_, >"'O-. ..-. IO~.Ook.-Mm” 0.0.... . . _..- RA\.\.¢.... .309. ... .........-... .......... . .. ... . ... . .. . . ... . . . . . ... . . . .. ... . . .. . . . n .. _ . _ ._ . . .. _ ._ .. . . .. .H...._... .. ......1 ....;................ . ~.... . ... .. ......... ..... .. £2. ..:...33...‘ a..... u...L...!fdfdmufifiauth. ...: . . . . . . a _ . . ... . .. a . .- u..c.c ,....0..u.5..fl.n..q..w_ ....53. M f . . . _ .. ._ .. _ u . . .._.... . ,. . . . . . r , 4.3”... . . . . . ...._ .. . .. .. . .. _ . . . . . . . ... . _ . . . .. . . . . . . . . . . ...... . . ..4...0¢”_3.U1M:”...¢o.. . ”mm”; .1 . . . . . . .. . _ . . . 75;...5553..7.....H.“..,. ... .. . . .. :5 . . . . . . . .g ..n .. . .fififififfir . . . . . . . . . . H. . . . . .. .2. . .../....émmah — p . . . m .A ”N, . . . ...._ 37/31.. “1...”. .. ...:Ja...a...ol.. a h... . . . . .. . . _ . . 56 . m. .5 _ M. .. _ .. v. ... . m F0 .2 F P. . . _. m . . _ 0, ...U . . . 4...“..5. H...H.,..,......u. pill. . . . .. . . . .. . ._ ...... .. H i . . . . ._ . H ..... - ..... ... ... ..u ‘ L E 900 .0 ll. . _ . .. ...: .... . ...”, ......W... w m .A .C .... . N .S . . , ,. .. w A v E e . . . . U _. . _ . . .. . ......m..;....u.....pu...... . R T ”S—_— g . 2 . . . . . . . r. c :9“. . .rn.c. 'Ho' ,.‘.. .0 nu S M 5...... flu nun & 9 . . ... . . w ..l . . . . _ . _ _ . . I'll M .. e S . . . _ . . . _ _ w. _ ...... H . _ . ._ « ..: B C ...... N . . . . _ ._ . on Up C M . M . . . . . . .__ .R 1. f .i s . . . .. ...4... . ..:. . r . m 0 N M s H .. _ . . ... _ , .... _ o c A 0 .3' C . . . u... .a I. ... . . . u c .m __ _. . u T .0 H . . L . . . . . Ill. . . C c H . - _.._. « f... .0 . . 6 ...L .S . . . M .3 ._ . _.. . “ . . , . u .. .. nw.~.«............ .H .-. .... .......... ......mm.“.....u......n,...n....7_.n ..M.......u..... 630.... .......m............. 2%.”...5? . .....vwwuumduwwddz ..M...._.$....w...23.. .....smw....f.w.n?...m..n “..:...” .. . I..F.4..........wfle.\.v... J.u.m...n..v..137695.42...... . - .. l I I! I! I l '1’ H ’l' 1- 6 7 9 8 7 m 1| I II!!! N! I!!! "I NUIIUIIIIUIIW III”! {II mm 129 ABSTRACT A CORRELATION ANALYSIS OF SELECTED URBAN PUBLIC SERVICE EXPENDITURES AND SOCIO-ECONOMIC CHARACTERISTICS OF MICHIGAN CITIES By Ismet Kilincaslan The determination of socio-economic variables which affect urban public service expenditures is a general defi- ciency in the understanding of public services. Given a long range land-use plan, the spatial distribution of ex- penditure requirements within urban areas is an essential prerequisite-for orderly and efficient growth. Determi- nants of urban public expenditures help policy makers de- velop a better anticipation of the actual and future costs of land-use plans. The need for a theory explaining the variations and difficulties in this formulation, difficulties in the measurement of the quality and level of services, urban- suburban diSparities in the consumption of these services and scale economies are the main issues in the study of public services. Simple correlation and regression analyses are u- sed to analyse the relations of five services- police, fire, sanitation, parks and recreation, highway- and eight Ismet Kilincaslan socio-economic characteristics of cities, with 1970 data. Economic variables are the most significant deter- minants for urban public services in cities of Michigan. There are few economies of scale, if any; there is a need for more qualitative data for further exploration of these: economies. Expenditures vary by function and structure of cities; suburban cities differ from single cities in their allocation and spending. A CORRELATION ANALYSIS OF SELECTED URBAN PUBLIC SERVICE EXPENDITURES AND SOCIO-ECONOMIC CHARACTERISTICS OF MICHIGAN CITIES BY Ismet Kilincaslan A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER IN URBAN PLANNING School of Urban Planning and Landscape Architecture 1972 ACKNOWLEDGMENTS I would like to express my appreciation to several persons who have contributed their knowledge and support to this achievement. I am especially indepted to: Prof., Dr. W. Paul Strassman, Department of Economics, Michigan State University, who initiated my interest in Urban Economics and gave encouragement. Prof., Dr. Carl Goldschmidt, School of Urban Planning, and Landscape Architecture, MSU, who offered valuable advice and criticism as my thesis advisor. Mrs. Ruth Hosley, Bureau of Municipal Finance, Michi- gan Department of Treasury, who provided me the-most recent data from her special files; and MSU Computer Laboratory personnel who assisted in every phase of the analysis. My wife, Tulay, who deserves special recognition and thanks for her patience and faithful support. ii TABLE OF CONTENTS INTRODUCTION O00............OOOOOOOOOOOOOOO......OOOOOOOO 1 CHAPTER I. ISSUES IN THE STUDY OF URBAN PUBLIC SERVICES.....t. 3 Need for a Theory, Difficulties Measurement of Urban Public Services Intrametropolitan Disparities Economies of Scale II. REVIEY OF SELECTED STUDIESOOOOOO0.0.0.000000000000014 British Studies Studies in the U.S. before 1945 Hawley Scott and Feder Brazer Fabricant, Fisher, Sacks and Harris Hirsch Pidot, Bahl III. SETTING FOR THE ANALYSIS ......OOOOOOOOO0.00.0.0.0 23 Methodology and SCOpe of the Analysis Validity of a Case Study Iv. ANAI’YSIS .....OOOOOOOOOOOOOOOOO0.000000000000000. 32 Urban Public Service Expenditures Police protection expenditures Fire protection expenditures Sanitation expenditures Parks and recreation expenditures Highway expenditures Effects of Independent Variables Social variables Wealth variables Capacity to Finance variables Effects of Suburbs on Central City Expenditures Economies of Scale V. CONCLUSIONS 0.0.0....0.0.0.0.......OOOOOOOOOOOOO. 60 Michigan Cities and Urban Public Services Implications for Planning Policies Summary and Recommendations for Further Study iii BIBLIOGRAPHY 0.0.0.0000.........OOOOOOOOOOOO...... 68 APPENDIX A. EXPENDITURES FOR 1970 IN MICHIGAN CITIES OF 10000 OR MORE 00000000000000.0000... 72 APPENDIX B. SOCIO-ECONOMIC CHARACTERISTICS OF MICHIGAN CITIES ....................... 75 APPENDIX C. SIMPLE CORRELATION AND REGRESSION ANALYSIS USED IN THE STUDY ............ 77 iv LIST OF TABLES TABLE 1. 2. 3. 4. 10. ll. 12. 13. 14. Municipal Expenditures in the U.S. in 1969-70 .......24 Simple correlation coefficients of Police Protection expenditures and Independent variables ..............35 Simple regression coefficients of Police Protection expenditures and Independent variables ..............36 Simple correlation coefficients of Fire Protection expenditures and Independent variables ..............38 Simple regression coefficients of Fire Pretection expenditures and Independent variables ..............39 Simple correlation coefficients of Sanitation expenditures and Independent variables ..............42 Simple regression coefficients of Sanitation expenditures and Independent variables ..............43 Simple correlation coefficients of Parks and Recreation expenditures and Independent variables....45 Simple regression coefficients of Parks and Recreation expenditures and Independent variables....46 Simple correlation coefficients of Highway expenditures and Independent variables ..............49 Simple regression coefficients of Highway expenditures and independent variables ..............50 Per capita expenditures of Public services in MiChigan NuniCipalj-ties 0000......00.00.00.00000000057 1970 Expenditures in.Michigan cities of 10000 or more .....OOOOOOOOOOOOOOOOOO......0.0.0.00000000000072 Socio-economic Characteristics of Michigan Cities....75 LIST OF FIGURES FIGURE 1. City Size Distribution in Michigan ................. 