which is structured so as to give larger per capita grants to gminas with greater infrastructure and housing investment needs. The specific context for use of the formula is the Communal Credit and Development Program (CCDP) that is now under development. The objective of that program is to stimulate gminas with the greatest needs to undertake investments, using their own funds and, where financially responsible, with debt financing. Results in the paper show that communal infrastructure needs among urban and urban-rural gminas as measured with the limited data available for this analysis are strongly concentrated in gminas with populations of 25,000 or less. Specifically, the investment needs indicators are the percent of the population in a gmina not served by water, sewer, and gas connections. (These measures have limitations, which are discussed in the text). On the other hand, uniform per capita allocation of a portion of the funds being distributed is made to all gminas to assist with addressing housing problems. This approach was used for housing assistance because no consistent pattern between housing problems (as measured here) and gmina size was identified. 1 This work was prepared for the Government of Poland, State Office on Housing and Urban Development, and was supported by the U.S. Agency for International Development. The authors work for the Urban Institute and are based in the Institute s office in Budapest. They wish to thank Michael Lee, Katie Mark and Krzysztof Jaszczolt for comments on drafts of the paper. All views are strictly those of the authors and are not necessarily those of any of the above mentioned institutions.

6 ii Support for Economic Growth and Institutional Reform Illustrative allocation formulas are developed using two procedures: Under the direct procedure deficit measures are employed in the allocation formula. Under the indirect procedure whose use is warranted when deficit indicators are not collected in a timely way or deficit indicators are available for only a subset of gminas; variables found to be strongly related statistically to the deficit measures are employed in the formula. The following chart illustrates the targeting of resources by gmina size and therefore community investment need under the direct procedure. (The pattern is very similar for the indirect procedure.) Even compared with an allocation in which all gminas receive equal per capita grants the targeting on smaller gminas with greater needs is impressive. The improvement in targeting is even more dramatic when compared to allocations made under competitions. CHART EX 1 Per capita grant direct allocation by gmina size Per capita grant allocation ,000 and below 10,001 to 25,000 25,001 to 50,000 50,001 and above By gmina size The analysis presented here is strictly illustrative, as it is constrained by the indicators of investment needs available to the authors. It is likely that Poland would want to use more comprehensive indicators of housing and communal infrastructure investment needs in preparing an actual allocation formula.

7 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program iii While this analysis was prepared in the context of the CCDP, policy makers should consider applying the approach more comprehensively. Poland has a steadily growing number of relatively small, special purpose programs to encourage local governments to undertake investments in housing and urban infrastructure. Funds are consistently allocated on a competitive basis a process in which smaller gminas fare badly; and this adversely affects target efficiency. The Government should consider consolidating many of these programs into a more general community development block grant program allocated to gminas on a formula basis. Gminas could use the funds for a range of community development purposes specified in the enabling legislation.

8

9 ALLOCATION OF COMMUNITY DEVELOPMENT FUNDS TO GMINAS BY FORMULA: ILLUSTRATIVE ANALYSIS FOR THE COMMUNAL CREDIT AND DEVELOPMENT PROGRAM INTRODUCTION In Poland, like many other countries, the national government supports local government investments in housing, communal services and economic development through a number of grant and loan programs, each targeted on a specific type of activity. Investments supported include extending and improving water and wastewater services, expanding infrastructure needed for new housing construction, housing rehabilitation and thermal modernization, and small area revitalization, among others. Because limited funds are available for each program, funds are typically distributed through an application process. This is usually a competitive process, but sometimes a formula-based allocation is employed. The application approach has the advantage of allowing the national government to target funds at those investments it believes important. But it has several disadvantages as well: Those gminas receiving funds through the application process are usually the better organized and managed gminas, not necessarily those with the greatest investment needs. In particular, small gminas receive comparatively little funding. Gminas find it difficult to make a rational plan for future investments, including the expected support from the national government, because the national government s priorities as reflected in the budget change from year-to-year and the outcome of the application process is uncertain. Gminas are not setting their own priorities and are therefore unable to address what they perceive to be their greatest needs. Instead they are using their own funds to match the relatively cheap or free national government funds that finance national priority investments. 2 An alternative approach is to distribute the national government funds to gminas or groups of gminas or powiats through a formula. Although some Polish programs have adopted this approach, the practice is still somewhat exceptional. For example, a formula is used to allocate the national funds for housing allowances to gminas; 2 The World Bank supports this argument and reviews the experience with decentralized decision making in World Bank, Entering the 21 st Century: World Development Report 1999/2000 (Washington, DC: author, 2000), pp ,

