Measuring the Role of Subnational Governments
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

To measure subnational governments, only external money flows are counted, excluding intra-level transactions in measuring a level of government and all intergovernmental transactions in measuring general government. Control, finance, and administration should be distinguished in measuring centralization and each level’s share of general government, administered expenditures being net of grants given to other governments and financed expenditures net of grants received. Disparate decentralization of finance, control, and administration brings vertical imbalance, measured by the portion of a government’s expenditures not covered by its own resources and by the ratio of intergovernmental grants to total government expenditures.

Abstract

To measure subnational governments, only external money flows are counted, excluding intra-level transactions in measuring a level of government and all intergovernmental transactions in measuring general government. Control, finance, and administration should be distinguished in measuring centralization and each level’s share of general government, administered expenditures being net of grants given to other governments and financed expenditures net of grants received. Disparate decentralization of finance, control, and administration brings vertical imbalance, measured by the portion of a government’s expenditures not covered by its own resources and by the ratio of intergovernmental grants to total government expenditures.

I. Introduction

For subnational government, as for most other organized activities, effective analysis requires measurement. Measuring the operations of a single subnational government is clearly critical to the successful conduct of its affairs. In a wider, national context, concerned with the equalization of burdens and benefits and the universal attainment of minimum standards, the comparability of measurement is important. Comparability, the measurement of an entire level of subnational governments, and measurement of all governments in the country take on additional importance in considering the appropriate assignment of resources and responsibilities, relations between levels of government, macroeconomic effects, and the lessons to be gained from international comparisons. At each level of analysis, though the same basic principles of measurement apply, additional issues arise.

This paper discusses these principles and the issues encountered in the measurement of an individual subnational government, a level of government, and all governments within a country. It then examines indicators of centralization and vertical imbalance utilized in international comparisons and presents a number of comparisons drawn from the International Monetary Fund’s Government Finance Statistics Yearbook (GFSY). 1/

II. Measuring a Single Subnational Government

The first issue in the measurement of a subnational government is the delimitation of its borders, not geographically but functionally. In both the Fund’s A Manual on Government Finance Statistics (GFS Manual) 2/ and the United Nations’ A System of National Accounts (SNA) 3/, government is defined by its function, that is, the promotion of public policy by the production of nonmarket goods and services and the distribution of income, supported primarily by compulsory levies on other sectors, i.e., households and enterprises. The distinguishing feature of government is thus not public ownership or control; that feature defines the public sector, which includes also enterprises under majority government ownership or control. What distinguishes government is that it operates by political means rather than through a market mechanism in its collection of receipts and its distribution of goods, services, and funds. As a result, while the provision of a service, such as water supply, with general revenues is counted as a government activity, its distribution through sale is counted outside government and can be reflected only in the measurement of the industry involved or of the public sector if the enterprise is under majority government ownership or control.

The second issue in measuring a subnational government is whether it is indeed subnational rather than an outlying branch of central government. The purpose of this distinction is to identify separately from the central government any units which because of their separate circumstances are likely to behave differently from the central government itself. The GFS Manual recommends that this distinction be based on whether the subnational government has its own separate existence, with sufficient discretion in the management of its affairs. As possible criteria which may indicate a government’s lack of such a separate existence, the Manual lists: 1) dependence for all or a substantial portion of its income on appropriations or allocations made at the discretion of another government, 2) the lack of its own officers, 3) the need to submit budget estimates to another government entity which may in turn raise or lower them, and 4) control of important aspects of its administration (e.g., sites, contracts, personnel, scope of activities) by another governmental entity. 1/ While these are the criteria recommended to the Fund’s government finance statistics correspondents reporting data for the GFSY. Judgments also take into account established political conventions within the country and difficult questions of degree. 2/

The single denominator by which government activities, other than regulation, are registered is money. To measure the extent or performance of particular government services, one may register units of input such as teacher hours, police time, or hospital beds -- or units of output e.g., students taught, crimes solved, or patients treated. Such units, however, cannot be added together to reach a total or to calculate the portion of the total devoted to each. It is important to keep in mind, therefore, that the basic measurement of each government’s operation is financial and that, whatever their wider implications, most measurements of government represent money flows.

To serve its purpose, measurement of the government’s activities should be comprehensive, covering all of its transactions with the rest of the economy households, enterprises, nonprofit institutions, and other governments. However, it should not include transactions within the same government, that is, intragovernmental transactions. While such transactions may appear in budgets or accounts focussing on control of particular units or subunits, they do not form part of a government’s receipts from or payments to the rest of the economy. If significant, they should appear as memorandum items.

The distinction between a government’s own revenues and the grants it receives from other governments can sometimes be difficult to draw, though it forms a key dividing line in the study of relationships between levels of government. There is no difficulty in attributing to a government the proceeds of a tax it collects and retains, but questions arise regarding taxes collected by one government and passed on to another. The GFS Manual recommends attributing tax revenue to noncollecting beneficiary governments when they (1) exercise some influence or discretion over the setting of the tax or distribution of its proceeds; or (2) automatically receive a given percentage of the tax collected or arising in their territory; or (3) receive tax revenues under a tax law leaving no discretion to the collecting government. 1/

To ensure continuity in the measurement of a government, four compilation tools are especially useful: (1) an inventory or listing of the various parts making up the government, referred to variously as institutional tables 2/ or the universe 3/; (2) a coverage note specifying which parts of the government are covered by the data; 4/ (3) a derivation table identifying the sources of the data and any adjustments made; 5/ and (4) classification keys relating the classifications used to those in the source documents. 6/

III. Measuring a Level of Government

To measure the combination of all governments at a particular level -- state or local one additional operation is necessary. This is the elimination of transactions between governments within this level, so as to portray only the level’s transactions with the rest of the economy, including other levels of government. Only if the total includes transactions with other levels of government -- grants received and grants given, for example -- can the portrayal of the overall operations of this level of government be complete. When significant transactions take place between governments at the same level -- as with horizontal equalization grants between laender in Germany, for example -- they should be shown as memorandum items outside the overall totals for the level.

Not all government units aggregated within the same level of government can be expected to be homogeneous, and separate data for tiers, sublevels, or types of local government can be useful in many circumstances. 7/ The structure of Canadian government, for example, has been represented by Anthony Scott as a federal pyramid, with a single block representing the single central government at the top, next down a wider band of blocks, each representing the government of a state or province with territories that do not overlap, and at the lower, local, government strata many units which to the architectural eye “look more like rubble than the foundation blocks of a great pyramid, for they differ greatly in size, sometimes overlap territorially, and do not have a uniform height from the ground” but each lies within the territory of only one provincial government. 1/ Notwithstanding such heterogeneity at the local level, in most presentations of data for local government, as in the GFSY, all tiers or varieties of local government data are aggregated under the single local government heading. A separate presentation of data for state governments, however, is most important.

Treatment of social security funds can also affect delimitation of the levels of government. Such funds operate mainly at the national level but in a few countries also at the regional or local levels. Because the analysis of social security fund operations is important both in itself and as an integral part of central government fiscal policy when the funds are national, data on social security fund operations are best presented both by themselves and as part of the level of government at which they operate.

Within each member country, operations of the European Communities constitute a separate level of government, not subnational but supranational. Though they are nonresident institutions, forming part of the rest of the world sector, supranational authorities’ operations must be taken into account. Otherwise, the tax burden in each member country would be represented as declining when taxes previously assigned to the national authorities are assigned to the supranational authorities. The GFSY treats supranational authorities as a separate level of government and includes them in general government totals.

Where the number of local governments in a country is large, the use of sample surveys for the collection of data, on a regular basis or between universal compilations in benchmark years, can be helpful. The use of such sample surveys was recommended by regional seminars on local government statistics convened under the auspices of the IMF and the World Bank for countries of Asia, Latin America, English-speaking Africa and the Middle East. 2/ Of the 67 countries providing local government data to the GFSY, however, the use of sample surveys is reported by only eight: Canada, Iran, the Netherlands, South Africa, Spain, Sweden, Switzerland, and the United States. 3/

IV. Measuring General Government

The combination of data for all governments in a country to obtain the totals for general government is no different from the procedures followed in combining individual governments’ data to reach the totals for a single level of government. In this case, however, all transactions between levels of government are eliminated to reach the totals for all government transactions with the rest of the economy households, enterprises, and nonprofit institutions, but no other governments within the country unless coverage is incomplete. Transactions with governments abroad, as with other nonresidents, would of course not be eliminated.

To compile general government totals when reports of intergovernmental grants by donor and recipient governments differ, the more valid figure is chosen and appropriate adjustments are made. Identifiable intergovernmental transactions other than grants, e.g., a government’s employer contributions to social security schemes at other levels of government, are also eliminated.

Following compilation of general government data, problems arise when attribution of general government totals to different levels of government is necessary. With intergovernmental grants eliminated, to whom should their eventual use be attributed-- the grantor or grantee government? This question arises, for example, in calculating the share of each level of government in total government operations for use in comparing centralization or decentralization in different countries or over time.

V. Measuring Centralization or Decentralization

Much of the quantitative comparison of the role of subnational governments or of federalism has centered on a concept of the centralization or decentralization of government, calculated as the ratio of the central government or of subnational governments, respectively, to the totals for general government. 1/ Such a ratio has been employed as a measure of subnational governments, to be related to possible explanatory factors -- e.g., industrialization, per capita income, geographic extent, urbanization, openness, and federal structure 2/-- and to possible effects, such as larger overall government operations. 3/

Though various versions and combinations of centralization ratios have been used, persistent reservations have been raised regarding their meaningfulness and significance. Thus, in a paper presented to the 1962 Congress on Centralization and Decentralization in Public Finance of the International Institute of Public Finance, Alan Williams wrote: “I would like to state categorically that I do not regard crude indices, such as the ratio of grant receipts to total expenditure, or of local authority expenditure to total expenditure of the public sector, as an adequate representation of the degree of centralization or decentralization in the public finances of any particular country.” 1/ More recently, Richard Bird referred to “such artificial and sterile exercises as comparing the degree of ‘centralization’.” 2/

The nature of such reservations regarding the measurement of centralization can perhaps be best understood in the framework of distinctions drawn by Kjeld Philip in 1954. 3/ Examining relationships between levels of government, he distinguished three aspects of each government function: 1) regulation or control, 2) finance, and 3) administration. He described the dynamics of gradual centralization in Denmark, Sweden, England, and the U.S., and showed how aspects of particular functions exercised initially by local governments were shifted gradually and separately to a higher level of government. The shift was led in some cases by central regulation, which strained local government resources and brought on grants to finance continued local administration of the function. 4/ The shift was led in other cases by central government grants, which were followed by regulation to ensure effective use of the funds in local administration. On the tax side, too, such separation could take place, with assessment, rate - setting, and collection taking place at different levels of government.

Much of the dissatisfaction with the measurement of centralization reflects the separation of regulation, finance, and administration in the exercise of some government functions. All regulation, whether of other levels of government, enterprises, or households, is difficult, and perhaps impossible, to measure. 5/ This prompts uneasiness, therefore, with any measure of financial or administrative decentralization, which cannot, by its nature, recognize central regulation or control. Both finance and administration can be measured. Because neither measure alone presents a full picture of how government activities are carried out, however, it is important to clarify the distinction between them.

The difference between finance and administration occurs because of intergovernmental grants, which are financed by the grantor government and administered, in providing goods, services or income to the public, by the grantee government. To measure which level of government administers government operations, expenditures are measured net of grants given to other levels of government. To measure which level of government finances government operations, expenditures are measured including grants given to other governments but net of grants received. 1/ Thus, a state government receiving grants from the central government and making grants to local governments, administers its expenditures net of the grants it gives and finances its expenditures net of the grants it receives.

Centralization ratios comparing each level of government’s tax revenue or general revenue, excluding grants received, are sometimes calculated to accompany centralization ratios for administered expenditures. While such revenue centralization ratios may approximate centralization ratios for financed expenditures, they omit expenditures financed by net borrowing and the use of cash balances, that is, the deficit, which may be more centralized than other government receipts in some countries.

Examination of eight published presentations on centralization shows that of the three concerned primarily with expenditures, one calculated the centralization of expenditures on an administered basis, 2/ and two on both a financed and administered basis. 3/ Of the other five, all calculated centralization ratios or shares on the basis of both administered expenditures and revenues. 4/

Centralization ratios calculated for government consumption, as measured in the national accounts, may be closer to administered expenditures than to financed expenditures in that they omit from the grantor’s expenditures intergovernmental grants given, along with all other transfers given. However, such ratios may seriously understate the expenditure role of central government, for which transfers in pursuit of redistributional objectives are usually far more important than for subnational governments. Also critical in such calculations is whether social security fund operations are counted within or outside central government.

In practice, centralization ratios have been calculated for both total government outlays i.e., expenditure and lending minus repayments (the latter insignificant for most subnational governments) -- and total revenues, as well as for their major components. Frederic Pryor, for example, comparing the centralization ratios of seven capitalist and seven communist countries as regards expenditures for education and for health and welfare, 1/ found the degree of centralization for education, health and welfare expenditures to be more dependent on institutional than economic factors. He concluded also that much of the change in the centralization of aggregate public expenditures was due to the changing composition of expenditures. 2/ Michael Wasylenko found that fiscal decentralization --of either revenues or expenditures -- while higher among industrial countries than among developing countries, did not vary much among industrial or developing countries, and that neither per capita GDP, urbanization, population size, nor territorial extent explained variations in either subgroup, though federal structure and openness of the economy did have an effect. 3/

VI. Measuring Vertical Balance

One result of the disparate decentralization of control, finance and administration is the difference between subnational governments’ own resources and their expenditures, which is sometimes referred to as vertical balance or imbalance. One measure of this difference is found in J.S.H. Hunter’s coefficient of vertical balance. This is calculated as one minus the portion of a subnational government’s expenditure provided for by receipts --of its own taxes, shared taxes, nontax revenues, unconditional grants, conditional grants, and net borrowing --controlled by the higher level of government. 4/ Calculation of the coefficient requires a quantified judgment as to the portion of each source of a government’s funds controlled by a higher level of government.

While this measure compares the control of finance with expenditures, it does not deal with the element of control applied by the higher level of government to expenditures administered and fully financed by the lower level of government. Such an examination was undertaken by Richard Murray. He measured the share of local government expenditures in Sweden for tasks delegated on an obligatory basis, on a regulated basis, and under grants with regulations tied to them, and found that about 20 per cent of total local government expenditures were altogether free from regulations and grant formulas. 1/

Another indication of the separation of finance and control from administration of government functions is to be found in the magnitude and character of intergovernmental grants. Intergovernmental grants are given for three basic purposes: 1) to correct vertical fiscal imbalance between levels of government, 2) to correct horizontal fiscal imbalance in the needs or resources of governments at the same level, and 3) to promote specific purposes. 2/

Without comparisons of individual governments, data aggregated by level of government cannot show the inequalities between governments within that level, nor the equalizing effects of the distribution of grants. 3/ Aggregated data can, however, show the overall importance of grants in each country in the ratio of total intergovernmental grants to total general government expenditures (see Table 1). Among industrial countries, this ratio varies from 6 per cent in Belgium to 27 per cent in Denmark.

Table 1.

Portions of General Government Expenditure Financed by Intergovernmental Grants

article image
Source: International Monetary Fund, Government Finance Statistics Yearbook, 1989, Vol. XIII (Washington, D.C.).

Includes supranational authorities share of general government expenditures in Belgium (2.2 percent), Denmark (2.2 percent), France (1.4 percent), Germany (1.8 percent), Luxembourg (2.7 percent), and United Kingdom (1.9 percent).

Data for general government do not include local government.

In addition, data presented in the GFSY for some countries identify the functions for which intergovernmental grants are given. Grants assigned to no particular function, which may be taken to represent more general support for the correction of vertical or horizontal imbalance, 4/ represent the majority of grants in France, the United Kingdom, Germany and Australia, but only 11 per cent in the United States and a little over 20 per cent in Canada and Denmark.

It is also possible to relate grants given for each function to total administered expenditures for that function, so as to measure the separation between finance and administration of that function, keeping in mind, however, the more general grant support. These data show substantial intergovernmental grant financing of education in Canada, the U.S., Australia, Zimbabwe and Switzerland, but quite little in the U.K., France, Germany, and Denmark. Intergovernmental grants cover some 20 per cent of total government health expenditures in the U.S. and Canada, but very little elsewhere. Grants cover almost half of social security and welfare expenditures in Denmark, and about one-tenth in the U.S., Canada, and Switzerland, but far less elsewhere.

Such data, where available, can usefully supplement the more basic data showing the magnitude of government education, health, and social security and welfare expenditures in each country and the shares administered by each level of government, appearing in Table 2.

Table 2.

Magnitude of General Government Expenditures and Portion Administered by Each Level of Government 1/

(Average of Latest Three Tears Available)

article image
Source: International Monetary Fund, Government finance Statistics Yearbook, 1989, Vol. XII (Washington D.C.).

Excluding Intergovernmental grants.

Data for general government do not include local government.

Includes supranational authorities share of general government expenditures in Belgium (2.2 percent), Denmark (2.2 percent), France (1.4 percent), Germany (1.8 percent), Luxembourg (2.7 percent), and United Kingdom (1.9 percent).

Similar analysis is possible on the receipts side. Table 3 compares the importance of grant receipts to state and local governments in different countries. It then shows both the magnitude, in relation to Gross Domestic Product, of revenues from all taxes, income taxes, property taxes, and domestic taxes on goods and services, and their centralization, as measured by the share of each accounted for by central, state and local governments. What emerges is the relative importance of subnational income tax revenues in Scandinavian and some other northern European countries, of subnational property taxes in countries with English-speaking administrative traditions, and of domestic taxes on goods and services at the state level in the U.S., Canada, India, and Brazil, and to a lesser extent in Australia, Germany, and Colombia.

Table 3.

General Government Tax Revenues as Percent of GDP Portion Attributable to Each Level of Government, and Grants as Percent of State and Local Revenue and Grants

(Average of Latest Three Tears Available)

article image
Source: International Monetary Fund, Government Finance Statistics Yearbook, 1989, Vol. XIII (Washington D.C.).

Includes supranational authorities share of general government total tax revenues for Belgium (1.5 per cent), Denmark (.7 per cent), France (.7 per cent), Germany (.9 per cent), Ireland (1.7 per cent), Luxembourg. (.5 per cent), Netherlands (1.4 per cent), and United Kingdom (1.2 per cent).

Data for general government do not include local government.

VII. Conclusion

Centralization ratios, coefficients of vertical balance, grants-to-expenditure ratios, and subnational shares of general government totals are internal dimensions, measuring the organization by level of general government. The magnitude of government operations, in relation to overall economic activity, is measured by their ratio to Gross Domestic Product, to regional or local product, or to other macroeconomic dimensions. The structure and content of operations -- of a single government, a level of government, or general government -- is measured by the ratios of receipt or payment components to the larger aggregates, such as the ratio of grants received to the sum of grants and own revenues, the ratio of each type of tax to total tax receipts, or the ratio of each function to total expenditures. With the results of such measurements, which bring precision to our ideas, analysis can proceed.

References

  • Bahl, Roy, “The Design of Intergovernmental Transfers in Industrial Countries,” Public Budgeting & Finance, Vol. 6, No. 4, (1986), pp. 322.

    • Search Google Scholar
    • Export Citation
  • Bird, Richard M., Federal Finance in Comparative Perspective, (Toronto: Canadian Tax Foundation, 1986).

  • Break, George, “Changing Roles of Different Levels of Government,” in Margolis, Julius, ed., The Analysis of Public Output, (New York: Columbia University Press, 1970), pp. 163218.

    • Search Google Scholar
    • Export Citation
  • Buglione, Enrico, and France, George, “Skewed Fiscal Federalism in Italy: Implications for Public Expenditure Control,” Public Budgeting & Finance, Vol. 3, No. 3, (1983), pp. 4363.

    • Search Google Scholar
    • Export Citation
  • Bureau de la Statistique du Quebec, Univers du Secteur Publique au Quebec, (Quebec: 1987).

  • Hunter, J.S.H., Federalism and Fiscal Balance, A Comparative Study, (Canberra: Australian National University Press and Centre for Research on Federal Financial Relations, 1977).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, A Manual on Government Finance Statistics, (Washington, D.C.: 1986).

  • International Monetary Fund, Government Finance Statistics Yearbook, Vols. XII and XIII, (Washington, D.C.: 1988 and 1989).

  • International Monetary Fund, IMF Survey, February 23,1987, pp. 5354, May 16, 1988, pp. 146147, (Washington: D.C.).

  • Kaufman, Herbert, Politics and Policies in State and Local Governments, (Englewood Cliffs, N.J.: Prentice-Hall, 1963).

  • Madhusudhan, Ranjana G., and Burkhead, Jesse, “Expenditure Trends in Selected Industrial Countries.” Public Budgeting & Finance, Vol. 7, No. 4, (1987), pp. 4956.

    • Search Google Scholar
    • Export Citation
  • Mathews, Russell, Revenue Sharing in Federal Systems, Research Monograph No. 31, (Canberra: Centre for Research on Federal Financial Relations, The Australian National University, 1980).

    • Search Google Scholar
    • Export Citation
  • Mawhood, Philip, “Decentralization: the Concept and the Practice,” in Mawhood, Philip, ed., Local Government in the Third World, The Experience of Tropical Africa, (Chichester: John Wiley & Sons, 1983), pp. 124.

    • Search Google Scholar
    • Export Citation
  • Murray, Richard, “Central Control of the Local Government Sector in Sweden,” in Gramlich, Edward M., and Ysander, Bengt-Christer, eds., Control of Local Government, (Stockholm: The Industrial Institute for Economic and Social Research, 1985), pp. 297345.

    • Search Google Scholar
    • Export Citation
  • Oates, Wallace E, Federal Fiscalism, (New York: Harcourt Brace Jovanovich, Inc., 1972).

  • Oates, Wallace E., “Searching for Leviathan: An Empirical Study.” American Economic Review, Vol. 65, No. 4, (1985), pp. 748757.

  • Philip, Kjeld, Intergovernmental Fiscal Relations. (Copenhagen: Institute of Economics and History, 1954).

  • Pryor, Frederic, “Elements of a Positive Theory of Public Expenditures,” Finanzarchiv, Vol. 25, (1967), pp. 405430.

  • Pryor, Frederic, Public Expenditure in Communist and Capitalist Countries, (London: George Allen and Unwin Ltd., 1968).

  • Regan, Edward V., “Albany Decrees Medicaid Spending and New York City Pays,” letter to The New York Times, (July 14, 1990), p. 20.

  • Scott, Anthony, An Economic Approach to the Federal Structure, Reprint Series No. 27, (Canberra: Centre for Research on Federal Financial Relations, The Australian National University, 1978). Reprinted from Berkowitz, S.D., ed., Canada’s Third Option. (Toronto: Macmillan).

    • Search Google Scholar
    • Export Citation
  • United Nations, Department of Economic and Social Affairs, Statistical Office of the United Nations, A System of National Accounts. (New York: 1968).

    • Search Google Scholar
    • Export Citation
  • Wasylenko, Michael, “Fiscal Decentralization and Economic Development,” Public Budgeting & Finance, Vol. 7, No. 4, (1987), pp. 5771.

    • Search Google Scholar
    • Export Citation
  • Williams, Alan, “Centralization and Decentralization in Public Finance With Special Reference to Central and Local Government in England and Wales,” Centralization and Decentralization in Public Finance, Congres de Istanbul, XVIIIe Session, International Institute of Public Finance, (Le Haye: 1964).

    • Search Google Scholar
    • Export Citation
  • Wolman, Harold, and Page, Edward, “The Impact of Intergovernmental Grants on Subnational Resource Disparities: A Cross-National Comparison,” Public Budgeting & Finance, Vol. 7, No. 3, (1987), pp. 8298.

    • Search Google Scholar
    • Export Citation
*/

This paper was presented at the Forty-sixth Congress of the International Institute of Public Finance (Brussels, August 27-30, 1990) on “Public Finance with Several Levels of Government.” It will be published in the proceedings of the conference. The author thanks John Brondolo, Barbara Jarvis and Chris Wu for statistical assistance and Beverly McKinney for documentary assistance.

2/

For other criteria, see for example Mawhood (1983).

2/

See end of each country presentation in IMF (1988).

4/

See end of each country presentation in IMF (1988).

5/

Ibid, and IMF (1986) 71-87.

2/

IMF Survey (1987) 53-4, (1988) 146-7.

3/

IMF (1989) coverage notes in country presentations.

1/

See, however, use of the ratio of subnational governments to central government in Bahl (1986) 5-7.

2/

A review of the literature on relationships between decentralization and various national characteristics is provided in Wasylenko (1987) 59-60.

3/

Philip (1954) 18, 22, 36-7.

4/

See, for example, attribution of New York State’s high per capita Medicaid expenses, which exceed those of all other states, to the fact that only in New York are increased services legislated by the state government but paid for by city governments. Letter from New York State Comptroller Edward V. Regan (1990) 20.

5/

Six nonquantitative indicators of the centralization of control are provided by Herbert Kaufman: 1) subordination to detailed central directives; 2) central selection, advancement, and disciplining of staff; 3) operation through a central treasury, requiring central authorization; 4) central inspection, book-keeping, and reporting requirements; 5) frequent referrals and consultations with the center; and 6) staff identification with the center, and central training of staff. Kaufman (1963) 12-3.

1/

Administered expenditures are referred to as “control of resources” and financed expenditures as “expenditures from own resources” in Madhusudhan and Burkhead (1987) 54-5. They are referred to, respectively, as “direct expenditures” and “own-financed expenditures” in Break (1970) 164.

1/

Pryor (1968) 177, 223.

2/

Pryor (1968) 295, and Pryor (1967) 417-8, 428-9.

3/

A study by Harold Wolman and Edward Page making use of disaggregated comparisons to examine the overall equalizing effects of grants to subnational governments found that such grants reduced disparities in Canada and Denmark quite extensively (by 73 per cent in each), moderately in Germany (by 38 per cent), and only marginally in England and the U.S. (by 15 and 8 per cent, respectively), and increased disparities in Australia. Wolman and Page (1987) 87.

4/

A comparison of the bases for determining the total of intergovernmental grants and for their distribution is contained in Bahl (1986) 3-22.