28 INTRODUCTION Yv,m i Cities are in an age of accelerating change; an increap sing migration of peOple from rural areas to cities is they main characteristic of the movement in communities. Another fact of urban life is technological change that raises s- tandards of living but often creates new challenges in pro- viding such urban services as hospital facilities, air pol- lution control, water supply and transportation. Another dynamic factor arises with increasing racial and economic disparities between central city and the suburb. In the face of these changes existing methods of p- lanning or performing public services must be continually appraised and modified to meet new circumstances. Urban communities,in general, suffer from the lack of overall planning. There are certain common services that all mu- nicipalities provide as centers of population, industry, and commerce; they include police and fire protection, traffic control, education, sanitation, street cleaning, and others. Economic activities in cities are becoming more service oriented and the demand for urban public ser- vices is constantly increasing A general deficiency in the understanding of public services is the determination of variables which affect expenditures. The relationships between socio-economic characteristics and expenditures of the municipality is clearly relevant in this determination. If we know the forces that tend to increase or decrease expenditures and think they are apprOpriate determinants, we can attempt to change by modifying the characteristics of the municipality or by making policies on the expenditures directly. Another question in public services that has long been interest to policy makers is whether or not there are economies of scale in the provision of these services. Al- though it is not the only consideration, it clearly is im- portant in determining whether consolidation or metrOpoli- ...___ ." tan growth is desirable. ’ In the following chapters, we will first introduce the main issues in the study of urban public services. In chapter two, a review of selected studies will be presented; following this, the data and methodology of the analysis and validity of the case study will be explained. Analysis of the expenditures, effects of the independent variables, economies of scale and effects of suburbs on the central city expenditures will be studied in chapter four. The last chapter will include conclusions and some recommmendations for further study. CHAPTER I ISSUES IN THE STUDY OF URBAN PUBLIC SERVICES Need for a Theory, Difficulties The appropriate line dividing public and private provision of goods and services has never been clear. Neverthless, certain general principles have been used by welfare theorists to justify public as opposed to private: supply. First, social, or collectively-consumed goods provide one of the best example for public supply. Social goods are goods which, once supplied, are available to all people, whether or not they have paid for themii They are “*also goods the consumption of which does not reduce the supply available to others:] National defense is usually given as the best example 6§3a pure social good. Another example at the local level, would be the control of air and water pollution. Goods such as these, once supplied to anyone, are equally available to all. Since people can- not be excluded from the enjoyment of social goods, welfare theorists argue that people will not engage in voluntary payments for them, which means that the market can not sa- tisfy such wants} The principal problem of a public expenditure theo- ry is to find some methed of efficiently allocating commu- nity resources between private and public provision of social goods. There are two related problems in this con- text. First, since there is no market by which preferences for social goods can be revealed, there is no guide which the government can use to calculate the required amount of goods, 1.6., the amount of resources to be withdrawn from private use for public purposes. SecOnd, social goods are consumed in equal amounts by all members of the community; therefore, there is no single most efficient solution to the complex problem of satisfying social wants. One different approach to the theory‘b public ex- penditure in an urban area is that of C.M. Tiebout. In it, “The consumer-voter may be viewed as picking that com- munity which best satisfies his preference pattern for pub- lic goods”, having been offered a range of choices among jurisdictions,each of which has its "revenue and expendi- 2 Thus the problem of getw ture patterns more or less set." ting individuals to reveal their preferences is solved, much as it is in the private market sector, provided that there are enough communities from which to choose and the other assumptions of Tiebout's theory hold. These other assumptions are: full mobility, including the absence of restraints associated with employment opportunities; full knowledge on the part of “consumer-votersfi; no intercommu- nity external economies or diseconomies associated with local public services; some factor limiting the Optimum x size of each community? given its set pattern of services and communities constantly seeking to reach or maintain this optimum size.3 All of these approaches represent efforts of for- mulating a public expenditure theory for urban services. Indeed, several students of urban public expenditure such as Siegel, Bahl, Wilensky, Brazer, etc., have mentioned the lack of an adequate theory in explaining the variations of expenditures from one urban area to the other? In 0p- position, they do argue also that in formulating these theories tHeir asumptions become unrealistic and many times they are no more than "an exercise in abstraction"? Brazer for example, in his article argues that "Tiebout's model can not be said to be even a rough first approximation of the real world. The most pressing fiscal problems of met- ropolitanism arise precisely because of the very factors he denies in his assumptions. Even if individuals had full knowledge of differences among communities in revenue and service patterns and were willing to move in response to them and their own tastes, income, zoning, racial and re- ligious discrimination and other barriers to entry to vap rious communities would restrict their mobility."6 Another fact is that families and individuals do extend their ac- tivities, in working, shopping, and playing across commu- nity lines,so that there is no clean-cut coincidence bet- ween one's place of residence and the place in which ser- vices are consumed and taxes paid. Employment opportunities do condition the choice of community of residence, par- ticularly for lower income families, and for all families commuting costs, like all transport costs, restrict choices. Another major limitation is that it seems impossible to separate clearly the demand from the supply side in an examination of the activity in the public sector. For example, the level and distribution of incomes can be vie— wed as a demand factor in that the quality and quantity of public services that a family desires is thought to be di- rectly related to their level of income. 0n the other hand large pr0portions of low income residents may necessitate more police protection. Even more complicated is the fact that higher income levels, generally mean higher revenue levels, giving the city government a greater capacity to supply public services. A Measurement of Urban Public Services [Government provides urbanites with tangible and in- tangible services. Direct government participation in rendering tangible services involves building and opera, ting public facilities. Among building or investment de- cisions are what plant to build; how, where and when to build it; among the Operating decisions are what quanti- ty and quality of services to render, how, where and whom they should be rendered? Our interest in public services is in those acti- vities whose objective is to satisfy urbanites' desires and thus enhance their welfare. Both pecuniary and non-pe- cuniary benefits of government services must be considered. The latter relate to cultural and artistic values of natu- ral and manemade beauty. Although in the study we are dealing with the measurable services, the prevalence of non-measurable services must be recognized. A few Services have a basic output unit with well— defined physical characteristics. The best example is water, where thecbasic output unit is a cubic foot of wa- ter delivered to the place of use having a certain set of socio-economic characteristics. Street cleaning, po- lice protection, fire protection services offer more or less the same degree of quality for different urban places. The municipal area served is the basic unit to be conside- red. Hospital services and education are not easy to measure because of the complexity and number of quality dimensions associated with the units; standards are very 'different from one municipality to the other. Peculiar services exclusive to few cities, such as museum, arbo- ratum, zoo,would not be appropriate for determinant anap lyses. For these peculiarities and for the difficulties of measurements, such services are excluded from the r- study.\; Iptrametropolitan Disparities Local government, like the firm, benefits not on— ly from its own actions but also from the actions of other governments. If no compensations is required, benefits resulting from actions of other governments may be clas- sified as positive externalities. .fThe nature and magnitude of the intrametr0politan externalities is one problem of the central city. The flight of higher income families and some industries to the suburbs has diminished the fiscal capacity of the central city. In addition, the suburban residents through an interaction with the core city, draw heavily on public services and multiply such city problems as traffic con— gestion and air pollution. The exporting of the tax base from the central city and the importing of service costs find much support in empirical research? Hawley found an inverse relationship between per capita expenditures by the central city and the prOportion of Standard Metropo- litan Statistical Area population living inside the cen- tral city. Brazer, in examining 1953 expenditure data of the central city and overlapping government units found evidence to support Hawley's work. More recently, Kee has concluded that spending by the central city is signi- ficantly and positively related to the ratio of the fringe area to the total Standard MetrOpolitan Statistical Area pOpulation.9 There are two obvious types of exploitation which can occur. Suburbanites may impose costs on the city in their role as workers-commuters; they may also impose costs by their use of the city's facilities for shopping, enter- tainment, and so forth. In both cases, the increased costs are likely to be in terms of roads, traffic control, police protection];0 A third exploitation is possible; this can be by discrimination. Central cities tend to have increasingly greater concentrations of low income, poorly-educated re- sidents, which increase demands for welfare, health, public housing and police protection, etc. To the extent that su- burbs force or reinforce this concentration by zoning re- gulations or discriminatory practices in real estate trans- actions, suburbs can be regarded as exploiding the central city}1 From the revenue side, commuters are likely to con- tribute in the following ways: non-residents income taxes, user fees; sales taxes, either directly or by the tax on output produced by commuters in resident industry and com- mercial establishments and increases in prOperty taxes from resident industries and commerce which employ commu- ters. Other non-resident users may indirectly contribute to the city's taxes by increasing the value of city pro- perty and also, directly contribute to taxes via sales taxes or user charges. lO . Economies of Scale KImThe question as to whether or not there are economies of scale in the provision of public services has long been of interest to policy makers. Although it is not the only consideration, it is important in determining whether met- r0politan growth in general, is desirable. As you will see in the next chapter,tsevera1 of the\ determinant studies have reported finding us observable e- conomies of scale. Scale economies exist when an increase in output is associated with a decline in the average cost per unit of output. This occurs whenever an increase in output allows for a more efficient combination of inputs 4 than was previously possible. Some capital equipment, for example, matfite extremely efficient if their utilization is very high. Thus, the negative relationship between per capita expenditures and population with scale economies is the following: as pOpulation increases, output must be in- creasing and if this increase is associated with a reduc- tion in cost (i.e., per capita expenditures), there is evi- dence of scale economies. However, this reasoning might be misleading. Population and per capita.expenditures may not be a good determinant for per unit costs. We could think that if there were scale economies present, we would not necessarily expect expenditures to decline. 0n the contrary, they may well increase instead if it happens that the demand for the particular service is elastic, i.e., 11 decreased costs per unit of output could lead to increased per capita expenditures for that particular service. {For the existence or absence of scale economies, it \- is necessary to estimate the long-run average cost func- tions for specific government services]:2 The difficulties in making such estimates are considerable. Defining out- put is the main problem to solve plus the costs of both inputs and outputs. Although output definition for sewe- rage or refuse collection is not to difficult; output for services such as police, fire, recreation is very difficult (er-uni l to define. KQuality differentials should be taken into ac- 1 (In—.4 count. “~. Sacks, in his article about scale economies}3 puts more emphasis on density rather than per capita expendi« tures and mentions: "...per capita measures do not show any regularities, may not provide a good method for projec- ting municipal expenditures in urban areas, the expendi- tures per square mile may in fact do so." Another factor that should be taken into conside- ration is the service level. ghen population increases, there is a shift in municipal services in quality, orga- sation and in process. At some instances, scale economies are absorbed by these shifts in the level of services. The introduction of indexes indicating quality and ser- vice levels may prevent this disappearence of scale econo- (I / mies./ / x””’* In spite of the conceptual and empirical problems 12 encountered in estimating scale economies there are quite a few studies done for municipal services. Hirsch studied cost functions for residential refuse collection for St. Louis City-County cities and municipalities. Nerlowe studied electricity and its supply and found evidence of significant scale economies. Will, using a different app- roach, based his study on engineering specifications which are related to service level and service requirements}4 I FOOTNOTES 1. See for a broader definition of social goods, R.Musgrave, The Theory of Public Finance, Mc Graw'Hill, New York T§§99pp06"'120 2. Charles M. Tiebout,"A Pure Theory of Local Expenditures," The Journalof Political Economy, 54(0ctober,l956),p.418. 3. The size at which its services can be provided at 10- west average cost. 4. B.N.Siegel, "0n the positive theory of State and Local Expenditures," Public Finance and Welfare, Kleinsorge Inc., (1965) pp.l75-185; R.W.Bahl, Mgtropolitan City Expendi+ tures, University of Kentucky Press (1968)p.69; G.Wilensky, Financing‘the Metropolis, Sage Publications (1970); H.E. Brazer. "Some Fiscal Implications of Metro- politanism," City and Suburb, Benjamin Chinitz,editor. Prentice Hall, (1970). 5. Brazer, Opus Cit.,p.132. 6. Ibid., p.133. 7. W. Hirsch, "The Supply of Urban Public Services," Issues in Urban Economics, John HOpkins Press, Baltimore, 9 8 p.477. 13 8. Amos Hawley,"Metr0politan POpulation and Municipal Go- vernment Expenditures in Central Cities," Journal of Social Issues (1951); H.Brazer, City Expenditures‘in the U.S., National Bureau of Economic Research occa- sionEI paper No.66,(l959). 9. Gail Wilenski, "Determinants of Local Government Ex- penditures," Financin the Metro olis, edited by J. Crecine, Sage pub. (1975) p.213. lO.Ibid., p.214. ll.Noll and Riew, Financing The MetrOpolis, Opus Cit., pp.481-5l5. 12.W.Z.Hirsch, "Cost functions of an Urban Government Service," Review of Economics and Statistics (Isbruany, 1965) pp.87:92. 13.Seymour Sacks, "Spatial and Locational Aspects of Local Government Expenditures," Public Expenditure Decisions in the Urban Community, John HOpkins Press (I9537p.180. 14.Hirsch, Opus Cit., p.90; M.Nerlowe, Returns to Scale. in Electricity Supply, Palo Alto Institute for Mathe- matical Studies in Social Sciences (1961); H.E. Will, "Scalar Economies and Urban Service Requirements," gale Economic Easays,v (Spring,1965). CHAPTER II REVIEW OF SELECTED STUDIES The purpose of this chapter is to review some of the most salient studies dealing with the relationship between municipal expenditures and socio—economic determinants. Interest in these determinant studies is a relatively re— cent phenomenon. Before 1950, the subject had been gene- rally ignored} since that time the number of article has been augmented. The intention, here, will be to review on- ly those articles or studies which are pertinant to the pre- sent.study. It is hOped that the review will contribute to a better understanding of problems associated with the s— tudy of the variation of municipal expenditures and their relationship with selected factors. British Studies The earliest study about municipal expenditures and services has been done in Britain. In 1910, C.A.Baker was interested in the cost of city management and population size. A second study on the question of municipal effici- cy and scale economies was prepared by Local Government Committee of the London County Council in 1914? Long after Baker's and London County Council's stu- dy, in 1942, H.S.Phillips and K.S.Lomax dealt with the sub- 14 15 ject of municipal expenditures within the context of econo- mies of scale. Phillips tried to determine what was effi- cient about municipal size in the light Of broad industri- al, social considerations. Lomax studying per capita ex- penditures, found that these are, primarily, a function Of the pOpulation? Studies in the United States before 1945 Two important studies have been prepared by Daven- port in 1926 and by Ridley and Simon, in 1938? The first was an analysis Of per capita Operational expenditures for local governments within the State of New York. Ridley and Simon, who undertook a major investigation of munici- pal activities measurement provide 8 striking exception. Rather than being an actual study in the expenditure pat- terns Of various municipalities this was a survey of sug- gested criteria for appraising city administrations.- Their Observations shed new light on the whole area of municipal activities. A group of economics, interested chiefly in public finance, has concentrated on the fiscal aspects Of munici- pal government? Mabel Walker's study attempts to get at the municipal scale economies. The emphasis is on the pro- portional distribution of the budget among the municipal departments of 175 cities above 30,000 pOpulation. Walker takes the average percentage of the total budget received 16 by each department in all cities and relates this to a few important influences, including time, wealth and population? The emphasis on actual expenditures has continued in contem- porary urban economists and gained much interest Of students of municipal activities after 1945. The best known re- searchers are the followings. Hawley Another new approach to the problem came about in 1951 with an article by Amos Hawley? The purpose of the Hawley's study was to test a hypothesis regarding the in- terdependence of populations lying within and without ur- ban centers. He was concerned with discerning what vari- ables had the greatest association with variations in mu- nicipal expenditures. The procedure he used to measure the degree of as- sociation was a correlation analysis, and in general Haw- 1ey's findings indicated "...that the municipal govern- ment costs of metropolitan centers vary with the sizes of their satellite pOpulations.....indeed, the association with satellite pOpulation is closer than with size of po- pulation in the cities concerned. That is true, moreover, of virtually every population variable employed, as well as Of such nondemographic factors as number of houses and housing density.” Hawley's paper raised many provocative questions, most of which have been considered further and its major 17 assumptions have served as a basis for other studies. Scott and Feder In 1957, the study for 196 California cities over 25,000 pOpulation by Scott and Feder8 consists Of a mul- tiple regression analysis of per capita expenditures. As independent variables they used per capita prOperty valu- ation, per capita retail sales, percent pOpulation increase and median number Of occupants in dwelling units. As scale diseconomies they found that expenditures tended to increase as pOpulation rose but at the same time the use of total, rather than departmental, expenditures limits the significance of diseconomies and obscures the relationship between specific services and city sociO-economic characte- ristics. Brazer The most comprehensive nationwide study on city ex- penditures was undertaken by H.E. Brazer who employed five different samples Of 1951 data; the large sample contained 462 cities, 5 smaller statewide groups, and a smaller num- ber of very large cities, including the overlying govern- ment unit. The analysis was made not only for total ge- neral Operating expenses, but also for police protection, highways, recreation, sanitation, general control and ot- hers. Among the independent variables tested were: pOpu- lation density, median family income, intergovernmental 18 revenues, pOpulation size, population growth rate and map nufacturing.- It was primarily the first three variables that were found to be statistically significant? Fabricant, Fisher, Sacks and Harris The abbve four students of public economy presen- ted fruitful expenditure determinant studies in a consecu- tive way]:O Using cross-section data for 1942, Fabricant found current expenditures of local governments strongly related to population density, urbanization, and income. Significant correlations were also found when these three variables were related to school, highway, public welfare, health and hospital, police, fire protection and general control expenditures. Fisher repeated the Fabricant ana- lysis with 1957 data, and found that the same variables no longer accounted for as much of the variation in spen- ding. Sacks and Harris modified the Fabricant approach by adding federal and state aid as additional indepen- dent variables. They found that level Of income and aid payments explained a large part of the variation in spen- ding, leaving the other variables insignificant. In 1964, Fisher categorized the determinants under three major headings: (l)economic variables-per cent Of families with less than $2000 income, and yield Of rep- resentative tax system as percentage of the U.S. average, (2)demographic variables-pOpulation density, urbanization, 19 percentage of population increase, (5)sociO-politica1 va- riables-index Of two—party competition, percentage of po- pulation over 25 years of age with less than five years schooling. Two of his conclusions were: "...variations in the state expenditures are the result of various political decisions: current decisions and decisions Of the past which have become embodied in constitutions, charters, statutes, ordinances, etc..... Level of expen- ditures, as measured by per capita, for 12 of the 15 categories are very significantly cor- 11 related with the seven independent variables." Hirsch Werner Z. Hirsch is, perhaps, the most productive student of urban public economy and services. His seve- 12 dealing with urban public economy and ser- ral studies vices supply, cost function studies for education and re- fuse collection, scale economies and government consoli- dation for public services provide broad explanation of the variations in governmental expenditures for specific functions. In his very recent study, he mentions: "...ex- penditure determinants studies, while not yielding bona fide cost functions, can advance our understanding of why expenditure levels differ among communities and among ser— vices. InSome cases, predictions based on these studies can turn out to be reasonably correct."13 The data which Hirsch compiled indicated that lar- ger cities spend more on governmental services and that per capita government expenditures increase with an increase 20 in population. However, neither economies or diseconomies of scale appeared and when correlation analysis was employed to measure the degree of association between pOpulation size and per capita expenditures it was discovered that pOpulas tion alone is not a particularly a strong factor in accoun- ting for variations in per capita spending for most services. Pidot, Bahl G. Pidot analyzed expenditures for the 81 largest metrOpolitan areas. His study is unique in that he used a principal component analysis to create six independent measures which were then assumed to descibe basic charac- teristics of a metropolitan area. He found that the "deg- ree of metropolitan development", "the level Of general wealth", "and an "index Of size" were important and in gene- ral were positively related to expenditures. He also found that state aid was significant for some functions, it was less important and less consistent in its effect}4 Roy W. Bahl did a study similar to Brazer's using 1950 and 1960 data for 198 central cities and in general adopted three different groups of determinants: demog- raphic, economic and financial ability. He found the same variables to be important as in Brazer's analysis}5 Data were analyzed cross sectionally for 1950 and 1960, and for the changes in per capita expenditures between 1950 and 1960. The conclusions were, in general, quite simi- 21 lar to those reached in the earlier studies: the level Of per capita central city expenditures is closely related to the size Of the central city population, relative to that of the entire Standard MetrOpOlitan Statistical Area. 2. 3. Social Issues (1951) pp.100-108. FOOTNOTES Up to 1950, studies were concerned with economies of scale, but later the effects Of socio—economic charac- teristics have been recognized. See C.A. Baker, "Population Costs in relation to City Management," Journal Of the Royal Statistical Society, (December,l9l0) p.73-79; and Com arative Munici a1 Statistics 1912-15, London County CounciI,(I915§. Hugh S. Phillips, "Municipal Efficiency and Town Size" Journal of the Town Flagging Institute, XXVIII(May-June, 1943); and K.S.’Lomax, "The Relationship between Expen- diture per head and Size of the po ulation," Journal of the Royal Statistical Society, CVIF1943) PP.5I-59. Clarence E. Ridley and H.A. Simon, Measuring Municipal AOtiVities, International City Managers Assn.,(l938). Gerhard Colm, "Public Expenditures and Economic Struc- ture in the U.S.," Social Research Vol.5 (February 1936); and(Hanscn and Perloff, State anerocal Finances, Norton Co. 1944 . Mabel)Walker, Municipal Expenditures, John Hopkins Press 1950 . Amos Hawley, "Metropolitan population and Municipal GO- vernment Expenditures in Central Cities," Journal of Stanley Scott and Edward L. Feder, Fgctors associated with Variations in Municipal ExpendituregLevels, Uni- versity of California Bureau Of Public Administration (February 1957). 22 9. Harvey E. Brazer, City Expenditures in the U. S.,Natio— nal Bureau of Economic Research, occasional paper No.66 (1959). 10. Solomon Fabricant, The Trend of Government Activity in the U. S. since 1900, National Bureau of Economic Research Inc. (1952); Gleen W. Fisher, "Determinants Of State and Local Government Expenditures," National Tax Journal, XIV (December 1961) pp. 549- 555; S. Sacks and R. Harris, "The Determinants of State and Local Government Expendi- tures," National Tax Journal XVII (March 1964) pp. 75-85. 11.Fisher, Opus Cit., p.354. 12. W. Z. Hirsch, "The Supply of Urban Public Services," in Issues in Urban Economics, John HOpkins Press (1968) pp 477-527, and*"EXpenditure Implications of Metropoli- tan Growth and Consolidation, " Rewiew Economics and Statistics (August 1959) pp. 232-211. 15.Hirsch, Issues in Urban Economics, Opus Cit., p.501. l4.G. Pidot, "A principal Component Analysis of the Deter- minants of Local Government Fiscal Patterns," Review of Economics and Statistics (Hay 1969) pp.176- l5.R.W. Bahl, Metro Olitan City Expenditures, University of Kentucky Press (1969). CHAPTER III SETTING FOR THE ANALYSIS Before going into Michigan's Municipal expenditures- let us have a brief look to the United States' aggregate budget.in order to have an idea of the distribution of reve- nue for diverse municipal functions. Revenue of all city governments during 1969-70 to- taled $32.7 billion, up 33 billion or 10.2% from the previ- ous year total. City expenditures totaled $34.2 billion in 1969-70, as against $50.5 billion in 1968-69. General expenditure-i.e., spending other than for utility and em- ployee retirement purposes- totaled $27.7 billion in 1969- 70 for the U.S. General expenditure for state and local government in 1969-70, totaled $190.8 billion. Thus, mup nicipal expenditures constitute, approximately, 80% of the total general expenditure.1 The five general expenditure groups that were stu- died: police and fire protection, parks and recreation, highways, sanitation represent 34.9% of the expenditures in the U.S. average for 1969-70. In the next page, you will find a summary of functional distribution of muni— cipal governments general expenditures for fiscal year 1969‘700 23 24 Function Amount Percent Per cap. ,Lmillions t) dollars Total General Exp. 27.6 100.0 209.8 Education 4.5 16.4 34.4 Police protection 2.9 10.8 22.7 Highways 2.4 9.0 18.9 Fire protection 1.7 6.4 13.3 Sewerage 1.4 5.3 11.0 Public welfare 2.2 8.0 16.7 Hospitals 1.4 5.3 11.1 Parks and Recreation 1.5 4.7 9.9 Sanitation other than sewerage 1.0 4.0 8.5 Others 8.5 50.1 62.9 Table‘l. MUNICIPAL EXPENDITURES IN THE U.S. IN 1969-70 25 Methodology and Scope of the Anatxsig Intercity variations in per capita expenditures of municipal services are studied in the following chapters using a cross-section approach with 1970 data. Some re- cent studies have used county aggregates, placing the em- phasis on the differences in per capita expenditures as mong county areas and ignoring overlapping political units? This approach is less useful to the planner because it treats the expenditures of city and county governments as aggregates. The problems central to this study are direct- ly related to the problems of coordinating fiscal and phy- sical planning, and the objective of the statistical analy- sis is to identify, where possible, those factors that con- tribute to differential per capita expenditures among city governments. Although the analysis centers on the actual expenditutes of city governments, it is based on a recogni- tion that economic and social areas, not corporate boundar ries, represent the most appropriate planning units. Some early students of municipal expenditures poinp ted out that the administration of a governmental unit, no matter how small involves some over-head costs which are relatively fixed and unavoidable? They suggested that incorporated municipalities required a minimum of 10,000 persons for efficient performance of municipal services. For this reason, and because of the lack of data for smal- ler communities, this study involves only cities with 26 10,000 population or more. All of these municipalities are 82 in total but the lack of 1970 data for socio-economic characteristics limited us to a number of 58 cities for consistency in the analysis. Thus, the purpose of the following analysis is to investigate primarily the expenditure determinants of se- lected urban public services in 58 municipalities of the State of Michigan. While identifying these socio-economic determinants and their correlations we will search for any observable economies of scale. As you will observe in the analysis part of the stu- dy the expenditures of five different groups of city have been taken into consideration. First, cities larger than 10000 which include 58 municipalities in Michigan. Second, cities larger than 50000 representing the aspects of middle size cities because they are different from small cities by their economic, social and political structure? Third, eight cities larger than 100000 represent the metropolises of Michigan which are: Ann Arbor, Dearborn, Detroit, Flint, Grand Rapids, Lansing, Livonia, Warren. At the fourth sample we tried to isolate the "single" or independent ci- ties which are simply those that are not located within a standard metrOpolitan area and stand by themselves without any dependence to another city. These 28 isolated cities are: Adrian, Alpena, Ann Arbor, Battle Creek, Bay City, Benton Harbor, Detroit, Escanaba, Flint, Grand Rapids, Hol- land, Jackson, Kalamazoo, Lansing, Marquett, Midland, Monroe, 27 Mt. Clemens, Mt. Pleasant, Muskegon, Niles, Owosso, Pon- tiac, Port Huron, Saginaw, Sault St. Marie, Traverse City, Ypsilanti. As the fifth sample we took suburban cities; these are those cities located within a standard metropolitan statistical area but is not a core city. These, respec- tively are: Allen Park, Berkley, Birmingham, Center Line, Dearborn, E.Detroit, E.Lansing, Farmington, Ferndale, Garden City, Grosse Point Woods, Hamtramct, Harper Woods, Hazel Park, Inkster, Lincoln Park, Livonia, Muskegon Heights, Plymouth, Roseville, Royal Oak, St. Clair Shores, Southfield, Southgate, Trenton, Troy, Warren, Wayne, Wyandotte, Wyoming. All of these suburban cities are 50 in total. After these categories of observation for variation of expenditures in different structures of city, several groups of city have been examined for economies of scale in each 10000 population bracket. Iigure an the next page summarizes the distribution of different city sizes. The data restrictions diminished the number of cities within 10000-20000 population bracket. We took this into consi- deration in our analysis while making generalizations. Another sampling was for the measurement of subur- ban exploitation on the central city services which we will explain in the next chapter. This sample consisted of the expenditures of central cities and socio-economic characteristics of its surrounding areas such as: 28 Number of Cities 281 24~ 20* Actual Distribution of Cities larger than 10000 16- ~——w 12- 8- __“__ 4. ' Po ulation (t ousands) 10 5O 50 7O 90 110 200 4r Nuber Of Cities 241 20« 16. Distribution Analyzed in the Study 12~ 8. 4‘ Papulation (thousands) O _.J_J__J__P:l ] 1“} r1 I I L‘1o 30 5 70 90 110 200 " Figure 1. City Size Distribution in Michigan 29 Detroit-E.Detroit Lansing-E.Lansing Detroit-Livonia Gr. Rapids-Wyoming Detroit—Dearborn Muskegon—Muskegon He. Detroit-Grosse Pt.Woods Pontiac-Birmingham The statistical technique used in this analysis is the simple correlation between per capita expenditures of public services and socio-economic characteristics of ci- ties. Along with this search of linear relationships among variables, we also undertook a linear regression analysis in an attempt to determine the degree of predic- tibility of the independent variables? The terms "independent variable","determinant", "socio-economic characteristics" are used interchangeably. In general, where simple correlation coefficients are sub- ject to analysis "independent variable" or simply "variable" is the preferred term; when analyzing regression coeffici- ent "determinant" or "proxy variable" has been used. Validity of a Cross-section Case Study Most of the students of urban public services con- centrated their efforts in Interstate Variations of Expendi- tures. Roy W. Bahl, in his book? suggests that: "One method of approach to certain of these questions is to abandon the macro statistical approach in favor of intensive case stué dies of specific states, or, better yet, metropolitan are- as." Hefpmther mentions that: "The advantages of a case study are numerous: a.quality variations within a given 30 metrOpolitan area or state are smaller; b. externalities associated with the public sector such as the urban-suburban exploitation hypothesis, mat be examined more intensively; c. more accurate data for longer periods of time may be col- lected from local sources; d. the problems in the data crea- ted by differing intergovernmental fiscal arrangements may be eliminated by confining the analysis to a particular state or Standard Metropolitan Statistical Area." The primary advantage of limiting the analysis to ci- ties within a state is that it eliminates much of the statis— tical"noise" due~to differences in functional responsabili- ties, and to histOrical or geographical peculiarities. Gail Wilensky mentions: "...case study apparently limits the.ge- nerality of the findings. This is unfortunate, but if the determinants of local expenditures are, in fact, specific to a particular state this information is useful."7 31 FOOTNOTES City Government Finances in 1969-70, Bureau of the census U.S. Dept. of Commerce, (1971) H.J. Schmandt and G. Ross Stephens, "Measuring Munici- pal Output," National Tax journal, (December, 1960) PP.369-375. Hansen and Perloff; Opus Citation, see p.21. W. Thomson, Introduction to Urban Economics , John Hop- kins press.(l965) p.24. Heath and Downie, Basic Statistical Methods, Harper and Row Inc.,(l970) chapter 7; and see Appendix C for more information about statistical method used in the analysis. Roy W. Bahl,(l969) p.157; Opus Citation see p.21. G. Wilensky,(l970) p.202; Opus Citation see p.15. CHAPTER IV ANALYSIS ”V‘ Urban Public Services Expenditures (“In the analysis capital outlays are eliminated, ope- rating expenditures are expressed on a per capita basis except for police and fire protection expenditures. These are more relevant than the aggregated expenditure figure, because it measures the normal day to day expenditures on a per capita basis. The five common functions those supported more or ‘ less to the same extent by all cities included in the stu- dy are: police protection, fire protection and other current expenses for sanitation, parks and recreation, and highways services. These municipal functions are expressed on indi- vidual bases to analyze more clearly the effects of the explanatory variables. It should be noted that any comparison of dollar fi- gures may not represent the true variation in per capita service among cities. Factors such as wage-rate variations and quality differentials may hide the true intercity diffe- rences in per capita levels;:]These differentials are maxi- mum when comparing cities' expenditures of different states. In our study concerning with one state this variation is 32 33 minimum; however, such a variation even very small exists. Intercity comparisons of quality levels is beyond the scope of this work, although adjustment for quality differentials would greatly enhance the interpretation of the results of any interarea analysis of public expendi- tures} _-._-/”.1 T‘s r ‘ 7 I / ZkPolice protection expenditures/J The police protection expenditures category includes both current and capital expenditures for the preservation of law and order and traffic safety, including highway po- lice patrols, crime prevention activity, police communicap tions, detention and custody of persons awaiting trial and the like? The average amount spent per city resident for police protection for 58 cities of Michigan was $ 22.56 in 1970. As outlined on the following pages the correlated variables are: pOpulation density, total retail sales, per- centage of negro head of household, intergovernmental reve- nue and property taxes. Total retail sales, intergovern- mental revenue and prOperty taxes variables are the most prevalent since their correlations exist in every set of observation. Percentage of negro is the second most sig- nificant variable; it is highly correlated (.9026) with the expenditures of cities bigger than 100,000. This re- sult is consistent with the general trend of negro popu- lation lecatiOn since the average percentage of negro 34 head of household living in big cities is the highest (9.96%), comparing to the average of the 58 cities which is 4.48%, the lowest. Population density variable is significant in single (isolated) cities and cities bigger than 100000 pOpulation. ([The importance of the non-white pOpulation and density in explaining variations in police expenditures may result from the relatively lower economic status of residents in the more crowded urban areas or of the neg- ro population. Further, higher population densities may lead to greater vehicular and pedestrian traffic control problems, thereby requiring a higher level of per capita expenditure for police protection:7 un- .‘ 35 OBSERVATION GROUPS CITIESa' CITIESBI 0111385 SINGLEd SUBURBANE- VAR]; ABLES )10 .000 > 50 . 000 > 100 .000 CITIES CITIES Population ' Density .2214 .3359* .7611* .5917* .2054 % of Household income ($30000 00748 ' 0551-2" .6078 -00892 00857 Total Retail 38133 0307” 0.7318" 0788]." 0468” 03499” Household Buying ' ‘ Income. .1725 -.5011 -.5402 -.2768 -sll68 Median.Home ' ' valuQB 01654 “02522 '07790* -01310 -01792 of N8 0 Head V, if Housgold 04942.. 08754" 09026. 081-93" 0 2978 4 ” Inter overnmenp tal Rgvenue 3058* . 6884* .79.14* .4465 5894* P t T roper y axes .5061* .6851* .7492* ..4286 ’.5912* 8.58 cities in total. b.18 cities in total. 0.8 cities in total:Ann Arbor,Desrborn,Detroit,Flint,Gr.Rapids, Lansing,Livonia,Warren. d.28 cities in total,exc1ude all Detroit suburbs and some depens dent cities such as E.Lansing,Muskegon Heights,etc. e.50 cities in total,include all Detroit suburban cities and some dependent cities mentioned above. * Denotes significance at the .05 level; Table 2.SIMPLE CORRELATION COEFFICIENTS OF POLICE PROTECTION EXPENDITURES AND INDEPENDENT VARIABLES. 56 OBSERVATION GROUPS 'SUBURBKEE' CITIES“ CITIES 5: CITIESc SINGLEa VARIABLES >10 .000 ) 50 .000 > 100 .000 CI TIES CITIES Papulation .0006 .0009 .0018 -.0019 .0005 Density (10.86) S of Household .0686 .6146 .7561 -.2035 .0846 incomet($5.000 (12.71) (8.96) (10.99) (8.23) Total Retail .0000 .0000 .0000 .0000 .0000 Sales ' Household Buying -.0004 -.0009 -1.1030 -.0011 -.0002 Income - (1257) Median Home -.1675 -.2417 ’4244 -.2293 -.1850 Values (10.56) (10.20) (18.85) (9.53) (10.23) % of Negro Head .2725 .4664 .0001 .4595 .1624 h of Household (19.15) (15.69) (15.27) (2176) Intergovernmene .0001 .0001 1 .0001 .0001 .0001 tal Revenue~ .' = ' Pr°P°rtY Taxes .0001 .0001 .0001 .0001 .0001 a.58 cities in total. b.18 cities in total. 0.8 cities in tota1:Ann Arbor,Dearborn,Detroit,Flint,Gr.Rapids, Lansing,Livonia,Warren. d.28 cities in total,exclude all Detroit suburbs and some depen- dent cities such as E.Lansing,Muskegon Heights,etc. 8.50 cities in total,include all Detroit suburban cities and some dependent cities mentioned above. Table 5. AND INDEPENDENT VARIABLES SIMPLE REGRESSION COEFFICIENTS OF POLICE EXPENDITURES (Standard error of estimate is in parenthesis where regression coefficient is significant) 37 Fire protection expenditures Per capita fire expenditures include current out- lays for "fire-fighting organization and auxiliary ser- vices thereof, inspection for fire hazards, and other fire~. prevention activities. Also included are costs of fire fighting facilities, such as fire hydrants and water, fur- nished by other agencies of the city government.”2 Average per capita fire protection expenditure in 1970 within 58 cities of Michigan is 315.45. Significant correlations between the city expendi- tures groups and other socio—economic variables are summap rized on tables in pages 58, and 59. Expenditures in 01- ties bigger than 100000 exhibit higher correlation and regression coefficient with fire expenditures than smaller cities. 0n the contrary, suburban municipalities do not exhibit any significant conrelation. This is, perhaps, due to the diversity of the single city functions such as; dormitory, industrial, touristic cities, etc. In general, the most important variables seem to be the wealth variables: percentage of households having an income of less than $5000 and buying income per house— hold. The former positively but the latter is negatively related to the expenditures. This negative relationships exists also for the median home values and fire expenditures which means that rich municipalities spend little for this servi0e. This correlation becomes stronger in big central 38 OBSERVATION GROUPS CITIESa CITIESB: CITIE§ SINGLE“ SUBURBAN;- VARIABLES )10 .000 ) 50 .000 7 100 .000 CITIES C I TIES Population Density L 3.0590 -.l592 -.l560 .0395 .2996 % of Household income (35.000 4220* .7245. .8925* .2120 .1686 Total Retail 38.188 .0382 01106 01835 -00658 -00556 HO sehold Bu in - Ingome y 8 -03316* '04254* -08678* -06168* 00459 Median Home ' Values -04058* -03344 -0804? -04739" -00952 of No o‘Head -- . ff Housggold .1777 .4102 .4865 .1912 -.0525 £23°§§3§§§:“°"' .0070 .0666 .1142 -.1125 -.2953 Property Taxes .0110 .0506 .0554 -.1026 .0444 1 a.58 cities in total. b. 18 cities in total. 0.8 cities in total: Ann Arbor, Dearborn,Detroit, Flint ,Gr.Rapids, Lansing,Livonia,Warren. d. 28 cities in tOtal, exclude all Detroit suburbs and some depenp dent cities such as E .Lansing,Muskegon Heights, etc. e.50 cities in total,include all Detroit suburban cities and some dependent cities mentioned above. # Table 4. * Denotes significance at the .05 level SIMPLE CORRELATION COEFFICIENTS OF FIRE PROTECTION EXPENDITURES AND INDEPENDENT VARIABLES. 39 OBSERVATION GROUPS SUBURBAN“— CITIESa CITIESF CITIESc SINGLEa VARIABLES )10.000 ) 50.000 ) 100.000 CITIES CITIES Population Density -' 00001 -00003 '00002 00000 .0007 % of Household .3583 .3313 .1343 .3333 .1462 incomet($5.000 (7.85) (5.20) (4.74) (5.10) (9.95) Total Retail .0000 .0000 .0000 .0000 .0000 Sales 30118911016. Buying ’00007 “00001 -0002]. -00017 .0000 Income - Median.Home ;.3807 -.2900' ‘-.7758 -.5720 -.0866 Values (9.24) (8.51) (8.18) (7.95) % of Negro Head .0907 .1978 .1554 .0759 -.0288 f of Household (12.47) (12.98) Intergovernmeny .0000 .0000 .0000 .0000 -.0018 tal Revenue . 7 ' Pr°PertY Taxes .0000 .0000 .0000 _.0000 .0000 1 a.58 cities in total. b.18 cities in total. 0.8 cities in totalenn Arbor,Dearborn,Detroit,Flint,Gr.Rapids, Lansing,Livonia,Warren. d.28 cities in tota1,exclude all Detroit suburbs and some depen~ dent cities such as E.Lansing,Muskegon Heights,etc. e.50 cities in total,include all Detroit suburban cities and some dependent cities mentioned above. Table 5 . SIMPLE REGRESSION COEFFICIENTS FOR FIRE EXPENDITURES . AND INDEPENDENT VARIABLES. (Standard errors of estimate are in parentheses where coefficients are significant) 40 cities (regression coeffficient is .7343 with standard error 4.74) which are mostly the settlements of poor peOple and lower grade housing. Sanitation expenditures The per capita sanitation expenditures include Ope- rating expenditures for sewage disposal, street cleaning, waste collection, and payments to other local governments for such services. Sanitary engineering, smoke regulation, and expenditures for other health activities are not inclu- ded in the analysis. For 51 cities of Michigan per capita sanitation current expenditures are distributed about a mean of $7.59. This average is slightly higher than 1960 average of 198 cities of the United States which was $7.40.3 Among the eight determinants, percentage of Negro head of household is the most significant one for five dif- ferent observations. The highest correlation, .57 at the 5% significance level (see page 42), is with cities having a population more than 50000. The regression coefficient is .2616 with a standard error of 6.06. The importance of this variable is not at the same degree for other observsp tions, it is significant however, for single isolated met- ropOlitan cities and cities bigger than 100000. As you might examine from the table in the page‘42, the‘other most important variables are intergovernmental revenue and prOperty taxes. The highest correlation (.6016 and .6425) 41 exists again between these variables and expenditures of cities bigger than 50000, the regression coefficients are nor very high, however. This means that the explanation as a proxy variable is less stronger than in the previous observations. ‘ Total retail sales is another important variable' which does Show some correlation with expenditures of su- burban cities. This is consistent with the existing con- ditions, for there‘is relatively less commercial activity in these suburban cities than the single central cities. Bahl and Brazer, in their study have found that per capita expenditures are positively related to DOPUIEP tion density? In our study we found also important relay tionships to population density with .5330 correlation coefficient except in the observation of suburban cities' expenditures. The absence of correlation in this case probably stems from the diversity Of settlement patterns, tastes of the residents and less densily populated charac- ter of these cities. The high correlation between density and single central cities proves this hypothesis. 42 OBSERVATION GROUPS 1 a.58 cities in total. b.18 cities in total. 0.8 cities in tota1:Ann Arbor,Dearborn,Detroit,Flint,Gr.Rapids, Lansing,Livonia,Warren. d.28 cities in total,exclude all Detroit suburbs and some depenp dent cities such as E.Lansing,Muskegon Heights,etc. e.50 cities in total,include all Detroit suburban cities and some dependent cities mentioned above. * Denotes significance at the .05 level fi—v , CITIES8L CITIESF CITIES‘5 SI NGLEa SUEUREEa VARIABLES )10.000 > 50.000 > 100.000 CI TIES CITIES Population . 7 i , on Density 3 53 533 .4285 .4652* .3228 % of Household -.1791 -.1268 _ - income1($5.000 .1698 .5025 .1900 Total Retail .3242* .5973* .8580* .4465* -.0598 Sales Household Buyine .0554 .1855“ -.0103 .0576 .0547 Income. - Median.Home .1200 .1467' 1.1132 .2530 .0102 Values % of Negro Head .1267 .5771* .5313; .4504i -.5802*i of Household Intergovernmenp ' .3569* .6016* .8521* .4696* -.1087 tal Revenue 1‘ J ' