Article

Financing Cities in Developing Countries

Author(s):
International Monetary Fund. Research Dept.
Published Date:
January 1974
Share
  • ShareShare
Show Summary Details

I. Introduction

The financing of cities in developing countries is a complex topic, which this paper only begins to explore. The paper is divided into six sections. Sections I and II attempt to provide a basic understanding of (a) the problems faced by city governments and (b) some of the factors that determine the extent to which city authorities will, and should, be expected to solve these problems. Section I introduces the reader to aspects of the urbanization process, explains why the process is accompanied by an increasing need for government revenues and expenditure, and indicates that cities have some potential for meeting their currently unmet revenue requirements. Section II investigates the effects that the division of responsibility among governmental levels may have on the provision of urban services. This section reviews the factors influencing decentralization, examines difficulties in measuring decentralization, and considers the special status and powers granted to authorities in the largest cities in several countries.

Sections III, IV, and V examine the patterns in city finances that have resulted as cities have struggled to finance much-needed services. Section III analyzes city government expenditure data for 27 cities in developing countries, and discusses the shortcomings of city data. This section also critically compares actual patterns of city expenditure with patterns predicted or recommended by experts on city finance. Section IV examines city revenues in 35 cities of the developing world and discusses the most likely sources for additional self-raised revenue. Section V deals with intergovernmental aspects of city revenues. Section VI summarizes the conclusions and problems unveiled in this study, and indicates specific areas in which further research is needed.

the problem and its causes

The continued poverty of cities is troublesome but not surprising. Cities are generally allocated revenue sources by higher levels of government. Local governments thus normally depend on the generosity of their too often poverty-stricken parents. But higher levels of government may not relinquish for local use those revenue sources which they themselves value, or consider potentially valuable, until forced to do so by crises at lower government levels. Even then, revenue sharing or grants may be relied upon when the higher levels of government are reluctant to relinquish power.

In many cities the need for public services has simply outstripped the ability to pay for them. Although the so-called push from the countryside may have been overemphasized, stagnant agricultural sectors have contributed to the flow of individuals to the cities.1 Rapid population growth made possible by medical advances has also contributed to high urban growth rates. However, the most important factor contributing to urban growth may well be the increased opportunity, productivity, and higher incomes that are drawing people to the cities. This might suggest that the urban centers should be able to pay for the services that they ought to provide. If they cannot, who can? Such a question oversimplifies the problem. One reason is that urban governments, relative to central governments, may very well have a shortage of administrative ability, and therefore may not be able to draw upon the available resources immediately. Another is that capital outlays needed during the process of rapid urbanization may be large, both absolutely and relatively. Municipal governments in developing countries may well have neither the authority nor the ability to raise the required revenues by borrowing in financial markets. In such cases, there is good reason for some reliance on financing by a higher level of government through grants or direct expenditure. Where administrative abilities permit, increased taxing powers or borrowing authority may be an alternative answer to this problem.

Complicating the financial squeeze on urban authorities is the nature of the activities in which they normally engage. For the most part, these activities are not technologically progressive. As the overall productivity of the economy increases, provided that wages in the various sectors of the economy move concurrently, the per unit cost of output will increase in that part of the economy which is not technologically progressive. 2 A study of local public service costs in the United States found that “rising unit costs have been a major (probably the single most important) source of recent increases in local public budgets.” 3 It concluded that “the historical record together with some further reflections indicate to us that local (especially large city) governments can probably expect that costs will continue to rise cumulatively and at a more rapid rate than those in the rest of the economy, even if there is no increase in either the quantity or quality of the services provided.” 4 This line of reasoning bodes ill for those who must purchase fire protection and education, police, hospital, social, and inspection services. The fact that the bulk of urban services is performed in the nonprogressive sector of the economy means that rising prices will put additional pressures on urban budgets.

urbanization and levels of city expenditure

That the accumulation of people in a small area leads to the need for additional public services cannot be questioned. The disposal of human wastes and garbage becomes a problem. Movements of individuals in an urban area must be coordinated by public transportation, or streets will become overcrowded. Streets must be maintained. Disease prevention and treatment services must be on a higher level, since the increased population density facilitates the spreading of infectious diseases. Crime prevention becomes more essential as strangers go unnoticed in urban neighborhoods and social pressures no longer provide an adequate deterrent. Urbanization is also likely to lead to increased expenditure on public utilities. Large capital outlays are required for water, electricity, and telecommunications systems, often run by the city government. Owing to the high unit costs that accompany small-scale production, many services are not possible in rural settings. But as urban populations increase, economies of scale can be realized and these services grow rapidly. 5 These and other factors lead to an increase in public expenditure as urban populations grow.

It is reasonable to expect the need for local government services to be concentrated more heavily in the large urban areas of developing countries than in those of more developed countries. First, relative to the concentration of population in urban centers, ownership of motor vehicles is much more heavily concentrated in the urban areas of the developing countries than in the urban areas of developed countries (Table 1). This leads to costs for traffic control, road and street maintenance, parking facilities, etc. Second, increased mobility and migration from villages to cities lead to the breakdown of extended family relationships—with a concomitant impact on the need for social welfare programs in urban areas. In highly mobile societies, this phenomenon has existed for some time and welfare programs have been developed. Third, piped water may become essential in order to avoid contamination—a significant change from rural and village water systems in developing countries. Local governments are at least partially responsible for dealing with these problems.

Table 1Selected Cities: Population and Automobiles in Urban Areas Compared with National Totals, Selected Years
CityYearPercentage of

Population
Percentage of

Automobiles
Africa
Nairobi19654.759.4
Far East
Bangkok19667.972.8
Bombay19661.114.5
Djakarta19693.836.6
Seoul197017.646.7
Latin America
Bogotá196911.635.6
Caracas196619.336.1
Mexico (D.F.)196814.845.5
Panama City197030.066.1
Rio de Janeiro19684.711.4
São Paulo19676.125.8
Developed countries
New York (state)19708.96.7
Ontario (province)196934.835.8
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Nairobi, Bangkok, Bombay, Djakarta, Bogotá, Caracas, and São Paulo, see No. 3; for Seoul, No. 23; Mexico (D.F.), No. 26; Panama City, No. 32; Rio de Janeiro, No. 7; New York, No. 55; and Ontario, No. 9.
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Nairobi, Bangkok, Bombay, Djakarta, Bogotá, Caracas, and São Paulo, see No. 3; for Seoul, No. 23; Mexico (D.F.), No. 26; Panama City, No. 32; Rio de Janeiro, No. 7; New York, No. 55; and Ontario, No. 9.

Because of these major changes that accompany a shift from a rural to an urban society, local government per capita expenditure in the large cities is frequently four or five times that of local government in general. As shown in Table 2, only in Uruguay, where a large number of urban services are provided by the Central Government, did the per capita expenditure by local governments in general exceed that by the large city. In Bangkok, Belo Horizonte, Lusaka,6 Recife, Seoul, and Stockholm, per capita expenditure by local authorities was less than twice the average local government expenditure in their respective countries. Although the variation is great, the per capita expenditure of most large cities exceeds that of other local authorities, and is frequently much larger.7

Table 2Selected Countries: Major City Per Capita Expenditure Compared with Total Local Per Capita Expenditure, Selected Years(In local currencies)
CountryPer Capita

Expenditure
Ratio of Major City Per Capita

Expenditure to Total Local

Government Per Capita

Expenditure
Africa(a÷b)
Kenya
(a) Nairobi, 196515.35
(b) Town and country councils, 19682.785.52
Malawi
(a) Blantyre, 19663.18 1
(b) All local authorities, 19660.516.24
Tanzania
(a) Dar es Salaam, 19654.36
(b) Town and country councils, 19650.884.95
Zambia
(a) Lusaka, 196711.07 1
(b) All local authorities, 19676.541.69
Far East
Korea
(a) Seoul, 19709,469.3
(b) All local authorities, 19707,177.11.32
Philippines
(a) Manila, 196465.0 1
(b) All local authorities, 196416.63.92
Sri Lanka
(a) Colombo, 196864.7
(b) All local authorities, 196812.55.18
Thailand
(a) Bangkok, 1968211.90
(b) All municipalities, 1968 2204.681.04
Latin America
Brazil
(a1) Belo Horizonte, 196844.31.38
(a2) Recife, 196841.01.28
(a3) São Paulo. 1968145.94.56
(b) All municipios, 196832.0
Colombia
(a) Bogotá (D.E.), 1967559.7
(b) All municipios, 1967170.93.28
Mexico
(a) Mexico (D.F.), 19682,367.8
(b) All municipios and states, 1968480.64.93
Panama
(a) Panama City, 19689.51
(b) All municipios, 19684.512.11
Uruguay
(a) Montevideo, 19682,950
(b) All municipios, 19683,9500.75
Developed countries
United States
(a1) New York City, 1969888.52.69
(a2) San Francisco, 19691,090.53.30
(b) All local governments, 1969330.2
Sweden
(a) Stockholm, 19675,001
(b) All primary and secondary communes, 19673,6171.38
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each country. For Kenya, see Nos. 21 and 22 (for 1969); for Malawi, No. 25; Tanzania, No. 47; Zambia, Nos. 61 and 63; Korea, Nos. 24 and 23; the Philippines, Nos. 40 and 42; Sri Lanka, No. 45; Thailand, Nos. 50 and 51; Brazil, No. 7; Colombia, No. 13; Mexico, No. 26; Panama, No. 36; Uruguay, No. 56; the United States, No. 54; and Sweden, No. 46.

These figures are for income rather than for expenditure.

Sanitary districts and changwats are not included. If they were included, the per capita figures for (b) would be much smaller.

Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each country. For Kenya, see Nos. 21 and 22 (for 1969); for Malawi, No. 25; Tanzania, No. 47; Zambia, Nos. 61 and 63; Korea, Nos. 24 and 23; the Philippines, Nos. 40 and 42; Sri Lanka, No. 45; Thailand, Nos. 50 and 51; Brazil, No. 7; Colombia, No. 13; Mexico, No. 26; Panama, No. 36; Uruguay, No. 56; the United States, No. 54; and Sweden, No. 46.

These figures are for income rather than for expenditure.

Sanitary districts and changwats are not included. If they were included, the per capita figures for (b) would be much smaller.

An indication of increased expenditure for urban services can be gleaned from data on local government expenditure and expenditure by the urbanized departments or provinces. 8Table 3 indicates the growth that has occurred in the relative importance of municipal expenditure of several developing countries in the past decade. These figures at least temper broad generalizations concerning the increased centralization of government in developing countries. Korea provides a clear example of rapid urbanization and an increase in the relative importance of local public expenditure. To a lesser extent the same phenomenon has been occurring in Colombia, the Philippines, Sri Lanka, and Tanzania, where urban growth has proceeded somewhat less rapidly. However, in Kenya and Turkey, increasing urbanization has been accompanied by declines in the relative importance of local authority expenditure as a share of total public expenditure, perhaps because higher levels of government are providing an increasing amount of urban services.

Table 3Selected Countries: Subnational Government Expenditure as a Percentage of Total Government Expenditure, 1955-70 1
YearBrazilColombiaKenyaKoreaPhilippines 2Sri

Lanka
TanzaniaTurkey
195526.9
195641.215.8
195744.117.8
195842.615.0
195949.942.814.5
196050.341.2(20.4)23.019.44.515.413.2
196147.534.9(17.0)21.419.54.616.49.1
196246.342.0(23.6)30.020.74.714.812.6
196344.742.2(23.8)33.618.34.616.911.2
196437.543.3(23.7)15.433.619.43.920.211.1
196550.340.6(22.4)18.535.321.519.3
196651.037.3(20.6)21.035.622.84.3
196753.340.6(24.0)19.437.322.15.2
196851.318.936.922.25.2
196919.038.122.6
197013.537.124.2
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each country. For Brazil, see No. 7; for Colombia, No. 13; Kenya, No. 22 (for 1972); Korea, No. 23 (for 1961-72); the Philippines, No. 42; Sri Lanka, No. 45 (for 1961-69); Tanzania, No. 47; and Turkey, Nos. 52 and 53.

Figures include expenditure by all subnational authorities. For Colombia, municipal expenditure as a percentage of total government expenditure is given in parentheses.

Philippine data are for revenues, including grants from the Central Government. Some of the data were obtained from official government sources.

Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each country. For Brazil, see No. 7; for Colombia, No. 13; Kenya, No. 22 (for 1972); Korea, No. 23 (for 1961-72); the Philippines, No. 42; Sri Lanka, No. 45 (for 1961-69); Tanzania, No. 47; and Turkey, Nos. 52 and 53.

Figures include expenditure by all subnational authorities. For Colombia, municipal expenditure as a percentage of total government expenditure is given in parentheses.

Philippine data are for revenues, including grants from the Central Government. Some of the data were obtained from official government sources.

As urban growth continues, urban public expenditure becomes a large part of total public expenditure—whether these expenses are met by a municipal, provincial, or national government. Evidence of concentration of capital expenditure in Dar es Salaam (in Tanzania) is striking. Of the development budget expenditure that could be allocated regionally in 1969/70 and 1970/71, 26.3 per cent and 24.8 per cent, respectively, were spent in Dar es Salaam, 9 which accounted for only about 2.6 per cent of the country’s total population in 1970. Even if the 58 per cent in 1969/70 and the 52 per cent in 1970/71 of development expenditure that could not be allocated by region were spent outside Dar es Salaam (an extreme assumption), concentration of development expenditure in Dar es Salaam would have been very high relative to population. 10 This may be partly due to the fact that Dar es Salaam is growing rapidly—during the 1960s its population grew more than two and a half times as rapidly as that of the nation. 11 It would not be surprising if similar concentrations of capital expenditure were found to exist in other countries with rapidly growing urban populations.

urban development and revenue potential

Urbanization increases the costs of government but simultaneously increases government revenues, as the cities are the focus of economic development. This has been true since industrialization led to the development of cities toward the end of the Middle Ages. The relationship between per capita gross domestic product (GDP) and the level of urbanization is striking, as shown in Table 4. In countries with per capita GDP below $100, 10.4 per cent of the population, on average, was urban; where the GDP was more than $500, the percentage was as high as 68.1.

Table 4Developing Countries: Relationship Between Per Capita Gross Domestic Product (GDP) and Urban Population
Number of

Countries
Per Capita GDPUrban Population

as a Percentage

of the Total 1
21less than $10010.4
22$100-19920.7
19$200-29939.4
11$300-49935.2
14$500 or more68.1
Sources: See the Sources Used in Preparing Tables, in the Appendix, Nos. 5 and 1.

Urban population includes all those living in cities of 100,000 or more in 1970, Table 5. Selected Cities: Population, Industrial Employment, Electricity Consumption, and Telephone Use in Urban Areas Compared with National Totals, Selected Years

Sources: See the Sources Used in Preparing Tables, in the Appendix, Nos. 5 and 1.

Urban population includes all those living in cities of 100,000 or more in 1970, Table 5. Selected Cities: Population, Industrial Employment, Electricity Consumption, and Telephone Use in Urban Areas Compared with National Totals, Selected Years

Some additional statistics provide further evidence of the concentration of what might be described as modernization in developing nations. Table 1 indicated the overwhelming concentration of motor vehicles in the principal cities of these countries. As well as generating a need for roads, these vehicles also indicate a concentration of wealth in the urban centers. In the Asian cities of Bangkok, Bombay, and Djakarta, and in Nairobi, Kenya, the concentration of motor vehicles is nine or more times what it would be if they were distributed evenly on a geographical basis according to population. In the other cities in Table 1, the concentration is somewhat less marked. Table 5 gives further evidence of the importance of cities as a focus for development. For example, although Saigon accounts for 10 per cent of national population, it accounts for 53 per cent of commercial and industrial employment. Cities are also the first places where modern means of communications and lighting are employed in a country. Table 5 indicates the degree to which the use of telephones and electricity is concentrated in several principal cities. Regional GDP figures normally indicate a concentration of production in urban regions. In the Colombian departments with the four largest cities, 1964 per capita GDP exceeded the national average by 50 per cent and was more than four times that in three of the more rural provinces.12

Table 5Selected Cities: Population, Industrial Employment, Electricity Consumption, and Telephone Use in Urban Areas Compared with National Totals, Selected Years(City totals as a percentage of national totals)
CityYearPopulationIndustrial

Employment
ElectricityTelephone

Subscribers
Africa
Dar es Salaam19652.244.01
Nairobi 2196863.81
Far East
Bangkok19696.777.2
Manila319617.859.24
Saigon196910.353.1579.81
Seoul197017.618.6628.842.9
Latin America
Antioquia (Medellín) 7196313.324.5629.2822.1
Cundinamarca (Bogotá) 7196314.430.8628.3840.8
Mexico (D.F.)196814.835.56(1965)15.71(1965)53.2
Panama City and Colón196838.683.4684.094.1
Rio de Janeiro19684.79.9624.6
São Paulo 7196818.850.6639.6
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Dar es Salaam, see No. 47; for Nairobi, No. 22 (for 1969); Bangkok, No. 51; Manila, No. 38; Saigon, No. 60; Seoul, No. 23; Antioquia and Cundinamarca, No. 12; Mexico (D.F.), Nos. 26, 27, and 28; Panama City and Colón, Nos. 33, 35, and 34; Rio de Janeiro and São Paulo, Nos. 7 and 8.

Electricity sold.

Greater Nairobi, including Mount Kenya area.

Metropolitan Manila.

Gross receipts for electric light and power.

Employment in industrial and commercial establishments.

Manufacturing employment.

These figures pertain to the departments or states rather than the cities.

Electricity generators in 1968.

Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Dar es Salaam, see No. 47; for Nairobi, No. 22 (for 1969); Bangkok, No. 51; Manila, No. 38; Saigon, No. 60; Seoul, No. 23; Antioquia and Cundinamarca, No. 12; Mexico (D.F.), Nos. 26, 27, and 28; Panama City and Colón, Nos. 33, 35, and 34; Rio de Janeiro and São Paulo, Nos. 7 and 8.

Electricity sold.

Greater Nairobi, including Mount Kenya area.

Metropolitan Manila.

Gross receipts for electric light and power.

Employment in industrial and commercial establishments.

Manufacturing employment.

These figures pertain to the departments or states rather than the cities.

Electricity generators in 1968.

The question still arises as to whether the wealth of cities is likely to allow them to deal effectively with their needs. It may not. Much of the government revenue derived from urban areas flows to higher levels of government. These levels of government are neither compelled to return the funds in the form of grants nor to spend the revenues on urban services. Indeed, political unrest in outlying provinces may dictate that the funds be spent on military or agricultural projects outside the urban centers.13 The political structure of the country may also have a rural bias, with underrepresentation of the urban sector and a continual subsidization of the rural sector by the urban centers. Countries where this has occurred include France, the United States, and perhaps India. 14 A major problem with such arguments is, of course, that it is exceedingly difficult to measure government expenditure that benefits a certain area relative to the revenues derived from that area. Such measurements must be based on tax burden and expenditure benefit assumptions, and the results are likely to be highly dependent upon these assumptions.15

The following section examines some factors affecting the division of responsibility among government levels and the effect that this might have on both the ability of city authorities to meet their needs and their reliance on higher levels of government.

II. Decentralization and Urbanization—An Overview

In developing countries, no level of government is likely to have a surplus of revenues or administrative talent. The question, which thus arises, is how best to meet the needs of the cities within the existing economic, political, and social constraints. In particular, to what extent can this end be most efficiently accomplished by “decentralization,” or the allocation of decision-making responsibility or power to subnational levels of government? An increase in the quantity of local government resources and/or local government authority to utilize those resources freely means greater decentralization.16

factors influencing decentralization

On the one hand, the devolution of increased powers to city authorities may prove beneficial for a number of reasons. (1) City officials, through their more intimate knowledge of the tastes and wishes of local inhabitants, may best be able to meet the needs of their constituents. (2) Provision of services by local authorities can also ensure greater diversity in the quantity and quality of government services. Even in developing countries this may be of some significance in the largest metropolitan centers. 17 (3) In many countries, tax consciousness is at a low level. In order to expand the public sector, it may be necessary to develop closer ties between tax revenues and benefits, and this can perhaps best be done at local levels. 18 (4) Government bureaucracies in many developing countries can stifle, as well as facilitate, the conduct of business. Decisions made at the local level may put an end to a long chain of command and the related paperwork and delays. (5) Administrative resources and initiative are limited in all developing countries, and an increase in local authority may allow local talents to be utilized more fully. 19 (6) Increased local responsibility may facilitate experimentation, resulting in greater efficiency in the provision of government services. All these arguments as well as equally important political and social reasons suggest that greater authority for city governments may improve the public services.

On the other hand, there are strong arguments in favor of increased centralization of government authority, as this may facilitate the achievement of greater economic stabilization, more equitable distribution of income, and greater uniformity in the provision of public services. 20 However, it seems unlikely that decentralization in any developing country will be so extensive as to seriously threaten the stabilizing role of the central government. In many developing countries, increased equality in incomes and wealth is viewed as a goal that is secondary to increasing the rate of economic growth, and, again, it seems improbable that the likely degree of decentralization will seriously reduce the central government’s ability to redistribute income if it wishes to do so. Another argument for centralization is that it can avoid spillovers from government activities and can increase the likelihood that the optimum level of public goods and services will be provided. Some people have also claimed that the centralization of fiscal powers in developing countries facilitates economic planning and, thus, nation building. However, the extent to which centralization facilitates planning is far from clear, and the need for nation building should diminish generally as colonial status recedes further into the past. 21 Finally, a high degree of centralization might also be justified on the grounds that subnational governments have little administrative capacity. This is true, but every effort should be made to use the administrative capacity that the lower levels possess; to some extent, decentralization may increase the amount of administrative resources that are available at the local level.

In addition, a number of forces exist that are likely to lead to increasing centralization, or to prevent any reduction in centralization regardless of the original situation. First, higher-income levels accompanied by increased mobility of resources lead to a demand for greater uniformity in the provision of public services throughout the nation. Second, if the central government uses the more progressive taxes, growing incomes (probably helped along by inflation) will lead to an increase in the share of the economy’s resources that are controlled by the central government. 22 Third, the elites of new states may overreact to any move toward local autonomy, the consequence being increased centralization. According to Geertz:

The vivid memory on the part of the elites of the New States of colonial divide-and-rule policies, as well as a fear of the strong divisive tendencies they quite accurately perceive both within themselves and in the mass of the population, leads them to regard any explicit and frank concern with internal diversity as subversive of the whole nation-building effort and to view, or try to view, their own society in a much too uniform and global, even stereotypic a manner. 23

A fourth factor that may inhibit any decrease in centralization is the likely concentration of administrative resources in the hands of the central government, following independence. The central government’s control over revenue sources makes it likely that it will pay the salaries and benefits that attract these scarce resources. Concentration of administrative resources may inhibit future decentralization even when nation-building objectives are met and greater decentralization becomes desirable. Finally, centralization is further facilitated by international aid. Foreign governments and international agencies deal with the central government, help to improve central government programs, and provide training for central government officials.

There thus exist significant noneconomic as well as economic forces that can lead to an excessive and perhaps increasing degree of centralization in developing countries, while forces in the reverse direction are not as apparent, despite the possible benefits of more decentralization. As a result, it is not unlikely that centralization of fiscal powers in many developing countries is already excessive, although in some it may still be insufficient.

problems in measuring decentralization

There is no simple index to measure the degree of decentralization of power. Power is an extremely “complex phenomenon and its distribution difficult to measure.” 24 This is true for a number of reasons that become increasingly apparent as countries are examined.

First, in a number of countries, decentralized autonomous government enterprises comprise an important sector of the economy. Since these enterprises are sometimes subsidized by the central government and sometimes pay net profits to the central government, there is a rationale for viewing them from a fiscal point of view as a part of the central government. An effort has been made to exclude autonomous government enterprises from the data examined in this paper. Had their gross revenues been included in central government revenues, centralization would have appeared much greater. For example, gross revenues of these autonomous agencies in Mexico (1969), Paraguay (1971), Venezuela (1969), and Peru (1970) were equal in size to 100 per cent, 46 per cent, 35 per cent, and 22 per cent, respectively, of all other central government revenues.

A second problem arises concerning the incompleteness of data on local government revenues and expenditure. The problem is particularly acute for rural local authorities, even in those countries where data on urban areas are readily available. For example, in India, Indonesia, and Pakistan, where rural populations are large, data are not readily available for the rural areas. Any index of decentralization derived from available figures will understate the importance of local authorities. In Viet-Nam, a country at war, a unique problem arises in that taxes are paid to the Vietcong. If estimates of taxes paid to the Vietcong are reasonable, the share of village revenues in the total government revenues rises from 1.9 per cent to 7.5 per cent, a substantial increase. 25

Third, decentralization may also refer to actual services provided, rather than merely to revenues raised and funds expended. Local authorities may be little more than employment services providing jobs for a number of local residents. An examination of Colombian municipalities found that expenditure on overhead 26 accounted for 40 per cent of total expenditure for all but the four largest municipalities, where it was 13 per cent. 27 An in-depth examination of local authorities in other developing countries is likely to show that overhead costs similarly account for a large part of the local budgets; however, where “community action” provides local services, the budget may understate the contribution of the local public sector. Indeed, to the extent that overhead expenditure helps to foster community action, it may be very productive.

The type of grant employed by central governments in the allocation of funds to local authorities is the fourth factor that will affect the extent to which power is decentralized in the government system. Several possibilities exist, including functional and unconditional grants. For functional grants, the freedom of local authorities to utilize funds as they see fit is strictly limited. A greater devolution of power accompanies the use of unconditional grants than any other form of grant. To gain an accurate picture of the degree of centralization in any nation, one cannot overlook the strings attached to grants—particularly where grants form an important part of total government revenue.

Fifth, the extent to which central governments can vary the revenues that they collect for local governments has a bearing on local powers. Sometimes the revenues distributed are subject only to the discretion of the central authorities. Local authorities have little influence over the amount of taxes collected for them in this manner. This is true in Brazil, India, the Philippines, and Thailand, as well as in other countries. Generally these revenues can be spent as the local authorities wish, and in this way they are similar to block grants. However, in some cases, such as the urban land tax (Iuran Rehabilitasi Daerah—IREDA) in Indonesia, the functions upon which the revenues have to be spent are clearly specified. Cases in which local authorities have some freedom to alter the rates of tax collected for them by the higher authorities appear to be concentrated in high-income countries. With rare exceptions, such as the gross receipts tax in Mexico City (which has only an upper limit), it appears that the local authorities in developing countries are not endowed with the statutory power to alter the rates on taxes collected for them by higher authorities.

Finally, one further word of caution is needed. Decentralization, at least to some extent, refers to the degree to which government remains close to the people. In geographically small countries, such as Panama, Sri Lanka, the Republic of China, and the United Kingdom, people are close to the seat of the central government. Representatives to the central government can be in closer contact with their constituents, and constituents can descend upon the seat of the central government much more easily. There appears to be no sound economic or political reason for the existence of a two-tier system of local government similar to those in the federal states. When speaking of the optimal degree of concentration, each country must be analyzed thoroughly, with recommendations based on this analysis and on relevant experience gained in other countries. 28 Because of these problems in measurement, any useful comparison of the degree of centralization in developing countries is left for a later study.

special treatment of capital districts

In many countries of the world, the urban area that serves as the seat of the central government is designated as a “special district” with powers exceeding those of most other urban areas. This seems particularly true in the Western Hemisphere.

The city of Rio de Janeiro, which formerly served as the seat of the Brazilian Government, coincides with the State of Guanabara. Several advantages stem from this. First, those taxes not allocated by the Constitution to be used by a particular level of government may be used by either the federal or state governments but not by municipalities. Second, the value-added tax (VAT), export taxes, and taxes on the transfer of real property are allocated to the states by the Constitution. This is particularly important in times of inflation, since VAT revenues keep up with the price level much better than do municipal revenues in general. Third, about half of the federal government revenues derived from taxes on electric power and fuel oil is apportioned directly to the states, and only about 10 per cent is apportioned to the municipalities. 29

In Venezuela, the Federal District (in which most of Greater Caracas is located) also combines the powers of state and municipal governments. The Constitution guarantees the distribution of 15 per cent of central government revenues to states, federal territories, and the Federal District—30 per cent to be in equal amounts and 70 per cent by population. 30 The situation in the Special District of Bogotá is similar. In Colombia, as in Venezuela, nearly all central government aid flows to the higher tier of local government (in Venezuela this is the state and in Colombia the department); Colombia also distributes 30 per cent in equal amounts and 70 per cent by population. 31

The Federal District of Mexico also benefits significantly from its special status. On equal footing with the states, it received more than 30 per cent of its revenues in 1963 from its surtax on the federal commercial receipts tax plus its share of national taxes on gasoline, beer, tobacco, and automobiles. 32 In Argentina, the Federal District of Buenos Aires also has a special status. Whereas Argentinian states are generally responsible for primary education, justice, police services, fire prevention, registration of persons and real estate, water supply and sewage, minor public works, and some public health services, the Federal Government provides these services for the Federal District. Outside the Federal District, state residents are subject to additional taxes on gross receipts and property in order to cover the cost of these services, but the Federal Government covers the cost for the Federal District. Since city officials in Buenos Aires do not have to worry about the addition of state gross receipts and property taxes to the local rates, they have greater freedom to raise the local rates. 33

In Asia, both Djakarta and Seoul are central government seats holding special status in a two-tier system of local government. Being on equal footing with the provinces enables Djakarta to levy taxes on motor vehicles and to receive a central government transfer in lieu of export taxes. In 1969/70, budget figures indicated that these two sources were expected to provide Djakarta with about 29 per cent of its total revenues. In Seoul, total funds from the acquisition, automobile, husbandry, entertainment, and license taxes are available for solving municipal problems without any split of these taxes between the sub-provincial and provincial levels of government. Chartered cities in the Philippines also benefit from some preference in that they receive one share of the Central Government’s distribution of tax revenues as a province and another share as a municipality.

The following sections suggest that the special status thus conferred on some large cities has enabled them to increase their revenues and expenditure. Most obvious is the special case of Singapore, where neither expenditure responsibilities nor revenue sources need be shared with higher government levels. Djakarta, Caracas, and Mexico City are other capital districts that appear to have benefited from their special status.

III. City Expenditure

problems with measuring expenditure patterns

This section examines city expenditure patterns in developing countries. Before the data are examined, some of their shortcomings are noted to ensure that conclusions drawn from the data are treated with the appropriate degree of caution.

First, expenditure frequently is not carefully and uniformly allocated by function. Many variations of this problem exist: (1) “General administration” in some cities may include expenditure on courts and police that is specifically allocated by function in other cities. (2) Where functional breakdowns do exist, different countries may use different terms to refer to expenditure for the same function, or the identical term to refer to expenditure for different purposes. For example, what part of public works expenditure is for roads, streets, water systems, school buildings, etc.? (3) In cities in at least two countries, separate supplementary payments are made to present and/or past employees as part of their remuneration, and these payments are not allocated by function. In Colombia, previsión social is designated as a separate function, whereas it is primarily part of the cost of providing other public services. 34 In Indonesia, additional payments to government employees are listed under kesedjahteraan pegawai in Djakarta but under different categories in other large cities. Again, this makes it impossible to estimate accurately the expenditure in particular categories. (4) The “other expenditure” category is frequently large in city data. This appears to be a particular problem in Brazilian cities where 45 per cent or more of total expenditure cannot be allocated by function on the basis of available data. In other cities, even though information on current expenditure may be available by function, capital expenditure information is not. As a result, it is difficult to ascertain the full amount of resources devoted to each function. Dick Netzer found the lack of adequate functional breakdowns in municipal finance data to be a major problem in analyzing local public expenditure in Colombia. 35 (5) In several instances, two or more functions were lumped together in the data relied upon for this study. Two of the more significant cases relate to Bangkok and Saigon: in Bangkok, total expenditure by the Education and Public Welfare Bureau were lumped together in the available data and have been placed under education; in Saigon, expenditure data on public health and social welfare were found to be combined and have been placed under public health. All these problems increase the difficulty of undertaking comparative analysis.

Second, it is often unclear which publicly run enterprise should be considered as part of the public sector and which should be considered as part of the private sector. For example, in the Indian cities, enterprises providing electricity and telephones may be controlled by the municipality, by the state, or by private corporations. In Lagos, Nigeria, the responsibility for providing water is divided between the city and the Central Government. In the three Colombian cities, municipal enterprises have the primary responsibility for providing electricity, telephones, and water. The finances of these Colombian municipal enterprises are recorded in such a manner that their revenues equal expenditure, and they make no net contribution to municipal revenues. By excluding the gross municipal expenditure on water, electricity, and telephones, the relative importance of other expenditure categories rises in Colombian cities vis-à-vis cities in other countries. It is not clear which enterprise in each city should be treated as equivalent to private businesses with expenditure excluded from the public sector, and which enterprise should be treated as part of the public sector with expenditure included. Available municipal data frequently do not provide enough information to allow a systematic and consistent treatment of this issue.

Third, subtle differences in the form of central government aid may lead to large apparent differences in the relative importance of a given function in overall municipal expenditure. For example, 28.5 per cent of total local expenditure is for educational purposes in Freetown, Sierra Leone. Most of this expenditure is financed by a grant from the Central Government to the municipal government to cover the cost of primary teachers’ salaries and the cost of capital improvements. In order for a salary to be covered by the Central Government, the teacher must be approved by the Central Government. In Nairobi, expenditure on education appears to be of relatively less importance, as it equals only 21 per cent of total expenditure. Nairobi received no aid for education from the Central Government in 1965. If, however, the Central Government of Sierra Leone had paid the Freetown teachers directly, rather than channeling the payments through the municipal government, expenditure on education in Freetown would have been a smaller percentage of total local expenditure than in Nairobi—probably less than half as large. In Colombo and in Mexico City the fact that the Central Government is responsible for education causes other expenditure to appear relatively more important.

Fourth, in analyzing local budgets it must not be concluded that the absence of a particular function from the local budget means that the function has been assigned to higher levels of government. In Seoul, Korea, as in U.S. municipalities, local school authorities exist separately from other local authorities, and education remains very much a responsibility of the local community. Therefore, in order to see the relative importance of different local expenditure, it is important to try to obtain information on all the local authorities that operate within the area.

expenditure patterns

Table 6 provides a few insights into municipal expenditure patterns, although, for the reasons mentioned, the data must be interpreted cautiously and only the broadest type of generalization should be derived from them. Apparently, four categories normally account for the bulk of total municipal expenditure—general administration, public health, education, and public works.

Table 6Selected Cities: Distribution of City Expenditure by Function as a Percentage of Total Expenditure, Selected Years
Function
CityYearGeneral administrationPublic healthEducationRoads and HighwaysSanitation and sewageWaterElectricityPublic works 1
Africa
Dar es Salaam196513.328.225.624.8
Freetown196428.32.1228.522.5
Lagos196313.730.021.020.131.9
Nairobi19659.519.820.917.47.8
Far East
Bangkok196827.814.917.0534.0
Djakarta197129.262.118.50.21.714.8
Semarang197027.365.36.011.1
Solo197033.463.93.91.48.3
Manila19679.07.836.86.26.7
Saigon196913.424.6725.618.1
Seoul19705.12.618.60.24.59.028.68
Singapore196911.510.020.94.3
South Asia
Ahmedabad196812.336.920.14.2
Bombay19685.431.319.69.7
Calcutta196836.327.93.86.8
Hyderabad196717.122.70.522.7
Madras19686.617.818.018.9
Colombo196833.524.35.818.5
Latin America
Belo Horizonte196821.15.210.2
Recife196831.52.95.1
Rio de Janeiro196840.911.416.2
São Paulo196810.55.210.6
Bogotá (D.E.) 11196926.31217.213
Cali 11196924.01246.313
Medellín 11196922.31216.913
Mexico (D.F.)196238.319.026.7
Venezuela (D.F.)196915.125.9146.410.79.915
Developed countries
New York City19692.110.020.63.32.81.1
San Francisco19692.66.315.02.11.24.2
Stockholm19682.318.712.210.92.89.018
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Dar es Salaam, see No. 47; Freetown, No. 43; Lagos, No. 31; Nairobi, No. 21; Bangkok, No. 50; Djakarta, Semarang, and Solo, No. 20; Manila, No. 39, p. 624; Saigon, No. 60; Seoul, No. 24; Singapore, No. 44; Ahmedabad, Bombay, Calcutta, Hyderabad, and Madras, No. 17; Colombo, No. 45; Belo Horizonte, Recife, Rio de Janeiro, and São Paulo, No. 7; Bogotá (D.E.), Cali, and Medellín, No. 13; Mexico (D.F.), No. 29; Venezuela (D.F.), No. 59; New York City and San Francisco, No. 54; and Stockholm, No. 46.

Wherever possible, public works and other public expenditure have been allocated by function; for example, school construction is included with education.

Includes housing as well as public health.

Includes drainage and building maintenance as well as roads.

Most of this expenditure is for parks and recreation.

Includes Education and Public Welfare Bureau.

Includes pensions, payments in rice, and some other fringe benefits of public servants.

Social welfare is included with public health.

Includes Han River construction, which accounts for more than 25 per cent of total public works.

“Defense and security” is a state expenditure.

Less than 0.05 per cent.

All the Colombian city percentages are much affected by the fact that expenditure by public enterprises for water, electricity, and telephones is excluded here.

Since previsión social is supplementary payments to present and past employees, it is included in “General administration.”

Valorization expenditure.

Primarily hospital expenditure but includes some other social welfare.

Includes development, public works, and rectifications.

Includes cemeteries, gardens and plazas, and the official press.

Includes other “utility” expenditure.

Includes heating plants and gasworks.

Includes ports.

Real estate administration accounts for much of this expenditure.

Function
Debt chargesCourts, police,

and public

safety
FireWelfareHousingTransport and

communications
Other

expenditure
Other capital

expenditure1
TotalCity
Africa
7.11.0100.0Dar es Salaam
6.612.0100.0Freetown
13.2100.0Lagos
1.215.67.94100.0Nairobi
Far East
1.94.4100.0Bangkok
1.20.731.5100.0Djakarta
5.72.914.427.3100.0Semarang
1.21.96.239.8100.0Solo
16.23.92.411.0100.0Manila
5.013.2100.0Saigon
0.41.57.41.820.2100.0Seoul
24.80.42.34.522.3100.0Singapore
South Asia
23.03.00.5100.0Ahmedabad
21.311.21.6100.0Bombay
3.95.216.1100.0Calcutta
2.325.59.2100.0Hyderabad
4.119.814.8100.0Madras
0.21.68.51.95.5100.0Colombo
Latin America
18.60.444.3100.0Belo Horizonte
14.645.9100.0Recife
0.2911.611.18.6100.0Rio de Janeiro
106.212.155.4100.0São Paulo
3.58.935.09.0100.0Bogotá (D.E.) 11
4.010.615.1100.0Cali 11
6.125.728.9100.0Medellín 11
3.712.3100.0Mexico (D.F.)
21.210.816100.0Venezuela (D.F.)
Developed countries
3.05.82.521.427.417100.0New York City
1.13.53.111.549.517100.0San Francisco
10.40.513.84.31915.220100.0Stockholm
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Dar es Salaam, see No. 47; Freetown, No. 43; Lagos, No. 31; Nairobi, No. 21; Bangkok, No. 50; Djakarta, Semarang, and Solo, No. 20; Manila, No. 39, p. 624; Saigon, No. 60; Seoul, No. 24; Singapore, No. 44; Ahmedabad, Bombay, Calcutta, Hyderabad, and Madras, No. 17; Colombo, No. 45; Belo Horizonte, Recife, Rio de Janeiro, and São Paulo, No. 7; Bogotá (D.E.), Cali, and Medellín, No. 13; Mexico (D.F.), No. 29; Venezuela (D.F.), No. 59; New York City and San Francisco, No. 54; and Stockholm, No. 46.

Wherever possible, public works and other public expenditure have been allocated by function; for example, school construction is included with education.

Includes housing as well as public health.

Includes drainage and building maintenance as well as roads.

Most of this expenditure is for parks and recreation.

Includes Education and Public Welfare Bureau.

Includes pensions, payments in rice, and some other fringe benefits of public servants.

Social welfare is included with public health.

Includes Han River construction, which accounts for more than 25 per cent of total public works.

“Defense and security” is a state expenditure.

Less than 0.05 per cent.

All the Colombian city percentages are much affected by the fact that expenditure by public enterprises for water, electricity, and telephones is excluded here.

Since previsión social is supplementary payments to present and past employees, it is included in “General administration.”

Valorization expenditure.

Primarily hospital expenditure but includes some other social welfare.

Includes development, public works, and rectifications.

Includes cemeteries, gardens and plazas, and the official press.

Includes other “utility” expenditure.

Includes heating plants and gasworks.

Includes ports.

Real estate administration accounts for much of this expenditure.

Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Dar es Salaam, see No. 47; Freetown, No. 43; Lagos, No. 31; Nairobi, No. 21; Bangkok, No. 50; Djakarta, Semarang, and Solo, No. 20; Manila, No. 39, p. 624; Saigon, No. 60; Seoul, No. 24; Singapore, No. 44; Ahmedabad, Bombay, Calcutta, Hyderabad, and Madras, No. 17; Colombo, No. 45; Belo Horizonte, Recife, Rio de Janeiro, and São Paulo, No. 7; Bogotá (D.E.), Cali, and Medellín, No. 13; Mexico (D.F.), No. 29; Venezuela (D.F.), No. 59; New York City and San Francisco, No. 54; and Stockholm, No. 46.

Wherever possible, public works and other public expenditure have been allocated by function; for example, school construction is included with education.

Includes housing as well as public health.

Includes drainage and building maintenance as well as roads.

Most of this expenditure is for parks and recreation.

Includes Education and Public Welfare Bureau.

Includes pensions, payments in rice, and some other fringe benefits of public servants.

Social welfare is included with public health.

Includes Han River construction, which accounts for more than 25 per cent of total public works.

“Defense and security” is a state expenditure.

Less than 0.05 per cent.

All the Colombian city percentages are much affected by the fact that expenditure by public enterprises for water, electricity, and telephones is excluded here.

Since previsión social is supplementary payments to present and past employees, it is included in “General administration.”

Valorization expenditure.

Primarily hospital expenditure but includes some other social welfare.

Includes development, public works, and rectifications.

Includes cemeteries, gardens and plazas, and the official press.

Includes other “utility” expenditure.

Includes heating plants and gasworks.

Includes ports.

Real estate administration accounts for much of this expenditure.

Expenditure on general administration is inevitably a large part of the cost of municipal government. Some evidence exists to support the proposition that administrative expenditure as a share of the total may vary inversely with the wealth of the city. This expenditure is subject to significant economies of scale, since, even in the poorest large city, an administrative system must be able to provide a minimum of health, sanitary, education, and maintenance services. Nevertheless, some of the figures are clearly not explained by this argument—for example, the relatively large share spent on general administration in Mexico City. Some of the differences among cities undoubtedly are due to improved classification schemes in some countries. In some cities the administration is more efficient and also overhead expenditure is likely to be more fully allocated to the various city functions.

Given these qualifications, evidence appears to support the proposition. Perhaps most obvious is the difference between the share spent on general administration in the 27 developing cities (20.8 per cent) and in the 3 cities in the developed world (2.3 per cent). Within geographical areas there also appears to be an inverse relationship between municipal per capita expenditure and the share of total expenditure used for general administration. For example, Seoul and Singapore have by far the highest per capita expenditure of the Far Eastern cities and spend a significantly smaller share for general administration. The same is true for Bombay among the Indian cities, Nairobi in Africa, and Sao Paulo in Brazil. 36 Expenditure for general administrative purposes as a percentage of total expenditure appears relatively comparable among the Colombian cities and also among the three high-income cities. Evidence appears to lend some support to the position that the poorer the city, the more closely it resembles only an employment service: that is, the percentage of municipal resources devoted to shuffling papers is large, while the percentage devoted to providing services to the public (including shuffling papers) is relatively small.

Revenues devoted to health services, as a share of total municipal resources, also vary tremendously among the municipalities of the world. Bangkok, Saigon, and the Indonesian cities all have the responsibility to provide clinics, dispensaries, and medical services. 37 In Bangkok and Saigon, where per capita expenditure is much higher than in Djakarta, the percentage of total municipal expenditure allocated for public health purposes is also much larger. South Asian variations in health expenditure as a percentage of total expenditure are relatively small, although the share in Ahmedabad is more than double that in Madras. In the three African cities of Dar es Salaam, Lagos, and Nairobi, public health accounts for between 20 and 30 per cent of total expenditure. In a fourth African city—Freetown, Sierra Leone 38—the Central Government has the responsibility for both public health and hospitals, whereas the other three cities provide clinics for maternity and child care, dispensaries, and other health centers.39

Between 3 and 5 per cent of total city expenditure in Brazilian municipalities is allocated to hospitals, medical centers, and dispensaries. Although the municipalities are entirely free to function in this area, the states and the Central Government are also active in the public health field. 40 In Colombia in 1969 the municipalities accounted for only about 2 per cent of total public health expenditure; the other 98 per cent was accounted for by collection of gambling taxes by departmental welfare agencies (15 per cent), the National Government (74 per cent), and departmental governments and miscellaneous sources (9 per cent). 41 In Colombia, central government grants for health go directly to the institutions providing the local health services and are not funneled through the municipal governments. 42 Expenditure by the Federal District of Venezuela on public health is a sizable 26 per cent of the total expenditure by the District, and nearly all of it is used to support hospitals. In Stockholm, New York City, and San Francisco, the responsibility for public health is shared by the city and higher levels of government. The share of total expenditure allocated to public health is 18.7 per cent in Stockholm, 10.0 per cent in New York City, and 6.3 per cent in San Francisco.

Expenditure for education ranges from one sixth to one fourth of total municipal expenditure in Bangkok, Djakarta, Seoul, and Saigon. Although the two smaller Indonesian cities of Semarang and Solo retain the responsibility for primary schooling, expenditure for education remains relatively unimportant, accounting for 6 per cent or less of their total expenditure. The city of Colombo in Sri Lanka spends nothing on education, since it is the responsibility of the Central Government. In India, expenditure for education by the municipal corporations varies from 0.5 per cent of total city expenditure in Hyderabad to 20.1 per cent in Ahmedabad; these variations may be partially due to the fact that city responsibilities for education are allocated by the states and may differ among them. It is evident that in the poorer municipalities, owing to lack of resources, either the burden of financing education rests with the states or education is not provided. 43 Evidence suggests that the low priority placed on education by the poorer municipalities is not compensated for by higher per capita expenditure by the state. In 1967-68 per capita state expenditure on education was lowest in Andhra Pradesh (Hyderabad) and West Bengal (Calcutta), and highest in Tamil Nadu (Madras) and Maharashtra (Bombay).44

In the African cities, the relative importance of expenditure on education appears fairly comparable, accounting for between 20 and 30 per cent of total city expenditure. The means of financing the expenditure vary considerably, however. Expenditure on primary schools in Nairobi is financed primarily by personal tax revenues and tuition fees. 45 In the towns and cities of Tanzania, a central government subvention accounted for about 85 per cent of current education expenditure. 46 In Lagos, 70 per cent of recurrent expenditure and 100 per cent of capital expenditure were met by grants; 47 in Freetown, the Central Government covered 100 per cent of the salaries of primary teachers and 60 per cent of capital expenditure.48

Municipalities in Colombia spend very little (as for public health) on education, which accounted for somewhat less than 5 per cent of their total expenditure in 1967. For the total educational expenditure by public authorities in 1966, municipalities accounted for only 6 per cent, departments for 37 per cent, decentralized public agencies for 19 per cent, and the Central Government for 38 per cent. 49 Educational responsibilities in Brazil are divided between the states and municipalities—the particular division varying from state to state. Generally the responsibility for primary education is shared by the municipality and the state, with private and state schools providing secondary education. 50 In 1968, municipalities accounted for 11 per cent of government expenditure on education; the state and federal governments accounted for 62 per cent and 28 per cent, respectively. 51 In Brazil, as in India, there appears to be some indication that the poorer the municipality, the smaller the percentage of total expenditure that is allocated to education. Recife, in Pernambuco (the poorest of the four Brazilian states considered), allocates 5.1 per cent of its total revenue to education, while the state allocates 3.0 per cent of its total; Belo Horizonte (in Minas Gerais) allocates 10.2 per cent, and the state allocates 16.4 per cent. In São Paulo, the municipality allocates 10.6 per cent and the state, 16.1 per cent. When resources are scarce, it appears that expenditure on education is severely affected.

It is difficult to generalize about other categories of expenditure; expenditure on public works deserves mention, however, as its importance has been emphasized in earlier sections. For some of the cities, outlays for public works are allocated by function in Table 6—as in Freetown, Lagos, and Nairobi. In other cities, at least some of the expenditure on these items has been included in public works, such as sidewalks and roads in Bangkok, and roads, sewers, sidewalks, parks, etc., in the Colombian cities. However classified, most cities in developing countries have had to allocate a significant part of their budget to capital projects. Such expenditure ranged from one third to one half of total expenditure in Bangkok, Mexico City, and Cali. In other cities, capital outlays may have been equally important, but because of the allocation of expenditure by function they are not so readily apparent.

Finally, Table 6 indicates that expenditure on welfare was significant only in Colombo, the Brazilian cities, and the cities in the developed countries.

per capita expenditure comparisons for 40 cities

Table 7 indicates that among the Far Eastern cities, those in Indonesia are lagging in per capita revenues and expenditure. In Djakarta, per capita revenues in U.S. dollars are only slightly in excess of half those in Bangkok and are almost one third those in Manila. The same is true for per capita expenditure. To the extent that Djakarta can provide the same services at a fraction of the cost, this type of comparison is not very useful. For example, salaries of school teachers and health personnel will depend in part upon per capita income and general living standards in the country; however, some of the services (such as the construction of main thoroughfares and water distribution systems provided by cities) require modern construction materials and equipment that may have comparable costs in terms of U.S. dollars in these cities. Seoul’s per capita expenditure in terms of U.S. dollars is, with the exception of Singapore, more than two and a half times that of any of the other Far Eastern or South Asian cities. Singapore is a city-state, has no legal restrictions imposed by higher levels of government on its revenue sources, and has full responsibility for all public services. This situation coupled with a relatively high per capita income level helps to explain why Singapore’s per capita expenditure in U.S. dollars is nearly four times that of Seoul.

Table 7Selected Cities: Per Capita Expenditure and Revenues in Terms of Local Currencies and U.S. Dollars, Selected Years
CityYearPer Capita

Expenditure
Per Capita RevenuesPer Capita Revenues

Net of Grants 1
Exchange

Rate Used

(currency

units per

U.S. dollar)
Per Capita

Income

(U.S. dollars)
Own

currency
U.S.

dollars
Own

currency
U.S.

dollars
Own

currency
U.S.

dollars
Africa
Blantyre1965/662.908.123.188.913.148.790.35711352
Casablanca1968111.421.93116.522.935.081882
Dar es Salaam19654.3611.744.2911.553.659.830.3715662
Fort Lamy19704,042.214.563,082.811.101,427.65.14277.71632
Freetown1963/645.4815.355.4815.353.6610.250.35711382
Lagos1962/635.0614.174.0311.292.426.780.3571702
Lusaka196711.0731.009.1225.540.35712452
Nairobi196515.3542.9915.3442.9615.3442.960.35711112
Far East
Bangkok1968211.910.16220.110.56177.58.5120.851462
Djakarta1970/711,761.04.702,153.05.741,688.04.50375.00892
Makassar1969/70768.22.05494.51.32375.00872
Semarang1969/70726.61.94685.11.83555.21.48375.00872
Solo1969/70750.22.00707.21.89494.81.32375.00872
Manila196465.016.5464.216.343.932452
Saigon1969878.77.451,368.011.601,368.011.60118.00
Seoul19709,469.329.979,866.931.228,757.527.71316.001792
Singapore1969355.5115.05442.3143.143.097792
South Asia
Ahmedabad1967/6863.38.3954.27.1850.46.687.547963
Bombay1970103.213.6589.411.8389.911.897.5591293
Calcutta1967/6832.64.3230.54.0426.33.487.5471133
Hyderabad1966/6719.62.5920.52.7120.32.687.576703
Madras1967/6849.06.4944.55.9043.85.807.547813
Colombo196864.710.8972.812.2655.39.315.9401472
Latin America
Belo Horizonte196844.311.5755.614.5229.77.753.831793
Recife196841.010.7044.111.5113.33.473.831483
Rio de Janeiro1968315.382.32302.879.06288.175.223.835703
São Paulo1968145.938.08145.938.0977.520.233.834343
Bogotá (D.E)1969504.928.16504.928.16434.024.2117.934743
Cali1969238.113.28205.211.44199.611.1317.933813
Medellín1969463.025.82435.724.30430.524.0117.933013
Dominican Republic (D.N.)196615.1715.1714.6714.676.446.441.002582
Guatemala City1962/638.768.768.768.761.002612
Mexico (D.F.)1963293.623.51338.127.07338.127.0712.494142
Montevideo1958122.712.03122.712.03122.712.0310.206532
Panama City19689.519.518.428.428.328.321.005672
Valencia196460.113.3652.411.644.508992
Venezuela (D.F.)1969216.148.02207.446.09109.124.244.509482
Developed countries
New York City1969888.5888.50828.8828.80458.5458.501.003,5212
San Francisco19691,090.51,090.50965.9965.90590.5590.501.003,5212
Stockholm19687,205.71,391.067,032.71,357.666,185.51,194.115,182,5712
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Blantyre, see No. 25; for Casablanca, No. 30; Dar es Salaam, No. 47; FortLamy, No. 10; Freetown, No. 43; Lagos, No. 31; Lusaka, No. 62; Nairobi, No. 21; Bangkok, No. 50; Djakarta, Makassar, Semarang, and Solo, No. 20; Manila, No. 40; Saigon, No. 60; Seoul, No. 24; Singapore, No. 44; Ahmedabad, Calcutta, Hyderabad, and Madras, No. 17; Bombay, No. 16; Colombo, No. 45; Belo Horizonte, Recife, Rio de Janeiro, and São Paulo, No. 7; Bogotá (D.E.), Cali, and Medellín, No. 13; Dominican Republic (D.N.), No. 14; Guatemala City, No. 15; Mexico (D.F.), No. 29; Montevideo, No. 57; Panama City, No. 36; Valencia, No. 58; Venezuela (D.F.), No. 59; New York City and San Francisco, No. 54; and Stockholm, No. 46. Per capita income figures most often were taken from No. 4. Indices to adjust for state or departmental per capita incomes in India were calculated from data given in No. 18; in Brazil they were calculated from No. 6; for Colombia the indices were taken from No. 11.

Cities with a share of certain taxes collected by the central government include Bangkong, Colombo, Djakarta, Manila, Mexico (D.F.), and Seoul.

Per capita income for the country.

Per capita income for the state or department in which the city is located.

Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Blantyre, see No. 25; for Casablanca, No. 30; Dar es Salaam, No. 47; FortLamy, No. 10; Freetown, No. 43; Lagos, No. 31; Lusaka, No. 62; Nairobi, No. 21; Bangkok, No. 50; Djakarta, Makassar, Semarang, and Solo, No. 20; Manila, No. 40; Saigon, No. 60; Seoul, No. 24; Singapore, No. 44; Ahmedabad, Calcutta, Hyderabad, and Madras, No. 17; Bombay, No. 16; Colombo, No. 45; Belo Horizonte, Recife, Rio de Janeiro, and São Paulo, No. 7; Bogotá (D.E.), Cali, and Medellín, No. 13; Dominican Republic (D.N.), No. 14; Guatemala City, No. 15; Mexico (D.F.), No. 29; Montevideo, No. 57; Panama City, No. 36; Valencia, No. 58; Venezuela (D.F.), No. 59; New York City and San Francisco, No. 54; and Stockholm, No. 46. Per capita income figures most often were taken from No. 4. Indices to adjust for state or departmental per capita incomes in India were calculated from data given in No. 18; in Brazil they were calculated from No. 6; for Colombia the indices were taken from No. 11.

Cities with a share of certain taxes collected by the central government include Bangkong, Colombo, Djakarta, Manila, Mexico (D.F.), and Seoul.

Per capita income for the country.

Per capita income for the state or department in which the city is located.

A large variation also exists in the per capita revenues and expenditure of the African cities. Per capita expenditure in Nairobi is five times that in Blantyre, three and a half times that in Dar es Salaam, three times that in Lagos, and more than double that in Freetown, keeping in mind that the years are not the same for all cities. Although separate expenditure data for Lusaka were not obtained, it appears from the revenue figures that per capita expenditure in Nairobi was about one and a half times as great as that in Lusaka.

Perhaps of greater interest is the disparity in urban expenditure within countries. Variations in responsibilities and costs, although sometimes still important, will be less than in the intercountry comparisons, and therefore the intracountry comparisons should provide a more accurate picture of differences in the services provided. For five Indian municipal corporations examined, per capita expenditure ranged from $2.59 (Hyderabad, 1966-67) to $10.43 (Bombay, 1967-68).52 Per capita expenditure by the other three corporations for 1967-68 had a wide range: $4.32 (Calcutta), $6.49 (Madras), and $8.39 (Ahmedabad). The differences do not appear to be compensated for by per capita expenditure made by the state governments. Indeed, although the per capita state expenditure varies considerably less than the per capita municipal expenditure, that of Maharashtra (Bombay) is highest ($14.32) and that of Andhra Pradesh (Hyderabad) is lowest ($8.88).

The variation is considerably less among the four Indonesian cities examined. In this case, since Djakarta is both a city and a province for budget purposes, the other cities need to be combined with their provinces in order to meaningfully compare per capita levels of expenditure. When this is done for 1969-70, per capita expenditure ranges from $4.58 in Djakarta to $2.75 in Makassar, $2.51 in Semarang, and $2.38 in Solo—that in Djakarta being less than twice that in the other Indonesian cities examined.53

Differences in per capita expenditure are large in both the Brazilian and the Colombian cities. In Brazil, São Paulo’s per capita expenditure ($38.09) is more than triple that of Belo Horizonte ($11.57) and Recife ($10.70). Rio de Janeiro, since it has a dual status as the State of Guanabara, can be compared with the others only if the per capita expenditure of respective state governments is added to that of the three cities just named. This results in the following per capita expenditure: São Paulo, $112.11; Rio de Janeiro, $82.32; Belo Horizonte, $30.44; and Recife, $26.55. 54 In Colombia, per capita expenditure by the municipal governments in Bogotá and Medellín is roughly similar, with Cali’s expenditure being about half that of the other two. Again, per capita expenditure by the departmental governments does not make up for these differences.

Some of the variation in per capita city expenditure within countries, as well as among countries, is undoubtedly due to variations in per capita income. For example, although supporting data are not available, average per capita income is probably somewhat higher in Djakarta than in the other Indonesian cities examined here. Per capita figures for the states of India indicate that of the states in which the five Indian cities examined are located per capita income in Maharashtra (Bombay) is the highest and in Andhra Pradesh (Hyderabad), the lowest. Likewise, per capita income in São Paulo (state) and Guanabara (Rio de Janeiro) is about three times that in Pernambuco (Recife) and Minas Gerais (Belo Horizonte). Data for the African cities are lacking, but it seems clear that the per capita incomes in Nairobi and Lusaka significantly exceed those in the other African cities and may be partly responsible for the higher level of expenditure in these two cities.

Finally, it may be of some interest to note that for two cities similar in size, Djakarta (1971 population of 4,576,009) and Rio de Janeiro (1968 population of 4,132,000), per capita expenditure by local governments was $4.70 and $82.32, respectively. Tremendous variation exists among the cities of the world, and it is beyond the scope of this study to explain the factors leading to the differences in the numbers presented in the preceding tables.

patterns of responsibility—theory and fact

Several students of local government have stressed the tendency of former colonies to follow the local government systems of the countries that previously controlled them. 55 This is said to hold relatively true for the division of governmental responsibilities and government revenue sources. Such generalizations may prove useful for some purposes, but for comparing the systems of municipal finance in major cities of developing nations, as is done in this paper, they may prove to be seriously misleading. For example, Froomkin stated that “the British and the French models are characterized by a close reliance of municipalities on central government subsidies for the operation of current services.” 56 How close is “close” may be debatable, but the major Indian cities examined have received on average less than 5 per cent of their revenues from higher levels of government; and in Sri Lanka (which is closer than India to the British system), Colombo receives one fourth of its income as grants. A high degree of dependence on national assistance is not common among the large African cities in the former British colonies. Blantyre and Nairobi received almost no aid, and for both Dar es Salaam and Lusaka, central government aid accounted for less than 20 per cent of total revenues. In comparison, grants from the Central Government accounted for one fourth of local authority revenues in London in 1967 and for 52 per cent of revenues for all local authorities in Great Britain. 57 Generalizations concerning similarities between the French and Latin American systems are dangerous, on the same grounds. Whereas the reliance of large cities on funds from higher levels of government is great in Brazil, the Dominican Republic, and the Federal District of Venezuela, it is low for the larger Colombian cities and for Guatemala City, Mexico City, and Panama City.

A second generalization that requires careful examination is that “underdeveloped countries… under the British influence have succeeded in building up their municipal services to a higher level.” 58 It is difficult to prove or disprove this statement, as it is clear that municipal services depend on factors other than the governmental system. It is also true that municipal services may be provided by different levels of government. Nevertheless, using the limited evidence mustered in this study, a comparison of per capita expenditure in Table 7 by the municipality of Fort Lamy, Chad (French) of $14.56 with expenditure by Freetown ($15.35), Lagos ($14.17), Dar es Salaam ($11.74), and Blantyre ($8.12) yields something less than clear support for the above-mentioned proposition. Although slightly smaller than in Freetown, per capita expenditure by Fort Lamy, made possible by large grants, is greater than that of Lagos, Dar es Salaam, and Blantyre. In Lusaka and Nairobi, where per capita expenditure is much greater than in Fort Lamy, per capita incomes are also greater.

Froomkin went so far as to propose the normative model in Table 8 for the responsibilities that should be allocated to municipalities in less developed countries. To begin with, it seems unlikely that one such model would serve equally well countries with differing governmental systems (Froomkin makes much of these differences). The model clearly suggests a high degree of centralization in a developing country. It proposes the elimination of municipal responsibility (where it exists) for police, schools, welfare, and corrections. Minor services of fire protection, libraries, recreation, and sanitation would be left with the municipalities, and health, highways, and hospitals would be shared with higher levels of government. The desirability of the centralizing tendencies of this proposal is, of course, subject to the reservations discussed in an earlier section.

Table 8Froomkin’s Model for Allocating Responsibilities to Municipalities for Current Services in Developing Countries
ServiceGovernmental

Responsibility
General controlMunicipality
PoliceOther
FireMunicipality
SchoolsOther
LibrariesMunicipality
RecreationMunicipality
WelfareOther
CorrectionsOther
HealthShared
HighwaysShared
SanitationMunicipality
HospitalsShared
Source: See Sources Used in Preparing Tables, in the Appendix, No. 2.
Source: See Sources Used in Preparing Tables, in the Appendix, No. 2.

In contrast to Froomkin’s model, it may be of some interest to examine the actual allocation of expenditure responsibilities to municipalities. This may not indicate the most desirable division of responsibilities, but it does indicate what has proved feasible. In most developing countries, municipalities have retained some responsibility for the provision of education, although generally financial aid is received from other levels of government. The Indian municipalities of Bombay and Calcutta both have primary responsibility for elementary education, while the states are responsible for secondary education. Primary responsibility for elementary education also rests with local authorities in Egypt, Kenya, Malawi, Nigeria, the Sudan, Israel, the Republic of China, Indonesia, Korea, Thailand, and Brazil. In most of these countries, significant amounts of central government aid are provided. In Sierra Leone, the Philippines, Colombia, Panama, and Venezuela the responsibility for education is shared by municipal and higher levels of government. 59 In the Philippines, education is provided by the Central Government up through the fourth grade; after that it is left to local authorities. Zambia, Sri Lanka, Mexico, Peru, and Greece are among the countries in which education remains the responsibility of the central government. In Argentina, primary education is the responsibility of the state governments.

Municipal governments in several countries—notably Nigeria, the Sudan, India, Mexico, Peru, and Venezuela—have partially financed their police services, while in other countries—Argentina, Brazil, and Colombia—higher levels of government have provided police services as Froomkin’s model recommends. Fire protection is a shared responsibility in Freetown, Sierra Leone, and a state responsibility in Argentina; public welfare is a shared responsibility in Nigeria, Zambia, Israel, Thailand, Peru, and Venezuela—all deviations from Froomkin’s model. Libraries and recreation seem generally to be municipal responsibilities (but frequently with aid from higher levels of government), as in Nigeria, Sierra Leone, the Sudan, Israel, Sri Lanka, Mexico, and Peru. The responsibility for providing highways and streets, public health facilities, and hospitals is in most countries shared by the various levels of government, and is therefore generally in accord with the Froomkin model.

Froomkin recommended that schools be the responsibility of higher levels of government, while libraries and recreation should be the responsibility of the local government. Evidence suggests that a strong case can frequently be made for at least partial local financing of schools. Education is valued by the local population; if they can see their taxes going toward this end, they may be more willing to pay them. In addition, the close connection that exists in many countries among schools, recreation, and libraries would seem to suggest that these three responsibilities rest with the same level of government. Schools are an important expenditure, and the need for city revenues would differ significantly depending on whether or not the cities are responsible for education.

Flexibility must be expected and encouraged in the development of local systems of government. Where federalism and/or local autonomy have traditionally been strong (Nigeria, India, Brazil), there may be good reason to place responsibilities and the means of finance in the hands of lower levels of government. Poor communications and resentment of the central government may also support the decentralization of governmental authority (Indonesia, the Philippines). However, the delegation of additional authority to lower government levels must be done at the appropriate moment, and not at a time when the local government authority may be used to cause internal disruptions. In still other countries, Froomkin’s model may offer the best allocation of responsibility. Although some problems and goals may be common to all developing countries, it does not follow that any particular allocation of responsibilities to municipal governments is appropriate for even a majority of these countries.

Evidence presented in this section highlights the variations existing in the division of responsibilities among governmental levels in different countries. Factors such as the division of constitutional authority, the distribution of administrative capabilities among government levels, the degree of urbanization, and the traditional roles in society all have some bearing on the optimal division of responsibilities. 60 The fact that variations in the division of responsibilities are so great among countries rightly brings into serious question the usefulness of a rigid model.

IV. City Revenues

revenue patterns

Table 9 depicts the basic patterns existing in municipal revenues around the world. Uniformity is clearly lacking.

Table 9Selected Cities: Sources of City Revenues, Selected Years(In per cent)
CityYearTaxes on Property

(1)
Sales and Excise Taxes

(2)
Vehicle Taxes

(3)
Entertainment and Gambling Taxes

(4)
Other Taxes 1

(5)
Total Taxes

(6)
Licenses, Fees, and Charges

(7)
Income from Interest, Investments, Rentals, and Sale of Property

(8)
Other Income 2

(9)
Total Nontax Revenue

(10)
Shared Taxes and Grants

(11)
Total (6), (10),

and (11)
Africa
Blantyre196660.40.360.728.010.238.21.1100.0
Dar es Salaam196536.94.141.044.144.114.9100.0
Fort Lamy19705.48.91.24.04.524.06.515.622.153.7100.0
Freetown196452.11.453.58.82.62.313.732.8100.0
Lagos196350.950.99.19.140.0100.0
Lusaka196750.05.255.227.327.317.6100.0
Nairobi196528.729.157.839.42.842.20.0100.0
Far East
Bangkong1968 315.039.1 36.0 35.0 310.275.33.00.12.15.219.5 3100.0
Djakarta1971 42.026.127.14.259.41.610.512.128.5100.0
Makassar1970 46.016.86.028.833.02.735.735.5100.0
Semarang1970 412.90.96.336.756.813.011.224.219.1100.0
Solo1970 48.70.33.412.024.49.535.745.230.6100.0
Manila196421.710.734.127.112.439.526.4100.0
Saigon1969 59.440.38.157.88.112.721.442.20.0100.0
Seoul197015.2 66.65.52.830.147.37.154.415.3100.0
Singapore196911.124.15.02.030.1 772.318.59.227.70.0100.0
South Asia
Ahmedabad196827.947.92.778.514.56.9100.0
Bombay197045.025.72.20.273.125.91.00.026.90.0100.0
Calcutta196865.00.46.371.714.413.9100.0
Hyderabad196732.02.829.163.935.11.0100.0
Madras196840.032.972.925.61.5100.0
Colombo196861.061.04.38.112.426.7100.0
Central America
Dominican Republic (D.N.)19660.70.61.334.18.542.656.0100.0
Guatemala City196320.711.33.518.554.038.27.846.00.0100.0
Mexico (D.F.)196326.232.482.39.570.415.414.229.60.0100.0
South America
Belo Horizonte196822.26.50.228.99.115.624.746.5100.0
Recife19689.95.60.215.75.30.28.914.469.9100.0
Rio de Janeiro19680.770.410.781.85.14.14.113.34.8100.0
São Paulo196817.210.90.428.59.90.514.224.646.9100.0
Bogotá (D.E.) 9196930.28.51.71.61.843.838.63.442.014.0100.0
Cali 10196967.212.01.62.61.384.78.60.33.612.52.7100.0
Medellín 11196935.113.70.01.11.351.222.84.720.247.71.2100.0
Montevideo195860.210.67.177.95.316.822.10.0100.0
Valencia196413.028.86.92.57.758.921.47.128.512.8100.0
Venezuela (D.F.) 12196917.72.61.021.30.30.530.531.347.4100.0
Developed countries
New York City196925.79.48.343.49.62.412.044.7100.0
San Francisco196938.44.80.443.610.07.417.438.9100.0
Stockholm196838.338.333.416.349.712.0100.0
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Blantyre, see No. 25; for Dar es Salaam, No. 47; Fort Lamy, No. 10; Freetown, No. 43; Lagos, No. 31; Lusaka, No. 62; Nairobi, No. 21; Bangkok, No. 50; Djakarta, Makassar, Semarang, and Solo, No. 20; Manila, Nos. 39 and 40; Saigon, No. 60; Seoul, No. 24; Singapore, No. 44; Ahmedabad, Calcutta, Hyderabad, and Madras, No 17; Bombay, Nos. 16 and 17; Colombo, No. 45; Dominican Republic (D.N.), No. 14; Guatemala City, No. 15; Mexico (D.F.), No. 29; Belo Horizonte, Recife, Rio de Janeiro, and São Paulo, No. 7; Bogotá (D.E.), Cali, and Medellín, No. 13; Montevideo, No. 57; Valencia, No. 58; Venezuela (D.F.), No. 59; New York City and San Francisco, No. 54; and Stockholm, No. 46.

“Other Taxes” may include property taxes, etc., to the extent that these are not already specified in the sources cited.

“Other Income” may include licenses, fees, charges, etc., to the extent that these are not already specified in the sources cited.

About 69 per cent of Bangkok’s taxes are shared taxes. Since the types of tax shown are specified in the available data, they are included here according to type. For Thailand, loans are lumped with grants.

The figures for Indonesian cities are budget, rather than actual, figures.

The Vietnamese Government does make grants to municipalities; however, in recent years none has been made to Saigon Prefecture. Column (8) includes revenues from excises as well as from renting.

Taxes on property include the acquisition tax, the planning tax, the fire service tax, and the property tax.

In Singapore, 21.1 per cent of revenues comes from taxes on income, thus accounting for more than two thirds of “Other Taxes.”

In Mexico (D.F.), the automobile tax is collected only in even-numbered years.

Includes 17.2 per cent for valorization.

Includes 43.6 per cent for valorization.

Includes 17.7 per cent for valorization.

Venezuela (D.F.) does have a property tax, which accounted for 8.7 per cent of total revenue in 1958. The 1969 data that were used included real estate taxes in “Other Income.”

Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each city. For Blantyre, see No. 25; for Dar es Salaam, No. 47; Fort Lamy, No. 10; Freetown, No. 43; Lagos, No. 31; Lusaka, No. 62; Nairobi, No. 21; Bangkok, No. 50; Djakarta, Makassar, Semarang, and Solo, No. 20; Manila, Nos. 39 and 40; Saigon, No. 60; Seoul, No. 24; Singapore, No. 44; Ahmedabad, Calcutta, Hyderabad, and Madras, No 17; Bombay, Nos. 16 and 17; Colombo, No. 45; Dominican Republic (D.N.), No. 14; Guatemala City, No. 15; Mexico (D.F.), No. 29; Belo Horizonte, Recife, Rio de Janeiro, and São Paulo, No. 7; Bogotá (D.E.), Cali, and Medellín, No. 13; Montevideo, No. 57; Valencia, No. 58; Venezuela (D.F.), No. 59; New York City and San Francisco, No. 54; and Stockholm, No. 46.

“Other Taxes” may include property taxes, etc., to the extent that these are not already specified in the sources cited.

“Other Income” may include licenses, fees, charges, etc., to the extent that these are not already specified in the sources cited.

About 69 per cent of Bangkok’s taxes are shared taxes. Since the types of tax shown are specified in the available data, they are included here according to type. For Thailand, loans are lumped with grants.

The figures for Indonesian cities are budget, rather than actual, figures.

The Vietnamese Government does make grants to municipalities; however, in recent years none has been made to Saigon Prefecture. Column (8) includes revenues from excises as well as from renting.

Taxes on property include the acquisition tax, the planning tax, the fire service tax, and the property tax.

In Singapore, 21.1 per cent of revenues comes from taxes on income, thus accounting for more than two thirds of “Other Taxes.”

In Mexico (D.F.), the automobile tax is collected only in even-numbered years.

Includes 17.2 per cent for valorization.

Includes 43.6 per cent for valorization.

Includes 17.7 per cent for valorization.

Venezuela (D.F.) does have a property tax, which accounted for 8.7 per cent of total revenue in 1958. The 1969 data that were used included real estate taxes in “Other Income.”

Although one British local tax expert has cited “dependence upon a single local tax, the rate,” [on real property] as “one of the more dubious English legacies conferred upon territories formerly subject to English influence,” 61 these territories do have at least one significant source of local tax revenue. The same cannot be said for all cities of the world. In none of the Asian or African cities that came under British influence (with the exception of the special case of Singapore) do taxes on property account for less than 27 per cent of total revenue (see Table 9). The average reliance on property taxes by the 12 South Asian and African cities plus Singapore is 43.2 per cent. This is all the more striking when it is seen that the 8 other cities of the Far East and the 10 Latin American cities examined derive, on average, only 11.4 per cent and 26.4 per cent (valorization included), respectively, of total revenues from taxes on property. Fort Lamy, the one city in French-speaking Africa that is covered here, raised only 5.4 per cent of its total revenues from a property tax.

The second major tax source relied upon by cities in this sample is some type of sales tax levied on gross business sales or assets, as in Brazil; 62 on some combination of gross sales, rent, square footage, and horsepower, as in Colombia; 63 or on gross receipts, as in Bangkok. Another form of indirect tax, the octroi, is levied by Indian municipal corporations. For those cities employing a sales or excise tax, the relative importance of this source varies tremendously—from 5.6 per cent in Recife (a Brazilian city of more than 1 million population) to 47.9 per cent of total revenues from the octroi in Ahmedabad, India. Many cities, however, ignore these forms of indirect taxation. These taxes are noticeably absent in the African cities in this study, except for Fort Lamy, which raised 8.9 per cent of its revenue from a turnover tax. In most countries, including India and Brazil, the more general types of sales tax are reserved for the central or state levels of government. 64

Four of the cities included in the sample levy some form of income tax. In 1969, New York City gained about 8 per cent of its total revenues from income tax, and Stockholm derived 38 per cent from that source in 1968; however, from the developing countries, Nairobi, Lusaka, and Singapore are the only cities in the sample that use some form of income tax.65 Nairobi until recently levied a tax on income at a fixed rate, and raised as much from this source in 1965 as it did from property taxes—about 29 per cent of total revenues. Lusaka raised about 5 per cent of total revenues from a personal levy.

Two other tax sources singled out in Table 9 are “Vehicle Taxes” and “Entertainment and Gambling Taxes.” Only in rare cases are these major municipal revenue sources. Most cities derive some revenue from vehicle taxes (including some African cities where vehicle taxes are lumped with “Other Taxes”). In Djakarta, vehicle taxes account for one fourth of total city revenue. This is due partly to the insignificance of property and sales taxation as revenue sources. The importance of vehicle taxes in Djakarta is further facilitated by the fact that it has the taxing powers of a province as well as those of a city. Semarang, Solo, and Makassar cannot utilize either of the motor vehicle taxes employed by Djakarta because they do not have provincial status. Other cities that draw somewhat disproportionately large amounts from vehicle taxation are Seoul, Montevideo, and Cairo.66 In Thailand and Tanzania, revenues from motor vehicle licenses are shared with city governments, and in the Philippines and Brazil, the municipalities are given a share of central government tax revenues from motor vehicle fuels.

Gambling and entertainment taxes have been a major revenue source for Djakarta and Makassar in Indonesia. These two cities gained 27.1 per cent and 16.8 per cent, respectively, of their revenues from taxes on casinos, lotteries, and other forms of entertainment. In no other city do these taxes appear to raise more than 7 per cent of total revenues. Indonesia is one of the few countries permitting taxation of lotteries and other gambling at the local level, and it is the gambling rather than the entertainment taxes that account for the large shares that were cited.67 Along with property taxes, entertainment taxes appear to be one of the revenue sources more or less universally relied upon by local governments. Administration is simplified by the fact that there are likely to be relatively fewer taxpayers, and checks for tax avoidance can be made more easily.

Licenses, fees, and charges vary tremendously in importance among the cities examined. This is due partly to a problem of classification. For example, the price for business licenses may be based on gross sales, or on some rough approximation of profits. Whether this type of levy is classified as a sales tax or a license may vary from country to country, without the available budgetary data providing a full explanation. A second problem arises in accounting for the revenues raised in the form of fees and charges made by municipal enterprises. Three treatments of these fees are possible: (1) gross fees and charges may be included in budget revenues; (2) net revenues from public enterprises may be included; and (3) revenues from public enterprises may be treated as income from private enterprises and thereby may be excluded from the municipal budgets. In general, it appears that net revenues are included. For this reason, in the Colombian cities of Bogotá, Cali, and Medellín only the net contribution to the city budget from public enterprises has been included. Had gross fees been included, licenses, fees, and charges in these three cities would have accounted for 73.5, 74.3, and 73.0 per cent of total revenues, respectively, rather than 38.6, 8.6, and 22.8 per cent. 68

The data indicate that charges and fees for services provided by public enterprises can be substantial. In particular, an examination of the budgets of Bombay, Makassar, and Seoul indicates that net fees and charges for public services can be an important part of the total revenue picture. The importance of fees in other cities may be obscured by their inclusion in “Other Income.” The objectives of the municipal government will obviously have much to do with the net contribution of these public enterprises.

Few details can be given about the “Other Income” category because even in most of the data sources it is not broken down. Only through familiarity with individual cities could one begin to piece together the components that make up “Other Income.”

Table 9 indicates that grants from higher levels of government generally form an important part of municipal revenues in all parts of the world. This topic is discussed at length in Section V.

suggestions for city revenues

Administrative resources have frequently been thinly spread, and revenue sources normally have not been developed to anywhere near their full potential. This may be due partly to the susceptibility of local officials to pressures from potential taxpayers. If the alternative to higher local taxes is an increase in central government taxes and greater aid to the cities, the city official may find that increase preferable. The large number of revenue sources frequently drawn upon also suggests that it may be easier to find new minor sources of revenue than to allow the effective rates of some taxes to grow to noticeable levels through rate changes or improved enforcement. This path of action holds the added attraction that more civil servants are needed to generate the same amount of revenues. It is only realistic to recognize that such forces are likely to have played an important role in the evolution of existing city revenue systems. As the demand for urban services becomes more urgent, however, and as the inadequacies of higher-level governments in meeting these demands increase and are recognized, the possibility of changing and improving city revenue systems is enhanced.

This appears to be the current situation, and it is one in which it is useful to seek the most promising avenues open to the cities in the developing world. Clearly, the appropriate avenues depend on the capabilities of city administrators and on other constraints (the total environment) on the city governments. The following suggestions, of course, will not be equally appropriate for all cities. In decreasing order of potential, the following taxes appear to hold the greatest promise for city governments in developing countries: taxes on property, on sales or trading licenses, on motor vehicles, and on amusements. 69 The potential of each of these taxes has been demonstrated effectively in various cities in the developing world: for example, the property tax in Lusaka, Nairobi, and Mexico City; taxes on sales or trading licenses in Bangkok, Mexico City, and the cities in Colombia and Venezuela; motor vehicle taxes in Djakarta and Montevideo; and amusement taxes (including gambling taxes) in Djakarta. It must always be kept in mind that the arguments for a strong “city” revenue base are inexplicably linked to those for the desirability of a measurable degree of decentralization.

The potential for property taxes and their ineffective use in many parts of the world has been noticed. For example, a recent intensive study of the property tax system in Djakarta, where property taxes raised only 2 per cent of total city revenues, concluded, not surprisingly, that property taxation could and should be strengthened. 70 A 1969 study of local finance in Korea recommended approximately a tenfold increase in property taxes over a five-year period. 71 After examining the local tax system in Viet-Nam, Milton Taylor found that

At the village and provincial levels, there is one broad base of taxation—real property—which could account for all of the revenue needs of both levels of government if it were properly exploited …. every effort should be made to develop and make more effective the tax on real property. 72

Weaknesses in property taxation have not been overlooked by experts examining this tax in South and Central America, either. In Colombia, 73 Mexico, 74 Panama, 75 and Venezuela, 76 experts have emphasized the importance of more effective use by local governments of property taxes as a source of local revenue. 77 If property taxes—unlike sales, income, or motor vehicle taxes—are not used by local authorities, they are probably ignored by other levels of government.

Increased use of betterment levies may also be a way to effectively tax property owners and to finance urban improvements. The major cities in Colombia have found that significant capital improvements can be financed locally if property owners benefiting from the projects are required to pay a tax equal to some part of the increase in the value of their property resulting from improved streets, parks, or sidewalks. 78 Similar forms of tax have been used in Ecuador and the Federal District of Mexico, and special assessments have been widely used to finance improvements in other developing countries, although they have generally raised only small amounts of revenue. Betterment levies and special assessments are a means of assuring that urban revenues will be used for specific purposes. They differ from some forms of earmarked taxes, as they tend to be levied for a definite period, with the objective of recovering the cost of a specific project. They do provide useful evidence, however, that earmarking of taxes might be used effectively to finance various city services.

Sales taxes at the municipal level have received less attention from technical assistance experts than have property taxes. This is due to the complex way in which existing taxes have frequently been administered, the difficulties in improving markedly upon the existing systems, and the allocation of this form of tax to higher levels of government. In Colombia, where the business and industry tax is highly complex, it raised 27 per cent of the combined tax revenues of the four largest cities in 1967. Rather than abolish such a lucrative tax, one fiscal expert recommended a 2 per cent municipal retail sales tax, with a few items, such as food and clothing, exempt. In addition, those businesses with gross sales below a certain level could be exempted. 79 In Bombay and Caracas, where complex forms of the sales tax have also provided significant city revenues, fiscal experts have argued, among other things, for simplified rate structures rather than for eliminating existing taxes. 80 On the one hand, the relatively high income elasticity of the sales tax is a particularly attractive feature for municipal governments; on the other hand, the potential for evasion and corruption in its use in developing urban areas cannot be denied, and appears to have been a major consideration when one group of fiscal experts recommended the elimination of the tax in Panama. 81 The types of sales and business license taxes used have also been charged with contributing to regressivity and with inhibiting the efficient allocation of resources.

Fiscal experts have been less critical of the use of the vehicle tax by cities for three reasons: (1) administrative problems of vehicle taxes are considered less serious; (2) various forms of the tax are viewed as a means of improving on existing resource allocation; and (3) vehicle taxes are frequently considered to be a progressive form of taxation.

Two frequent recommendations are for higher rates and for increased uniformity of rates (particularly in small countries).82 As with property taxes, there appears to be general agreement concerning the usefulness of motor vehicle taxes as a source of city revenues. Much of this agreement may well be based on the heavy concentration of motor vehicles in the major cities in developing countries, the revenue potential of congestion charges,83 and the costs generated by the existence of these vehicles. There also appears to be increasing recognition that only through effective vehicle taxation, which may well include congestion charges, can developing cities hope to control the use of private transport and to create a viable system of public transport in the long run.84

Amusement taxes and the like generally are administrable locally and are not open to serious question on grounds of equity. Levied at relatively low rates on nonessentials, this form of taxation is perhaps least likely to be usurped by higher levels of government. If exploited effectively, it can serve as a significant and dependable revenue source.

There is a case to be made for the use of many other city taxes, as well as for the proper pricing of public services, and most cities will no doubt continue to use a wide range of revenue sources. In the long run, however, cities should probably begin to concentrate their scarce administrative resources on the collection of a small set of taxes, rather than continuing the never-ending search for new sources of revenue, which invariably yield only small amounts.

V. Some Intergovernmental Aspects of City Revenues

Left to their own devices, cities may well spend less than would be desirable from a national and/or local point of view. Several possible reasons exist for this: (1) cities may have an unmotivated, inefficient, or inadequate tax administration; (2) the richest revenue sources may be pre-empted by higher governmental levels, and cities may be left without adequate revenue to meet their needs; (3) spillovers or external economies may be generated by urban expenditure; (4) fear of tax competition from neighboring localities may inhibit the use of either some forms of tax or tax increases; or (5) cities may simply have inadequate fiscal capacity to meet their needs. In the developing countries all but the fifth reason provide strong support for the continual flow of funds from central and state governments to the major cities. The fifth argument seems unlikely, since wealth in developing countries tends to be concentrated in the major urban areas. With the possible exception of mineral-rich countries, if urban areas are too poor to finance basic public services, it seems unlikely that the remaining, poorer, part of the country is in much of a position to help.

intergovernmental transfers as a source of city revenues

As Table 9 indicates, grants from higher levels of government form an important part of city revenues in all parts of the world. Saigon and Singapore are the only cities in the Far East sample that do not rely heavily upon intergovernmental grants or shared taxes. Manila, although it received only 1.1 per cent of its revenues as grants, received more than 25 per cent of its revenue from its share of central government taxes. The six other cities of the Far East received, on average, 23 per cent of revenues from grants. 85

Grants from higher levels of government appear relatively unimportant in the Indian municipal corporations that were examined, exceeding 10 per cent of total revenues only in Calcutta and 6 per cent in Ahmedabad. In general, Indian states concentrate their grants on the rural local authorities; the urban authorities receive substantial grants in few states. The other South Asian city, Colombo, received substantial aid equaling about 24 per cent of total revenues—in the form of general grants and grants for particular public projects.

In the cities of Africa, aid varies tremendously in importance. Nairobi and Blantyre receive no aid and very little aid, respectively. This may be explained partly by the fact that Nairobi has successfully employed a revenue source not used by other cities—the personal levy; and in Blantyre, Malawi, the Central Government has assumed the major burden of educational expenditure. 86 In two cities relying heavily on central government aid, Freetown and Lagos, major expenditure on education was required of the local authorities and was financed largely by central government grants.

In Latin America, the major Colombian cities that couple a sound property tax system with the use of valorization levies receive, on average, 6 per cent of their total revenues from grants. The Federal District of Mexico and Montevideo, Uruguay, both with effective property tax systems, received no grants from their central governments. On the other hand, the Brazilian cities of Belo Horizonte, Recife, and São Paulo receive an average of 54 per cent of total revenues from grants. Rio de Janeiro as the State of Guanabara receives little in intergovernmental grants, owing to the grant allocation formulas. The Federal District of Venezuela, which has a weak property tax system, for years has received almost half its total revenue from grants.

The preceding paragraph suggests a possible relationship between the extent to which property taxation is exploited as a revenue source and the degree to which grants are relied upon. The relationship appears to be particularly strong in South America, although it may exist elsewhere. 87 Where the relationship does not appear to hold, as in Nairobi, it is due to relatively heavy reliance on another productive revenue source—in this case, the personal income levy. Admittedly, what is “cause” and what is “effect” in this general relationship remains unclear.

In the three developed cities, grants are an important element in revenues but to a lesser extent in Stockholm than in the U.S. cities. One possible explanation offered by A.H. Marshall is that “there is surely some connection between the existence of a local income tax in Finland and Sweden and the financial independence of local government in these countries.” 88 This is not intended to suggest that local authorities in less homogeneous countries could as easily employ a local income tax to the same extent as is done in Sweden. 89

shared taxes

Two basic means are used to transfer revenues to the cities: tax sharing and grants. Both methods are used extensively in both developed and developing countries. City governments in the Philippines and Thailand have both relied heavily upon taxes collected and shared by the central governments; shared taxes in Manila and Bangkok have accounted for 25 per cent and 50 per cent, respectively, of total revenues in recent years. Taxes shared with the local authorities in the Philippines include the following: 5 per cent of the income, estate, gift, and inheritance taxes; 50 per cent of the residence tax; 23 per cent of the agricultural products tax; 17 per cent of all other internal revenue taxes; 30 per cent of the excess income tax; and a share of taxes from mining, franchises, and forestry. 90 The 1967 Decentralization Law specifies that 70 per cent of these allotments are to be distributed on the basis of population and the remaining 30 per cent on the basis of land area. Taxes shared in Thailand are somewhat simpler, in that they are a surcharge on existing central government taxes and are distributed according to origin. The shared taxes are derived from a 10 per cent increase in (1) the business tax, (2) the liquor and beverage tax, and (3) the entertainment tax. In addition, parts of the motor vehicle registration fees and rice export tax are shared.

Shared taxes also play an important role in Brazilian city finances. During the 1960s, 15 per cent of the federal income tax, 10 per cent of the federal consumption tax, 12 per cent of the federal tax on motor fuels and lubricants, 30 per cent of the excess state tax collections, per município, and 40 per cent of any new taxes that the state might collect were earmarked for distribution to the municipalities.

Tax sharing helps to ensure that the more efficient central government tax administration can be relied upon to generate revenues from sources that may not legally be open to use by local authorities; and since the taxes are generally levied nationwide, the problem of tax competition can be avoided. Tax sharing may also be a means of increasing both tax consciousness and the linkages between revenues and expenditure. In this way, taxes collected even by the central governments may be linked to services provided by local governments. Aside from the possible undesirability of increased centralization accompanying any transfer program, tax sharing as utilized in the countries just discussed appears to have much to recommend it.

functional and unconditional grants

Different forms of intergovernmental grants have been used even more widely than tax sharing as a means of providing additional funds to cities and other local authorities. Basic forms of grants are the functional grants that are allocated for a specific purpose and the unconditional grants that the local authorities may use as they see fit. Shared taxes may be similar to either form, depending on whether or not they are earmarked for a specific purpose.

Functional grants may be used to increase expenditure in areas where local underexpenditure has resulted from the existence of external economies. Such aid may take the form of unit grants covering the partial or full cost of each unit up to x units of the service, or it may be in the form of percentage grants covering a fixed percentage of the local expenditure on a particular project. In both forms the objective, and usual result, is to expand a particular government service. The states in Nigeria have used functional grants fairly widely to provide urban and rural water supplies, road maintenance, local police, veterinary services, health services, public parks, and education. Tanzania has also relied heavily upon functional grants. In 1965, 85 per cent of recurrent expenditure and 50 per cent of capital expenditure on education in urban areas were to be covered by grants from the Central Government. Grants were to cover 50 per cent of health expenditure, while the percentage of road maintenance and construction to be covered depended on the grade of road. Percentage grants for specific purposes are also heavily relied Upon by local authorities in the Sudan for education, administration, road, and drainage services. In Sri Lanka, grants for specific purposes are given to local authorities for cost of living allowances for government officials, maternity and child welfare work, hospitals, reading rooms and libraries, rest houses, road repairs and drain construction, playgrounds, parks, and housing.

For some time Colombian cities have relied on functional grants in the fields of education and health. Education grants from the Central Government in 1966 were the equivalent of 1.15 times the teachers’ salaries in 1960. 91 Because of differing rates of expansion in school systems and differing rates of increase in salaries, the aid as a percentage of total education expenditure varied from department to department.

Unconditional grants have been used more sparingly than functional grants in the developing countries. Indeed, until revenue-sharing was enacted in 1972, the U.S. Federal Government, as well, eschewed the use of unconditional grants. Indonesia is a country in which subnational levels of government, including the main cities, have relied heavily upon unconditional grants. Djakarta received 23 per cent of total revenue in this form in 1970. Although an elaborate basis exists for distributing national grants—including population, number of government officials, area, kilometers of roads, and hectares of paddy land—the grants appear to be more closely based on the number of government officials in a city or region. 92 Unconditional grants of this nature could clearly be improved upon by establishing a set of criteria for allocating grants in line with chosen objectives.

Uganda is another country in which unconditional grants have been widely used—in this case with the objective of achieving minimum standards in services provided by local authorities. These grants accounted for from 0 to 70 per cent of current expenditures in the districts of Uganda. 93 In India, general purpose grants are common in the distribution of state funds to rural districts. Unconditional grants are much less commonly made to Indian urban authorities, although they have been used in the Delhi Territory 94 and in the States of Gujarat, Kada, Madhya Pradesh, and Uttar Pradesh.

effects of grants

The question arises as to whether grants, either functional or unconditional, will meet their objectives of stimulating local government expenditure in the desired way. The question of the effects of grants on both revenues and expenditure in the United States has been pursued at some length, but there is a dearth of research on this question in developing countries. Evidence generally supports the belief that functional grants lead to increased local expenditure in the aided categories. 95 Evidence is mixed concerning whether grants lead to distortions in local budgets by causing an absolute reduction in expenditure in unaided categories. 96 The extent to which functional aid leads to an increase in the aided categories depends on the “income elasticity” of demand for that function. Should the income elasticity be sufficiently great, the increase in expenditure could be greater than the grant, and budget distortion would occur as funds are drawn away from other expenditure. Income elasticities for the functions examined in the United States—education, health, highways, and public welfare—have generally been sufficiently low that grants have led to a freeing of resources for other purposes. Other studies have found that grants to local governments for particular purposes have also led to lower local taxes. 97

What then can be said about the desirability and effectiveness of grants to urban governments in developing countries? First, shared taxes or grants are essential, owing to externalities, tax competition, limited administrative resources, and limited revenue sources. Since these factors will differ among cities, the right amount of aid or tax sharing will also differ. Second, functional grants will permit higher levels of government to increase their control over the provision of city services. If it is believed that local authorities are in a better position to determine the optimal mix of city services and that spillovers are minor, unconditional rather than functional grants should be made. If the opposite conditions hold, the opposite conclusion follows. Third, the distribution of grants among localities affects the provision of services and the attractiveness of living in particular areas. Grants may favor smaller cities and rural areas—a result that may or may not be desirable, depending on the development plan of the country. Given the objectives of a nation, some amount and distribution of tax sharing and/or governmental aid is optimal. Neither spillovers, nor administrative capacity, nor problems of tax competition suggest the complete centralization of revenue-raising power in any country. The question, then, is what is the proper level of intergovernmental transfers. This is a topic of extreme relevance to city finances in developing countries and an area in which further research is greatly needed.

intergovernmental transfers—are large cities favored?

Intergovernmental transfers from a significant share of the revenues of many of the major cities examined in this study. In fact, it might be suggested that the largest cities have received more than their share of intergovernmental aid—to the detriment of other cities and local authorities. 98 Although the larger cities may have increased needs, evidence suggests that they are nevertheless in a better position to meet their needs than are somewhat smaller cities. It would be unfortunate if, in such a case, government transfers to local authorities accentuated the problem by giving generous aid to only those cities that already have considerable revenue-raising potential. For example, if per capita grants to Djakarta significantly exceeded those to Makassar, Semarang, and Solo, this might be viewed as undesirable. In fact, for 1969-70 per capita budgeted transfers for Djakarta were Rp 447, while those for Makassar, Semarang, and Solo (when lumped with per capita transfers to their respective provinces) were Rp 612, Rp 326, and Rp 409, respectively. There is at least no clear bias in favor of the principal city.

Table 10 presents some additional data concerning the relationship between central government grants and city size. For Brazil, the capital cities of the states are listed in order of population size. From this, it can be seen that there is no tendency for per capita grants to increase with size;99 however, it is also obvious that special treatment has been accorded to São Paulo (which is by far the largest city)—per capita grants to São Paulo are nearly two and a half times the average. In Colombia, as previous data indicated, grants form a small part of municipal revenues. Nevertheless, the Central Government makes substantial transfers to the departments for the provision of health and education services. Table 10 indicates that these grants do not favor those departments in which the largest cities are located. Per capita transfers to Antioquia, Atlántico, and Valle del Cauca are significantly below the average, and those to Cundinamarca, in which Bogotá is located, are only slightly above the average. Article 53, Legislative Act 1 of 1968—which distributes 30 per cent of central government aid in equal parts among the departments, federal territories, and the Special District, and 70 per cent on the basis of population—ensures that per capita grants to the departments with large populations will continue to be below average.

Table 10Selected Countries: Per Capita Transfers to Subnational Authorities by Central and/or State Governments, Selected Years(In per cent)
Brazil (1968)ColombiaIndonesia (1969/70)Philippines (1970-71)Tanzania (1965)
CityTransfer (cruzeiros)Population (thousands)DepartmentTransfer 1 (pesos)ProvinceTransfer (rupiahs)CityTransfer (pesos)Population 2 (thousands)Town or CityTransfer (shillings)Population 3 (thousands)
São Paulo68.45,685Antioquia20Atjeh643Manila33.41,731Dar es Salaam12.7227
Belo Horizonte25.81,167Atlántico29Bali265Quezon City12.2605Tanga30.057
Recife30.81,100Bolívar23Bengkulu719Cebu10.4382Mtwara /Mikindani13.923
Pôrto Alegre39.7933Boyacá27Djakarta (D.C.I.)407Davao12.6343Moshi44.121
Salvador25.2892Caldas27Djambi654Caloocan15.1221Lindi31.615
Fortaleza20.8846Cauca33Java, Central197Batangas41.2126Arusha43.915
Curitiba33.0617Córdoba14Java, East117Toledo122.897Iringa29.714
Belém23.7571Cundinamarca32Java, West226Mbeya31.710
Goiânia17.8345Chocó65Jogjakarta318
Niterói17.8345Guajira38Kalimantan, Central992
Manaus25.1304Huila31Kalimantan, East1,148
Natal30.6256Magdalena20Kalimantan, South438
São Luís19.6240Meta51Kalimantan, West426
Maceió24.4221Nariño27Lampung719
Teresína21.4190Norte de Santander34Maluku580
João Pessoa21.9189Santander28Nusa Tenggara, East382
Aracaju26.4156Tolima34Nusa Tenggara, West255
Florianópolis23.9132Valle del Cauca25Riau475
Vitória37.7127Sulawesi, Central
Cuiabá de Larga29.687Sulawesi, North635
Pôrto Velho12.582Sulawesi, South
Macapá26.872Sulawesi, Southeast339
Rio Branco43.672Sumatra, North339
Boa Vista25.937Sumatra, South176
Sumatra, West365
Average28.0Average31Average470.2Average35.4Average29.7
Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each country. For Brazil, the transfers were calculated from figures taken from No. 7; for Colombia, see No. 11; Indonesia, No. 19. For the Philippines, the transfers were calculated from data taken from No. 41; the population data were taken from No. 37 and were adjusted for an annual urban growth rate of 4.3 per cent. For Tanzania, the figures were derived from grant information in No. 49 and from No. 47; the 1957 government census figures were adjusted by town and city growth rates that were estimated in No. 48.

A sum of current transfers by the Central Government in 1966 for health and education (including auxilios regionales granted by Congress) divided by the departmental population in 1964. Data from Dirección Nacional del Presupuesto and Departamento Administrativo Nacional de Estadística.

In 1970.

Estimated.

Sources: See the Sources Used in Preparing Tables, in the Appendix, as specified for each country. For Brazil, the transfers were calculated from figures taken from No. 7; for Colombia, see No. 11; Indonesia, No. 19. For the Philippines, the transfers were calculated from data taken from No. 41; the population data were taken from No. 37 and were adjusted for an annual urban growth rate of 4.3 per cent. For Tanzania, the figures were derived from grant information in No. 49 and from No. 47; the 1957 government census figures were adjusted by town and city growth rates that were estimated in No. 48.

A sum of current transfers by the Central Government in 1966 for health and education (including auxilios regionales granted by Congress) divided by the departmental population in 1964. Data from Dirección Nacional del Presupuesto and Departamento Administrativo Nacional de Estadística.

In 1970.

Estimated.

The situation in the Philippines is somewhat mixed. Per capita transfers to the seven cities in terms of total allotments are seemingly U-shaped, falling with population and then rising. Per capita transfers vary tremendously, and for the two small cities—Batangas and Toledo—are substantially in excess of per capita transfers to Manila. The data generated for Dar es Salaam and seven other Tanzanian towns indicate that intergovernmental transfers do not favor the largest city in that country. Indeed, per capita transfers by the Central Government to all the smaller towns exceed those to Dar es Salaam and, on average, are two to three times as great.

Although the evidence presented in Table 10 is not conclusive, it seems that any argument that grants-in-aid consistently favor the largest or larger cities of a country is clearly suspect. Of the cities examined, only São Paulo and perhaps Manila appear to be favored.

Some of the states in India have made a conscious effort to distribute state aid in such a way as to favor the smaller urban areas. Gujarat has a scheme under which per capita grants range from 30 paise for municipalities with populations over 100,000 to 60 paise for those with populations between 20,000 and 50,000. A similar scheme exists in Madhya Pradesh. In Kerala, major corporations and municipalities receive Rs 1.00 per capita from the state, and minor municipalities and townships receive Rs 1.50. The appropriateness of this type of discrimination in favor of the smaller municipalities will certainly vary from country to country.

VI. Concluding Observations

If the complex problems of cities are to be approached intelligently, there is a real need for developing information that does not exist at present. Data on urban areas are sparse and frequently abysmal. Some countries have begun to collect data on a regional but not on an urban basis, and the importance of city problems is only beginning to be recognized. So, severe data constraints hamper urban research in developing countries. The research needed on this subject is, as suggested in this paper, indeed varied and complex. Given the propensity of trained economists to work in areas where well-developed data already exist, the problems mentioned here have, too frequently, received inadequate attention.

Urban life requires public expenditure that rural life does not require, and large cities may have needs that smaller cities do not have. Little information is available concerning the per capita cost of providing a certain level of services in different settings. If a state or national objective is to attain some degree of equity in the provision of public services, information of this nature would help to establish a rational system of intergovernmental aid. In addition, information concerning “tax effort” in cities of various sizes would be essential for this purpose.

In both developed and developing countries, concern is voiced over the extent to which the urban sector subsidizes the rural sector, or vice versa. This is a complex problem, made more so by the division of governmental responsibilities within urban areas. Little information exists about expenditure by central and provincial governments in different localities. The same is true for revenues. This lack of information makes it difficult to ascertain accurately whether the urban or rural sectors are being favored by higher levels of government, and to determine what resources should be made available to either sector. Unfortunately, lack of information coupled with the urgent need for action leads to uninformed discussions and frequently to inappropriate decisions on these or other problems.

Much additional research is also needed on the connection between economic development and trends toward centralization. Evidence on this issue is mixed. In some countries, such as Korea and Sri Lanka, local government expenditure as a percentage of total government expenditure has been increasing steadily. In Korea, the change has been rapid as well as steady; in many other countries, a trend is much more difficult to discern. In all countries concerned, there are important questions that need answers. For example, why has local expenditure, as a share of the total, grown so much more rapidly in Korea than in the Philippines, when urban growth rates have been similar? In Latin America, why has municipal expenditure as a share of total government expenditure decreased in Brazil and Venezuela, over the past 15 years, while the proportion has increased in nearby Colombia? These differing trends, on closer examination, may well turn out to be due largely to individual special circumstances. Even so, the question of how these differing trends have affected the provision of city services remains open, and important.

Further work is also needed even to begin to answer the difficult question of the extent to which international aid—from foreign governments, international organizations, and private aid-giving institutions—has contributed to a further centralization of governmental powers.

Provision of public services by city governments in developing countries was found to vary tremendously. In some countries, the largest city or cities have additional powers and responsibilities. An unanswered question concerns the circumstances under which additional authority should be granted to cities, since it seems to have been beneficial in some cases. Studies providing additional evidence on the causes and effects of preferential status for a number of large cities should prove useful in determining what steps should be taken in the future.

Although difficult data problems severely limit the usefulness of comparative analysis of expenditure patterns, it is clear that public works, health, education, and general administrative activities account for most city expenditure. Factors partially accounting for large differences in per capita city expenditure include different divisions of responsibility among government levels and differences in city per capita incomes, tax effort, and the rate of urbanization.

Cities rely on a tremendous number of revenue sources, but only a few are of major importance. Taxes on property are heavily utilized by the African and Asian countries with a British heritage. In addition, Mexico City and the Colombian cities have developed relatively effective property tax systems. Sales taxes of various types are a significant revenue source for Manila, Saigon, several Indian cities, Mexico City, and the cities in Brazil, Colombia, and Venezuela. Other common revenue sources include taxes on motor vehicles and on entertainment. In Djakarta, motor vehicle taxes account for about one fourth of total revenues, and entertainment and gambling taxes account for another fourth. The most obvious partial answer to the financial plight of most cities is for them to increase the effectiveness of their existing major taxes rather than to seek out new revenue sources. Improved use of real property taxes and betterment levies is the most common need.

Since cities remain dependent on shared taxes and grants received from central governments, an important and largely unanswered question is the extent to which grants and shared taxes enable higher levels of government to exert control over the cities. Another important question is the extent to which inept central government control can be avoided without any loss in city government revenues through a transfer of revenue sources, rather than of revenues, to the cities. It appears that intensive case studies are the only way to begin to answer these difficult questions.

In general, data limitations confronted in this study make for rather tenuous conclusions. Nevertheless, it can be said with certainty that a host of circumstances—historical, political, economic, and social—determine the optimal as well as the existing division of revenue sources and expenditure responsibilities in each country, and that there is no one mix of revenue sources and expenditure responsibilities that is appropriate for every city in the world.

APPENDIX: Sources Used in Preparing Tables
General

    KingsleyDavisWorld Urbanization 1950-1970 Volume I: Basic Data for Cities Countries and Regions Institute of International Studies (University of CaliforniaBerkeley1969) Table C (pp. 113-38).

    JosephFroomkin“Fiscal Management of Municipalities and Economic Development,”Economic Development and Cultural ChangeVol. 3 (July1955) p. 315.

    International Bank for Reconstruction and Development“Urbanization, Sector Working Paper,” in World Bank Operations: Sectoral Programs & Policies (Johns Hopkins University Press1972) pp. 405-513especially p. 485.

    International Bank for Reconstruction and DevelopmentWorld Tables (mimeographed, Washington1971).

    United Nations Statistical OfficeStatistical Yearbook 1970Sales No. E/F 71.XVII. 1 (New York1971).

Brazil
    Canada
    Chad
    Colombia
    Dominican Republic
      Guatemala

        John O.Rees“The Organization and Financing of Local Governments in Guatemala,”U.S. Agency for International Development Public Administration Division (mimeographed, WashingtonAugust1962).

      India
      Indonesia
      Kenya
      Korea
      Malawi
      Mexico

        OliverOldman and othersFinancing Urban Development in Mexico City: A Case Study of Property Tax Land Use Housing and Urban Planning (Harvard University Press1967).

      Morocco
      Nigeria

        G. OkaOrewaLocal Government Finance in Nigeria (Oxford University Press1966).

      Panama
        Philippines
        Sierra Leone
        Singapore

          Memorandum on the Estimates of Revenue and Expenditure of Singapore1969/70 (Singapore1970).

        Sri Lanka
        Sweden
          Tanzania
          Thailand
          Turkey

            State Institute of StatisticsBudget and Final Accounts of Special Provincial Administrations and Municipalities1956-64Publication No. 581 (Ankara1969).

          United States
          Uruguay
          Venezuela
          Viet-Nam
          Zambia

          Mr. Smith, an economist in the Fiscal Affairs Department, received his graduate training at Harvard University and the University of California at Berkeley, and taught at the University of Alberta, Canada, prior to joining the Fund. He is the author of Local Income Taxes: Economic Effects and Equity and has contributed a number of articles on subnational public finances to economic journals.

          N. V. Sovani and David R. Kamerschen have made rather convincing arguments that the “pull” of the urban areas has contributed much more to urban growth than has the “push” from the countryside. See Sovani, “The Analysis of ‘Over-Urbanization’,” Economic Development and Cultural Change, Vol. 12 (January 1964), pp. 113-22, and Kamerschen, “Further Analysis of Over-urbanization,” Economic Development and Cultural Change, Vol. 17 (January 1969), pp. 235-53. William Alonso, in “Urban and Regional Imbalances in Economic Growth,” Economic Development and Cultural Change, Vol. 17 (October 1968), pp. 1-14, argues that urban growth has played a crucial role in economic development.

          William J. Baumol, “Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis,” American Economic Review, Vol. 67 (June 1967), pp. 415-26.

          D. F. Bradford, R. A. Malt, and W. E. Oates, “The Rising Cost of Local Public Services: Some Evidence and Reflections,” National Tax Journal, Vol. 22 (June 1969), p. 201.

          Ibid., p. 202.

          One study estimated that per capita costs to install piped water in urban areas would be $10-15 in Africa, $12 in India, $42 in Sri Lanka, $40 in Brazil, and $25 in Jamaica. These figures are large, relative to annual city per capita expenditure for recent years (see Tables 2 and 7 in this paper). See John M. Henderson, Report on Global Urban Water Supply Program Costs in Developing Nations, 1961-1975, International Cooperation Administration (mimeographed, Washington, June 1961), pp. 5-6.

          The comparison in Zambia is based on revenues rather than expenditure.

          Some evidence that the gap between “expenditure requirements” and actual expenditure is greater for large cities than for medium-sized or small cities is found in a study by the Zakaria Committee in India. See Committee of Ministers constituted by the Central Council of Local Self-Government, Government of India, Report of Committee on Augmentation of Financial Resources for Urban Local Bodies (Delhi, Ministry of Health, 1965).

          The shortcomings of these data are discussed in the section, problems with measuring expenditure patterns.

          William R. Spence, “Design and Implementation of a Development Project System in Tanzania” (doctoral dissertation, University of Toronto, 1972), p. 215.

          Ibid.

          Kingsley Davis, World Urbanization, 1950-1970, Volume I: Basic Data for Cities, Countries, and Regions, Institute of International Studies (University of California, Berkeley, 1969), p. 144.

          Dick Netzer, “Some Aspects of Local Government Finances in Colombia,” Economic Development Report, No. 51, Development Advisory Service, Harvard University (mimeographed, Cambridge, Massachusetts, 1967), p. 29.

          This was true in Indonesia in the late 1950s, for example.

          For India, the existence of a rural bias is at least debatable. Michael Lipton develops an argument in support of an urban bias in “Strategy for Agriculture: Urban Bias and Rural Planning,” Chapter 4 in The Crisis of Indian Planning: Economic Planning in the 1960s, ed. by Paul Streeten and Michael Lipton (Oxford University Press, 1968), pp. 83-147. Lipton argues that urban bias has led to underinvestment in the agricultural sector as well as to insufficient concern for the welfare of the rural population.

          For a comprehensive analysis of the way in which shifting assumptions affect the distribution of tax burdens and expenditure benefits, see Luc De Wulf, “Fiscal Incidence Studies in Developing Countries: Survey and Critique” (unpublished, International Monetary Fund, February 15, 1974); also, see R. S. Smith, “Are Nonresidents Contributing Their Share to Core City Revenues?” Land Economics, Vol. 48 (August 1972), pp. 240-47.

          For discussions of centralization and decentralization, see Cambridge University Overseas Study Committee, Summer Conference on Local Goverment in Africa (Cambridge University Press, 1961), pp. 8-10, and Wallace E. Oates, “An Economic Approach to Federalism,” Chapter 1 in Fiscal Federalism (New York, 1972), pp. 3-20.

          Most discussions of centralization become involved in rather complex terminology: centralization-decentralization, concentration-deconcentration, devolution-delegation. Decentralization refers to the division of total political and administrative powers of government among separate authorities, and it implies some degree of local authority such as is found in federal systems. Deconcentration is more a matter of administrative convenience, with political powers tending to be quite centralized. Delegation of power generally implies greater deconcentration, not necessarily greater decentralization. Devolution of power implies greater decentralization.

          Charles M. Tiebout first developed this line of thought in “A Pure Theory of Local Expenditure,” Journal of Political Economy, Vol. 64 (October 1956), pp. 416-24. It has been seriously questioned recently by Lyle C. Fitch in “Scale and Effectiveness of Urban Government: A Commentary,” Chapter 9 in Modern Fiscal Issues: Essays in Honor of Carl S. Shoup, ed. by Richard M. Bird and John G. Head (University of Toronto Press, 1972), pp. 164-86, especially pp. 166-68. Fitch argues that even in industrialized countries only the upper-income groups are likely to take advantage of the wider choice in available public services and to “vote with their feet” by moving.

          This point is made in the following articles: Alison Martin and W. Arthur Lewis, “Patterns of Public Revenue and Expenditure,” Manchester School of Economic and Social Studies, Vol. 24 (September 1956), pp. 203-44; R. S. Smith and Theodore M. Smith, “The Political Economy of Regional and Urban Revenue Policy in Indonesia,” Asian Survey, Vol. 11 (August 1971), pp. 761-86; John H. Adler, “A Reply to Mr. Bauer,” in Readings on Taxation in Developing Countries, ed. by Richard M. Bird and Oliver Oldman (Baltimore, Revised Edition, 1967), pp. 372-82. (This originally appeared as Section I, “Government Receipts and Expenditures in an Underdeveloped Economy,” of “The Economic Development of Nigeria: Comment,” Journal of Political Economy, Vol. 64 (October 1956), pp. 425-34.) Adler states that: “… the willingness to pay taxes directly depends on the immediacy of the benefits of the government services which these taxes ‘buy.’ The rural areas of Nigeria abound with examples of virtually voluntary increases in taxes—where the taxpayer directly benefits from new or improved communal facilities such as schools, hospitals, community centers, and roads” (p. 380).

          Clifford Geertz carefully analyzes the important role that local talent can play in the development process in Peddlers and Princes (University of Chicago Press, 1963). This point is also developed by David C. McClelland, The Achieving Society (Princeton, New Jersey, 1961), pp. 431-32.

          It may be possible to establish equal fiscal capacity and to facilitate greater homogeneity in the provision of public services among subnational governments without increasing centralization. For a discussion of the possibilities, see James M. Buchanan, “Federalism and Fiscal Equity,” American Economic Review, Vol. 40 (September 1950), pp. 583-99, reprinted in Readings in the Economics of Taxation, Vol. 9, ed. by Richard A. Musgrave and Carl S. Shoup (Homewood, Illinois, 1959), pp. 93-109.

          Geertz, Peddlers and Princes (cited in footnote 19), p. 154.

          Alan T. Peacock and Jack Wiseman offer another reason to expect increasing centralization: “the major and inescapable problems of urban life… have been understood and largely solved during the period, leaving local governments with the relatively less onerous task of maintenance of established standards.” Unfortunately, the same cannot be said for urban governments in developing countries. See Peacock and Wiseman, The Growth of Public Expenditure in the United Kingdom (London, 1967), p. 119.

          Geertz, Peddlers and Princes (cited in footnote 19), p. 155.

          James W. Fesler, “Approaches to the Understanding of Decentralization,” Journal of Politics, Vol. 27 (August 1965), p. 537, cited by Richard M. Bird in The Growth of Government Spending in Canada, Canadian Tax Foundation (Toronto, 1970), p. 168.

          Milton C. Taylor, Local Government Finance in Vietnam, Vietnam Working Paper No. 1, Development and Resources Corporation (mimeographed, New York, August 1969), p. 9.

          Defined as a residual where public utilities, development expenditure, debt service, education, and health expenditure are subtracted from total expenditure.

          Netzer, “Some Aspects of Local Government Finances in Colombia” (cited in footnote 12), p. 17.

          For a framework upon which to base such an analysis, see Richard M. Bird and Douglas G. Hartle, “The Design of Governments,” in Modern Fiscal Issues: Essays in Honor of Carl S. Shoup (cited in footnote 17), pp. 45-62.

          Carl S. Shoup, The Tax System of Brazil: A Report to the Getulio Vargas Foundation (Rio de Janeiro, 1965), pp. 13-14.

          Pan American Union, Constitution of the Republic of Venezuela, 1961 (Washington, 1961), pp. 38-39.

          Organization of American States, Constitution of the Republic of Colombia, 1886 (Washington, 1971), p. 45.

          Oliver Oldman and others, Financing Urban Development in Mexico City: A Case Study of Property Tax, Land Use, Housing, and Urban Planning (Harvard University Press, 1967), p. 29.

          Jorge Macon, “Urban Growth and Horizontal Equity—The Case of Buenos Aires,” in Problèmes de finances publiques urbaines (Issues in Urban Public Finance), Institut International de Finances Publiques (Saarbrücken, 1973), pp. 355-64. (Paper delivered to a meeting of the International Institute of Public Finance in New York, September 12-15, 1972.)

          Netzer, “Some Aspects of Local Government Finances in Colombia” (cited in footnote 12), pp. 129-30.

          Ibid., pp. 126-34.

          Rio de Janeiro, being a state as well as a municipality, has revenue resources and expenditure responsibilities that differ from those of other Brazilian cities.

          United Nations, Department of Economic and Social Affairs, Seminar on Central Services to Local Authorities, E/CN.ll/664/Rev. 1 (New York, 1964), pp. 128-29 and 198.

          Public Administration Service, “Local Government Finance and Organization in Sierra Leone: A Reconnaissance Report” (mimeographed, Chicago, 1965), p. 25.

          R. G. Penner, Financing Local Government in Tanzania, Economic Research Bureau (Dar es Salaam, 1970), pp. 26-28; G. Oka Orewa, Local Government Finance in Nigeria (Oxford University Press, 1966), pp. 18-21; City Council of Nairobi, Abstract of Accounts for the Year Ended 31st December, 1965 (Nairobi), pp. 3-4.

          United Nations, Department of Economic and Social Affairs, Decentralization for National and Local Development, Sales No.: 62.II.H.2 (New York, 1962), p. 138.

          Dragoslav Avramovic, Economic Growth of Colombia: Problems and Prospects, Report of a World Bank Mission to Colombia in 1970 (Baltimore, 1972), p. 429.

          Netzer, “Some Aspects of Local Government Finances in Colombia” (cited in footnote 12), p. 53.

          Although municipalities in West Bengal (Calcutta) were responsible for universal education, only one half to two thirds of the children attended. See Annmarie Hauck Walsh, The Urban Challenge to Government: An International Comparison of Thirteen Cities (New York, 1969), pp. 139 and 240-47.

          This comparison is only for the five states in which Ahmedabad, Bombay, Calcutta, Hyderabad, and Madras are situated.

          City Council of Nairobi, Abstract of Accounts (cited in footnote 39), pp. 3-4.

          Penner, Financing Local Government in Tanzania (cited in footnote 39), p. 19.

          Orewa, Local Government Finance in Nigeria (cited in footnote 39), pp. 18-19.

          Public Administration Service, “Local Government Finance and Organization in Sierra Leone” (cited in footnote 38), pp. 15-16.

          Calculated from Dirección General del Presupuesto, División de Estudios Presupuestales, Boletín, No. 56 (Bogotá, 1969), pp. 172, 264, and 348.

          United Nations, Decentralization for National and Local Development (cited in footnote 40), pp. 138-39.

          Calculated from Instituto Brasileiro de Estatística, Anuáriá Estatístico do Brasil, 1970 (Rio de Janeiro, 1970), pp. 714-37.

          As Table 7 indicates, per capita expenditure in Bombay in 1970 was $13.65.

          T. M. Smith and R. S. Smith, “Municipal Finance,” Bulletin of Indonesian Economic Studies, Vol. 7 (March 1971), p. 118.

          Calculated from data taken from Anuáriá Estatístico do Brasil, 1970 (cited in footnote 51), pp. 41 and 734.

          Harold F. Alderfer, “Four Basic Patterns,” Chapter 1 in Local Government in Developing Countries (New York, 1964), pp. 1-16; Joseph Froomkin, “Fiscal Management of Municipalities and Economic Development,” Economic Development and Cultural Change, Vol. 3 (July 1955), pp. 309-20 (also reprinted in Readings on Taxation in Developing Countries, ed. by Richard M. Bird and Oliver Oldman (Baltimore, Revised Edition, 1967), pp. 415-30).

          Froomkin, “Fiscal Management of Municipalities and Economic Development” (cited in footnote 55), p. 313.

          William A. Robson, “The Great City of Today,” Part One in Vol. 1, Great Cities of the World: Their Government, Politics and Planning, ed. by William A. Robson and D. E. Regan (London, 1972), p. 103.

          Froomkin, “Fiscal Management of Municipalities and Economic Development” (cited in footnote 55), p. 314.

          The distinction between these five countries and the previous group is difficult to define, but it appears that in the five the central government plays a more active role in setting standards, selecting personnel, etc.

          For a comparison of intergovernmental participation in urban services in 12 cities of the world, see Walsh, The Urban Challenge to Government (cited in footnote 43), p. 132. Walsh’s findings for Calcutta, Casablanca, Davao, Karachi, Lagos, Lima, Łódź, Paris, Stockholm, Toronto, Valencia, and Zagreb are consistent with the situation described here. Local participation is frequently found in the provision of education, public welfare, and police services; streets and highways and public health are usually shared; and sanitation and sewerage, fire protection, and parks and recreation are normally the sole responsibility of the local authorities.

          A. H. Marshall, New Revenues for Local Government, Fabian Research Series, No. 295 (London, June 1971), p. 5.

          Kenneth T. Cann, “The Structure of Local Government Finance in Brazil with Comments on Its Relationship to Community Development,” Land Tenure Center (mimeographed, University of Wisconsin, 1964), p. 22.

          Malcolm Gillis, “Reform of Municipal Indirect Taxes, Service Taxation, and Stamp Duties,” in Fiscal Reform for Colombia: Final Report and Staff Papers of the Colombian Commission on Tax Reform, ed. by Malcolm Gillis (Harvard Law School, 1971), p. 648.

          The reason that Rio de Janeiro (with 70 per cent of total revenues from sales taxation) has not been singled out in the foregoing discussion is that it is also the State of Guanabara with the right to employ a VAT. Because of its special status, the Federal District of Mexico is also able to levy a broad-based tax on commercial receipts.

          Ibadan, Nigeria, is another city that uses the income tax, reportedly having received 34.5 per cent of its total revenues from income taxes in 1964-65. See Leslie Green, “Ibadan,” in Vol. 1, Great Cities of the World (cited in footnote 57), p. 468.

          Cairo, which has not been included in our sample, obtained 10 per cent of total revenues from motor vehicle taxation in 1967—more than from any other local tax source. See Fouad D. Abdel-Tawab, “Cairo,” in Vol. 1, Great Cities of the World (cited in footnote 57), pp. 287-91.

          Juanita D. Amatong in a study, “The Revenue Importance of Government Lotteries” (unpublished, International Monetary Fund, April 16, 1968), found that lotteries were used as a revenue source by 58 national governments, state (and provincial) governments in 8 countries, and local governments in Japan and Indonesia.

          While net revenues from water are included for most cities, the figures for Guatemala City and Seoul reflect gross revenues.

          Froomkin, in “Fiscal Management of Municipalities and Economic Development” (cited in footnote 55), p. 319, suggests that the most promising revenue sources for cities are “(1) local property taxes, (2) taxes on activities specifically concentrated in towns, such as amusements, and (3) licenses for specific activities.”

          Daniel Holland and Oliver Oldman, “Djakarta Real Estate Tax Study: Final Report” (unpublished study conducted by the International Bank for Reconstruction and Development for the Government of Indonesia, October 1972). See also Smith and Smith, “Municipal Finance” (cited in footnote 53), where a tenfold increase in effective property tax rates is recommended.

          Robert R. Nathan Associates, Inc., Chapter I, “The Property Tax,” in Local Government Finance in Korea (mimeographed, Washington, 1971), pp. 10-37.

          Taylor, Local Government Finance in Vietnam (cited in footnote 25), pp. 21 and 22.

          Richard M. Bird, “Local Government Finance,” Chapter 5 in Taxation and Development: Lessons from Colombian Experience (Harvard University Press, 1970), pp. 147-84.

          Oldman and others, Financing Urban Development in Mexico City (cited in footnote 32).

          Joint Tax Program of the Organization of American States and the Inter-American Development Bank, Fiscal Survey of Panama: Problems and Proposals for Reform (Baltimore, 1964), pp. 207-208.

          Carl S. Shoup and others, The Fiscal System of the Federal District of Venezuela: A Report (Baltimore, 1960), Chapter 4, pp. 27-75.

          Ursula K. Hicks has also stressed the importance of developing an effective tax on urban realty. See her Development from Below: Local Government and Finance in Developing Countries of the Commonwealth (Oxford University Press, 1961), p. 348.

          See William G. Rhoads and Richard M. Bird, “The Valorization Tax in Colombia: An Example for Other Developing Countries?” Chapter 10 in Land and Building Taxes, ed. by Arthur P. Becker (University of Wisconsin Press, 1969), pp. 201-37.

          Gillis, “Reform of Municipal Indirect Taxes, Service Taxation, and Stamp Duties” (cited in footnote 63), pp. 650-56.

          Carl S. Shoup and others, The Fiscal System of Venezuela: A Report (Baltimore, 1959), pp. 330-32; Roy Bahl, “Public Finance and the Efficiency of Urban Development,” Annex IV, Report on Bombay, International Bank for Reconstruction and Development (unpublished, Washington, April 30, 1971), pp. 52-53.

          Joint Tax Program, Fiscal Survey of Panama (cited in footnote 75), p. 207.

          Shoup and others, The Fiscal System of Venezuela (cited in footnote 80), p. 343.

          For some estimates of revenues that can be generated through congestion charges, see Anthony Churchill, Road User Charges in Central America, World Bank Staff Occasional Papers, No. 15 (John Hopkins University Press, 1972), Chapters 7 and 8, pp. 106-74.

          For a discussion of an optimal road tax system in Central America, see Churchill, op. cit.

          Bangkok’s aid from the Central Government, in particular, is understated. In addition to grants, another 50 per cent of its revenues is collected for the municipality by the Central Government.

          Annual Report of the Ministry of Local Government for the Year 1966 (Zomba, 1967), pp. 9-10.

          Spearman’s coefficient of rank correlation between taxes on property as a percentage of total revenues and grants plus shared taxes as a percentage of total revenues for nine South American cities in Table 9 proves to be significantly greater than zero at a 99 per cent confidence level. (Rio de Janeiro is omitted from the sample, since as a state it can and does rely heavily upon a VAT; and 1959 data is used for Caracas.) If the three Central American cities are added, the confidence level remains at 99 per cent. For the groups of Far Eastern, South Asian, and African cities, the rank correlation coefficient between these two series is not significantly greater than zero.

          A. H. Marshall, Local Government Finance, International Union of Local Authorities (The Hague, 1969), p. 22.

          For an examination of local income taxes used in the United States, as well as consideration of the theoretical effects of these taxes, see R. Stafford Smith, Local Income Taxes: Economic Effects and Equity, Institute of Governmental Studies (University of California, Berkeley, 1972).

          Eduardo Z. Romauldez, Angel Q. Yoingco, and Antonio O. Casern, Philippine Tax System (Manila, 1970), p. 249.

          Netzer, “Some Aspects of Local Government Finances in Colombia” (cited in footnote 12), pp. 58-59.

          Smith and Smith, “The Political Economy of Regional and Urban Revenue Policy in Indonesia” (cited in footnote 18), pp. 772-73.

          Mark A. Haskell, “Local Government Finance in Uganda: Taxes, Expenditures and Grants-in-aid,” Eastern Africa Economic Review, Vol. 4 (June 1972), pp. 75-83.

          Marshall, Local Government Finance (cited in footnote 88), pp. 238-42.

          Empirical studies supporting this point include those by Seymour Sacks and Robert Harris, “The Determinants of State and Local Government Expenditures and Intergovernmental Flows of Funds,” National Tax Journal, Vol. 17 (March 1964), pp. 75-85; David L. Smith, “The Response of State and Local Governments to Federal Grants,” National Tax Journal, Vol. 21 (September 1968), pp. 349-57; Thomas O’Brien, “Grants-in-Aid: Some Further Answers,” National Tax Journal, Vol. 24 (March 1971), pp. 65-77.

          Smith, in ‘The Response of State and Local Governments to Federal Grants,” argues that aid has not led to distortion and that expenditure in the unaided as well as in the aided categories has increased. O’Brien, in “Grants-in-Aid: Some Further Answers,” finds that aid has caused reduced expenditure in the unaided categories, or that distortion occurs.

          See George A. Bishop, “Stimulative versus Substitutive Effects of State School Aid in New England,” National Tax Journal, Vol. 17 (June 1964), pp. 133-43; Ronald G. Bodkin and David W. Conklin, “Scale and Other Determinants of Municipal Government Expenditures in Ontario: A Quantitative Analysis,” International Economic Review, Vol. 12 (October 1971), p. 471.

          Froomkin, in “Fiscal Management of Municipalities and Economic Development” (cited in footnote 55), p. 310, expresses particular concern for those cities that have less financial autonomy and less political clout than the largest cities in the country.

          “Spearman’s coefficients of rank correlation between city size and per capita grants for the cities in Brazil, the Philippines, and Tanzania in Table 10 are not significantly in excess of zero.

          Other Resources Citing This Publication