10 Support for Economic Growth and 2 Institutional Reform interestingly, the size of the grant depends on gminas expenditures for housing allowances and their tax effort. But discretionary grants, including those allocated through competitions, dominate. The conditional grants being transferred to gminas from the national government are allocated by the voivods on a discretionary basis. 3 Similarly, the grants and subsidized loan funds from the national and voivodship environmental funds are allocated through competitions in which the grantors priorities and other criteria are announced to potential beneficiaries. But formula allocation is not a panacea. Some of the formula allocations are not consistent with rational policies. Such cases often result from a combination of formulas inherited from the socialist period and politically driven modifications made over a number of years. The grants to gminas in the education sector have been a prominent example of such a case. However, beginning in fiscal year 2000 a new more efficient allocation algorithm is being employed. Another example, from farther afield, is the equalization grant being given to local governments from regional governments in the Russian Federation. As one evaluation of the program states: [the grants] create major disincentives for urban local governments to encourage economic growth and try to collect taxes more effectively. 4 The housing and communal infrastructure sector is singled out as the area most seriously disadvantaged by these incentives. 5 Nevertheless, there is general movement in Central and Eastern Europe toward greater use of formula-based allocations. Such grants are more common for funding ongoing expenses, and few funds for investment purposes are directed to local governments by formula. 6 The general idea for the (CCDP-see below) is to use a formula for distributing the matching grants that are part of the program. The allocation will be explicitly related to communal investment needs. In principle, the formula could also take tax effort into account. In the type of program envisioned in this paper, each gmina or set of gminas could use the funds allocated by the formula within one or two years to support the specified range of investments. If a variety of investments can be funded, then this mechanism is often called a block grant program. 7 Clearly a significant task is to 3 See A. Levitas, The Political Economy of Fiscal Decentralization and Local Government Finance Reform in Poland, (Warsaw: Research Triangle Institute, processed, 1999), pp L. Frienkman, D. Treisman, and S. Titov, Subnational Budgeting in Russia: Preempting a Potential Crisis (Washington, DC: World Bank, Technical Paper no. 452, 1999), pp Ibid. 6 J. Dunn and D. Wetzel, Fiscal Decentralization in Former Socialist Economies: Progress and Prospects (Washington, DC: World Bank, processed, 1999). 7 In Poland, the flexibility afforded to gminas by formula allocations to spend the funds on a range of investments is especially valuable because of the small share of gmina revenue that is actually raised from local sources. See A. Levitas, op. cit.

11 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 3 develop a rational formula for allocating the funds, that is, one that channels more funds to gminas with greater communal investment needs. This paper illustrates the process of developing an allocation formula. The context is the development of a funds allocation formula for the CCDP whose development is now underway. The paper begins with a few notes about the possible use of a formula in CCDP. It then presents an analysis of the variation in community development investment needs among gminas using the data we were able to assemble readily. Following this, we present two formulas based on the analysis and show how funds would be distributed among gminas using them. While the analysis and the formulas are strictly illustrative at this point, due primarily to-data limitations, the results presented here clearly demonstrate that the formula-based approach to fund allocation for community development is indeed feasible in Poland. Formula Allocation in the CCDP The CCDP is being designed to improve the quality of life and investment climate in smaller gminas (under 50,000 population), particularly with respect to housing and 8 community development needs. The underlying thrust of the project turns on stimulating gmina spending on badly needed investment projects. Of special importance are those investments that advance the economic development of gminas and contribute to the quality of everyday life among the inhabitants. These include such items as housing rehabilitation and infrastructure to support both housing development and commercial activity, namely sewage, water, gas, and roads. To stimulate small gminas to undertake the most urgent investment projects and to reduce the ensuing financial burden, one element of CCDP is a grant program under which the State Office on Housing and Urban Development (SOHUD) could cover up to 20 percent of project costs. A special feature of the program is that the available subsidy funds will be allocated annually to groups of powiats on a formula basis. The smaller gminas in each powiat group will then have priority access to these funds for a one-year period. In this way SOHUD will be able to target funds directly at those gminas with the greatest need. Should those gminas in a powiat group fail to present qualifying projects within one year, then the funds would be returned to SOHUD for reallocation. Through an analysis of community development needs among urban and urbanrural gminas, a formula for allocating the matching grants can be constructed. The next section describes the framework for the preliminary analysis of variations across gminas with regard to investment needs. 8 The program is described in R. Struyk, Communal Credit and Development Program: Concept (Warsaw: Urban Institute Report to the State Office for Housing and Urban Development, February 2000).

12 Support for Economic Growth and 4 Institutional Reform Framework for Analysis The first step in constructing a formula for allocating relatively more funds to those communities with greater investment needs is to identify community development investment need indicators that have reliable and frequently updated data sources, preferably across all gminas. We employ a data set on communal needs, which was constructed for another purpose. 9 This data set is adequate for the explanatory purposes of this paper. But more refined analysis may well be necessary for actual implementation of a formula. Indicators that can be constructed from variables in the data set range from sewage hook-up deficits to dwelling stock shortages and are taken from a variety of sources, including the State Statistics Office (GUS). The data set is stronger for urban and urban-rural gminas than for rural gminas. Hence, our analysis is restricted to the 871 gminas in this group. 10 (A possible basis for allocating funds to rural gminas is discussed later in the paper.) While far from ideal, particularly with respect to housing deficits and housing rehabilitation needs, the data set contains sufficient direct measures of investment needs and variables hypothesized to be correlated with these needs to permit an illustrative analysis. From this data set, two classes of variables have been defined. The first are the indicators commonly used to define different levels of urban development. These include the coverage or the share of the population or dwelling units not having access to basic infrastructure services (e.g., water and sewage), and three variables on the quantity of housing relative to the population: persons per dwelling unit, persons per room, and square meters of housing per person. In addition, two direct indicators of investment needs are included in the data set, which cover sewage and water networks. These are defined as the percentage shortage in the length of the respective networks in relation to the existing coverage of each type of infrastructure. 11 Second, a group of associated or explanatory variables was constructed from the data set. These include size of place (as measured by population), unemployment rates (at the powiat and voivodship levels), region of the country, age distribution of the gmina s population, change in population over time, and the rate of increase in the housing stock relative to the increase in population. One might ask, if information on the community development needs of gminas, such as infrastructure deficits, are available, why are the associated or explanatory 9 This data set was used for the paper J. Sierak, A. Galazka, and Z. Grzymala, Investment Needs of Polish Gminas: The Situation as of 31st December 1997 (Warsaw: Paper Prepared for the Metropolitan Development Authority, 1998). 10 Data for urban-rural gminas cover both portions of each gmina. For this reason figures presented here on deficits will differ from published statistics for deficits in urban areas, which exclude the rural portion of urbanrural gminas. 11 These are defined more precisely in Sierak et al., op. cit., Chapter 4, and are further discussed below. In fact, these measures were generally found not to perform as well as straightforward measures generated by GUS.

13 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 5 variables needed? The answer is that it is often the case that data on the direct measures of community development needs are only gathered infrequently. In other cases reliable data on investments will be available only for a sample of gminas. Where direct and timely measures of investment needs are not available for all gminas, the use of variables known to be highly correlated with investment needs can serve as useful proxies. In the following we use direct measures in one formula allocation procedure for distributing national government funds to gminas to assist them in meeting their communal investment needs. In a second procedure we use regression analysis to identify proxies to be used in constructing a formula that can be employed for the same purpose. Exploring indicators possibly associated with community investment needs This section reviews our initial thinking and hypotheses about a range of variables included in the data set that may be associated with community investment needs. After introducing these indicators, the next section examines the simple relation between them and several direct measures of community investment needs. Beyond the limited set of indicators in our data set, another limitation is that the time period covered is comparatively short, Unless otherwise stated, data are for Dynamics, such as population change, are for Mean values of variables are shown in Annex I. We expected size of place, as measured in terms of population, to figure heavily in the analysis of needs variations: specifically, that the smaller the gmina, the greater its basic community investment needs. Decentralization in Poland resulted in the creation of nearly 2,500 gminas. Our analysis, which excludes the rural gminas due to data inconsistencies, covers 871 gminas. Of these, 29 percent have populations of less than 10,000 inhabitants. An additional 45 percent of the selected gminas fall in the category of 10,000 to 25,000 inhabitants (Chart 1). Since the primary purpose of the analysis is to identify those gminas whose infrastructure needs are the greatest, use of the conventional categorization of gminas in which one category includes all gminas with a population of 50,000 or less was considered too gross. Based on preliminary analyses, we settled on the categorization in Chart 1.

14 Support for Economic Growth and 6 Institutional Reform CHART 1 Size of Gminas 25,001 to 50,000 15% 50,001 and above 11% 10,000 and below 29% 10,001 to 25,000 45% We also hypothesized that the age distribution of the population could be correlated with various forms of community development needs. On the one hand, the aging of the population in a gmina relative to other gminas suggests a stagnating or declining local economy over a number of years: economic prospects have been insufficient to retain a normal age distribution. In Poland, with its low labor market mobility, this is an especially strong indicator of local economic problems. 12 Gminas with economic problems are also those typically short on funds for investments in community infrastructure. In part this results from older populations generally requiring more in the realm of services, i.e., higher gmina operating costs, which crowd out investments. So we anticipate that gminas with a high rate of post-working age populations will have comparatively large infrastructure and other deficits. A high share of pre-working age population may also indicate infrastructure needs. Here the logic is the reverse of that for the post-working age population. A high share of young people indicates a growing local economy, one result of that is the arrival of young workers and their families seeking employment. Depending on the supply of new, fully-serviced housing by the gmina and private developers, such growth may produce housing and infrastructure deficits overcrowding and excessive demands on existing infrastructure nets. 12 An analysis of Poland s labor mobility at the beginning of the transition period is provided in S. Mayo and J.I. Stein, Housing and Labor Market Distortions in Poland: Linkages and Policy Implications, Journal of Housing Economics, vol. 4, no. 2, 1995, pp

15 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 7 Following the reasoning just outlined, two related but more general variables were also tested: Percentage change in population (1997 and 1994) Net population immigration as a percentage of base year population (1997 and 1994) Broad region of the country was also explored as a possibly useful shorthand or summary indicator of community investment needs. Gminas were classified by location according to the standard practice of defining voivodships as being in the western, central/southern and eastern parts of the country. These roughly corresponded with distributional mappings of gross national product (GNP) and unemployment data that were available to us at the voivodship level, as well. 13 Basic patterns This section presents descriptive information on the level of community investment requirements suggested by our need variables and on the relationship between these needs and the associated or explanatory variables. We begin with a review of the need variables. The variable ratio of population not served by an infrastructure service to total population was used in three cases: Water hook-ups Sewage connections Gas service These basic deficit measures are the central needs indicators used in the analysis. Other indicators were also tested, including three housing ratios for 1997: Dwelling units to population Persons per room Square meters of housing per person Data on GNP and unemployment rates for 1997 at the voivodship level are from Ministry of Economy, The Assumptions of the National Strategy for the Regional Development in Poland During the Period Between (Warsaw: author, processed, 1999), see Annex of Tables. 14 All of these indicators suffer from a common measurement problem. Poland has a large number of housing units, which are classified as being under construction and therefore are not included in the housing statistics. But many of these units are in fact completed and occupied. This likely produces substantial errors in the housing needs measures. Data on units under construction are not included in our data set.

16 Support for Economic Growth and 8 Institutional Reform We examined the variance in these indicators with gmina size. While moderately greater deficits were identified in smaller gminas for dwellings-to-population, this pattern was not repeated for the other indicators. We thus decided that the preferred method was to simply provide each gmina with the same per capita allocation of funds, as described further below. We studied two other measures developed by Sierak et al., investment needs in water and sewage service. These are expressed as percentages of the existing service level, e.g., kilometers of pipes in the water distribution system. Size of place. The initial hypothesis that smaller gminas have greater infrastructure and housing deficits was borne out by the data as shown in the following three charts. Data are for the 866 urban and urban-rural gminas with populations of 100,000 or less. Each chart has the same format. Each vertical bar shows the situation for a certain range of deficits. For example, the left-hand bar in the following chart is for sewer hook-up deficits up to 30 percent. Each bar displays the percentage of gminas in each of the four population size categories with deficits in this range. The sum of each bar is 100 percent. CHART 2 Sewage hook-up deficit by gmina size 100% 80% 60% 40% 20% 50,001 and above 25,001 to 50,000 10,001 to 25,000 10,000 and below 0% Lowest thru thru thru thru highest Percent of sewage deficit From the chart on sewage connection deficits, one can see that gminas with populations of less than 10,000 have the highest incidence of this problem followed by gminas in the 10,000 and 25,000 population category. Indeed, the two vertical bars for the highest deficit categories (on the right side of the chart) contain almost exclusively gminas under 25,000 population. Cross-tabulations (see Annex II) provides further details. Of the ninety gminas with populations of 50,000 and above, 84 are among those with the highest ratios of sewage hook-ups to the population. This can be

17 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 9 compared to the 49 percent of all gminas with populations of "10,000 and below that carry the largest deficits in sewage connection. Only 5 percent of the smallest gminas are among those with the highest rates of connection. A potentially serious limitation of the sewage deficit indicator should be noted. In low-density areas, septic tanks can provide fully adequate service at a cost far below that of a piped sewage system. A similar argument can apply to piped water and gas systems. Thus, to some degree, the deficit figures overstate the effective deficits in low density, typically smaller gminas. CHART 3 Water hook-up deficit by gmina size 100% 80% 60% 40% 20% 50,001 and above 25,001 to 50,000 10,001 to 25,000 10,000 and below 0% Lowest thru thru thru thru highest Percent of water deficit In the chart for water connections by population size, gminas with populations under 25,000 again clearly have the greatest deficits. Further analysis reveals that of the 213 gminas that occupy the lowest ranks for water service, 105 fall in the population category of 10,000 and below (while an additional 107 are in the 10,000 to 25,000 range). Out of the 89 gminas that represent the population category 25,000 through 50,000, only one gmina suffers from such a low connection rate (for these figures, see the cross-tabulations in Annex II). Gas connection deficits demonstrate a similar pattern as well.

18 Support for Economic Growth and 10 Institutional Reform CHART 4 Gas deficit by gmina size 100% 80% 60% 40% 20% 50,001 and above 25,001 to 50,000 10,001 to 25,000 10,000 and below 0% Lowest thru thru thru Percent of gas deficit As expected, a significant correlation was found between the estimated investment needs for sewage and water networks computed by the earlier study (see Sierak et al.) and the deficit defined by the share of population not connected to water or sewage networks. 15 The Sierak et al., definition of investment need is illustrated using sewage service levels as the example. The following chart shows the distribution of investment needs among gminas with a population of 10,000 or less and among all gminas. Investment need is computed as the increase in kilometers of distribution facilities required meeting the standard expressed as a percentage of the facilities (kilometers) currently in place. Percentages of over 100 are possible where less than half of the necessary services are currently provided. Consider the following example. A gmina has a total need of 52 km for a sewage pipe collection network. It now has only a 7 km net. Its deficit score is 642 [((52 7)/7) * 100]. 15 The simple correlation coefficient is about 0.7.

19 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 11 CHART 5 Sewage investment need distribution among gminas with populations of 10,000 and below 401 thru highest 41% 401 thru highest 27% Lowest Sewage Investment Needs thru 25 8% Lowest 26 thru 92 thru 25 24% 19% 93 thru % 93 thru % 26 thru 92 25% Further analysis of sewage investment needs in the form of the following crosstabulation supports the hypothesis that there is a correlation between size of place and infrastructure needs. Of those gminas with investment needs exceeding 400 percent 79 percent fall in the population categories of 25,000 and below. On the other hand, the majority of gminas occupying a good position with respect to this form of investment need are concentrated in the largest population categories (Chart 6). 16 Similar broad patterns were found for water investment needs The smaller sample size results from the investment measures not being defined for all places. 17 These investment needs measures are not employed further in this paper as some problems were encountered in the investment figures for some of the gminas. These variables also turned out not to be significantly correlated with most explanatory variables when used as dependent variables in the type of regression analysis reported below.

20 Support for Economic Growth and 12 Institutional Reform CHART 6 Age of the population. As argued above, variation among gminas in the age structure of their populations can signal relatively healthy or stagnant local economies. Places with a large share of post-retirement populations are likely characterized by economic performance below the national average and vice-versa for those with a high share of young persons among their residents. Both phenomena carry weight in the analysis of infrastructure deficits. The following two charts illustrate this point. Each vertical bar represents a particular incidence of an age category of the population. For example, in the following chart the left-hand bar is for gminas with a post-working age population constituting less than 12 percent of the total population. The segments of the bar are the four different rates of sewer hook-up deficits defined for the analysis, ranging from under 30 percent to 82 to 100 percent. For sewer hook-ups the highest deficit incidence (the two upper segments in the bars) are for gminas with the highest rate of post-working age population. Similar patterns can be found for water and gas deficits as well (see Annexes III and IV).

21 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 13 CHART 7 Sewage hook-up deficit by post-working age population 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Lowest thru thru thru highest Percent of post-working age population thru highest thru thru Lowest thru As for the relationship between the pre-working age population and infrastructure deficits, the example of sewage deficits shown below should suffice. The higher the incidence of young people (right-hand bars), the greater the share of gminas with high deficits. CHART 8 Sewage hook-up deficit by pre-working age population 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Lowest thru thru thru thru highest Percent of pre-working age population thru highest thru thru Lowest thru 30.00

22 Support for Economic Growth and 14 Institutional Reform Population change. Is usually considered to have a direct effect on community development needs for the kinds of reasons provided above when discussing the aging of a local population and the growth of young people. 18 As a more general variable, it might permit summarizing various growth phenomena associated with population change in a single variable. We examined the relation between gmina investment needs and (a) total population change, 1994 to 1997 and (b) net migration, 1994 to No significant pattern was identified. Region of the country. Two separate variables for region were created with the aim of identifying systematic variance of gminas infrastructure and housing deficits among large areas of Poland. The first was coded according to conventional categories of West, Center/South, and East voivodships (the classification is shown below in the following chart). In fact, large within-region variation in deficits among gminas makes the pattern much less clear-cut than we had anticipated. Chart 9 demonstrates the regional distribution of sewage deficits for gminas. For each region a set of four bars shows the percent of gminas with a range of sewer deficits. The percentages represented by the four bars sum to 100. All three regions have a substantial share of gminas in both the lowest and highest deficit categories. CHART 9 Sewage deficit by Regional Distribution Lowest thru thru thru thru highest Robert R. Nathan, Report on the Allocation of Community Development Funds to Small Cities (Washington, DC: Brookings Institution, Report to the U.S. Department of Housing and Urban Development, processed, 1978). 19 Options for treating tax effort in this type of formula are discussed in H. Bunce and R. Goldberger, City Need and Community Development Funding (Washington, DC: U.S. Department of Housing and Urban Development, Office of Policy Development and Research, 1979).

23 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 15 Region 1 (West) Pomorskie Kujawsko Pomorskie Zachodno-Pomorskie Lubeskie Dolnoslaskie Wielkopolskie Opolskie Region 2 (Center/South) Malopolskie Lodzkie Mazowieckie Silesian Region 3 (East) Warminsko-Mazurskie Podlaskie Podkarpackie Lubelskie Swietokrzyskie A second regional variable simply codes each of the sixteen voivodships separately in order to identify those that have the greatest and least infrastructure needs. Again, the intra-voivodship variance among gminas in community development needs made impossible the identification of clear distinctions among the voivodships for our investment needs indicators. It should be emphasized, however, that mean differences in investment needs do exist among voivodships and regions. But our analysis highlights that size-of-place is much more important than region in identifying community investment needs. Unemployment rates. We had anticipated that economically distressed gminas as indicated by high unemployment rates could have higher investment needs. This would result from their lower tax bases constraining investments. Unemployment data are only available at the powiat and voivodship levels. No significant pattern was found between gmina investment needs and unemployment rates at these higher levels of government. Regression Analysis While the kinds of patterns just discussed provide some basic insights, this kind of information suffers from the limitation that the pattern we may observe for one variable, for example, that between post-working age population and sewer hook-up deficits, may really be based on some third factor, e.g., gmina size, with which both are correlated. Regression analysis allows us to take account of such interrelations and do a better job at identifying the specific contribution of each variable. At least as importantly, regression analysis permits us to examine simultaneously multiple determinants of community development investment needs. Based on the foregoing analysis, two classes of explanatory variables were given primary attention in the regression model; (gmina size and population age distribution). For gmina size, a set of four dummy variables were created to capture the non-linear relationship between the service deficits and size of place. The non-linear pattern is

24 Support for Economic Growth and 16 Institutional Reform evident in Chart 10 for sewer hook-ups. We also experimented with a number of other variables. CHART Percentage of sewage deficits by size of gmina population Sewage hook-up deficit Gmina population Chart 11 presents the estimated regression models that proved most serviceable. The models do a reasonable job in explaining the variation in infrastructure among gminas (as measured here). For water and sewer deficits, about 60 percent of the variance are explained and about one-half of the variance in gas service deficits is accounted for by the independent variables. CHART 11 Basic Regression Model Results (Dependent variables: infrastructure service and housing deficits) VARIABLE: SEWAGE WATER GAS Unstandardized Coefficient (Constant) (16.09) (-20.32) (3.04) Gmina size: 10,000 and below (12.53) (7.48) (12.00) Gmina size: 10,001-25, (11.22) (8.60) (11.00) Gmina size: 25,001-50, (4.33) (3.21) (4.60) Post-working age population (17.69) (20.96) (8.85)

25 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 17 VARIABLE: SEWAGE WATER GAS Unstandardized Coefficient Pre-working age population (15.14) (18.59) (12.15) Adjusted R Square std. Error of estimates ( )=t-statistic. As noted, the variable for size of the gmina was recoded into four dummy variables, which correspond to the population categories used throughout the analysis (i.e., 10,000 and below, 10,001 and 25,000, etc.). The regression results show that relative to gminas with over 50,000 population (the omitted category in the regression), the gminas in the smallest two categories have much larger (and statistically significant) deficits. Further, these deficits are generally at their largest in the smallest places. There are smaller but still statistically significant differences between gminas with populations between 25,000 and 50,000 and larger gminas, after controlling for the effects of the age distribution of the population. We can illustrate these findings using the results for water deficits. Gminas in the smallest two size categories have gmina deficits about 21 percentage points greater than gminas of over 50,000 population. Similarly, those in the 25,000 to 50,000 population categories have deficits averaging 8 percentage points greater than the largest gminas. The consistency of this pattern across all three deficits is striking. The population age variables are also highly significant in all three models. This pattern supports the hypotheses stated earlier. Formula Allocation of Funds Broadly speaking, there are two approaches to developing an allocation formula: The direct procedure uses directly observed indicators of investment deficits in the allocation formula. The indirect procedure uses the proxies in the allocation formula that come from regression or other types of analysis, such as that presented in the last section. In the following two sections both are presented. Each section has two parts. The first outlines the derivation of the allocation formula. The second applies two variants of the formula for allocating funds under a hypothetical PLN 200 million communal credit program to eligible gminas. These are referred to below as Case 1 and Case 2. A final section suggests how funds might be allocated to rural gminas.

26 Support for Economic Growth and 18 Institutional Reform Note the following key points: Case 1 Funds are only allocated to gminas with a population of 100,000 or less. Case 2 Funds are allocated in principle to gminas. However, the actual allocation in practice would be to groups of powiats. The funds to be allocated to any powiat group would simply be the sum of amounts allocated by the formula to the eligible gminas within this powiat group. How the groupings are determined is beyond the scope of this paper. Finally, the formulas developed here included only measures of gminas need for communal infrastructure investment. They could as well take into account the tax revenue capability of gminas, allocating smaller grants to gminas with greater tax bases per capita. For simplicity of presentation we have not included these factors. 19 Direct Procedure Allocation formula The procedure for determining allocations to gminas is shown below. In the formulas in general, lower case variable labels are for individual gminas and upper case for the sum of the values of a variable over all gminas. The basic computation procedure is as follows: score(i) = a1 * (def1(i)/def1) + + a3 * (def3(i)/def3) + a4 * (pop(i)/pop) fund(i) = score(i) * FUND score (i) = an intermediate value, in this case the share of total funds that should go to the ith gmina def1(i) = the ith gmina s deficit for infrastructure 1, i.e., this is the count of households or population without service DEF1 = sum of def1(i) over all gminas In the formula used to allocate the national funds for housing allowances, the size of the grant depends on gminas expenditures for housing allowances and their tax effort. pop(i) = the population of the ith gmina

27 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 19 POP = total population of all eligible gminas fund(i) = the funds allocated to the ith gmina FUND = total funds to be distributed, i.e., PLN 200 million Housing investment need is accounted for by a simple share-of-population variable because we were unable to establish robust statistical relations between housing needs and gmina size. Note that the form of the equation, including the fact that the four coefficients sum to 1.0, gives the result that score (i) is the share of FUND to the ith gmina. We compute fund (i) under two assumptions so we can explore the sensitivity of the allocations to modest changes in the parameters. case 1: a1 = a2 = a3 = a4 =.25 case 2: a1 = a2 = a3 =.3; a4 =.1 Illustrative allocations The results of the allocations for the two cases are shown in the following four charts. The mean per capita grant is PLN 14.6 under the first set of weights, which gives equal weight to all four types of needs. (The mean is computed as the unweighted average of the per capita grants to the gminas receiving grants) Chart 12 shows the distribution of grant amounts among gminas by decile. The average grant received by gminas in the eighth grant decile is about twice that received by those in the second decile. Thus, the difference between an allocation based on a simple per capita allocation procedure and this procedure is very substantial. Chart 12 also gives some information on the difference between the two cases under the direct procedure, i.e., Case 1 under which all four types of needs are weighted equally and Case 2 under which housing receives a weight only one-third as large as each of the three infrastructure variables. As one would expect, the variance in per capita average amounts is greater among the deciles under Case 2 because we are weighting the uniform per capita allocation for housing in which every person gets the same grant in effect less in Case 2 than in Case 1.

28 Support for Economic Growth and 20 Institutional Reform CHART 12 Mean Per Capita Grant Amounts and Grant Amounts by Decile Case 1 Case 2 Mean Deciles * Note: The small difference in the means results from using the unweighted mean values of the grants to the gminas in the calculations. More information on the sensitivity of the per capita grants to changes in the formula weights is provided by Charts 13 and 14. Chart 13 gives the distribution of the absolute values of the differences between the two allocations. The larger differences, i.e., those in the three deciles with the largest differences, are equivalent to about 10 percent of the average per capita grant a large sum. Chart 14 illustrates the differences in allocations that would be experienced by a set of towns selected at random. Again, in some cases these approach 10 percent. This sensitivity underscores the need to examine trial allocations carefully and not to change the formula from year-to-year without very good reason. CHART 13 Distribution of absolute differences in per capita direct allocations to gminas under Case 1 and Case 2 Mean 0.90 Percentiles

29 Allocation of Community Development Funds to Gminas by Formula: Illustrative Analysis for the Communal Credit and Development Program 21 CHART 14 Differences in Per Capita Direct Allocation Between Case 1 and Case 2 for Selected Gminas Gmina Name Case 1 Case 2 Difference ANNOPOL miasto BYTÓW miasto CZELAD CZEMPIÑ miasto DUKLA miasto FROMBORK miasto GRYFINO miasto HAJNÓWKA miasto HEL HRUBIESZÓW miasto KRZEPICE miasto KRZESZOWICE miasto LUBIN miasto LUBLINIEC MOSINA miasto NOWOGARD miasto OZORKÓW miasto PABIANICE miasto RABKA miasto SEJNY miasto TUREK miasto TUSZYN miasto USTKA miasto USTROÑ USTRZYKI DOLNE miasto ZAKOPANE Finally, Chart 15 shows that the formula approach succeeds in targeting more of the available grant funds to smaller gminas than would a simple per capita allocation. The per capita grants of places under 10,000 population are about double those that would be received by places of 50,000 or more. Chart 15 is for Case 1. The same pattern exists for Case 2, except the difference in allocations is slightly greater.

30 Support for Economic Growth and 22 Institutional Reform CHART 15 Per capita grant direct allocation by gmina size: case 1 20 Per capita grant allocation ,000 and below 10,001 to 25,000 25,001 to 50,000 50,001 and above By gmina size Indirect Procedure The question addressed here is: how can one employ the results of the regression analysis in allocating available funds among gminas? An examination of programs in other countries that use formula driven allocations reveals that such results are used as a general basis for the allocation formulas. Seldom are such regression models employed directly as allocation formulas. 20 We follow this general approach, although the regression results are more explicitly used in our procedure. Formula derivation The construction of the formula proceeds in two parts that correspond to the separate allocation for the three infrastructure deficits, on the one hand, and the allocation for the housing component, on the other. As under the direct procedure, we 20 The Community Development Block Grant program in the U.S. has a well-studied and well-documented indirect formula based on extensive analyses of community development needs. See, for example, Harold L. Bunce and Robert L. Goldberg, op., cit., and Report on the Allocation of Community Development Funds to Small Cities, op. cit. On the other hand, Hungary uses an elaborate formula to allocate funds to its local selfgovernments, but it is not based on behavioral analysis of the type presented here but rather on rule-of-thumb per capita assistance amounts for different services, e.g., 75,600 Hungarian forints per secondary school student. (In February 2000, the exchange rate was HUF 255 = $1.) Local governments have the freedom, however, to spend the funds as they see fit, not necessarily on the purposes for which allocations are nominally made. For a description and critique of this system, see Wm. Fox, Intergovernmental Finance in Hungary: Summary and Evaluation. (Washington, DC: Urban Institute, Report Prepared for USAID, processed, 1998).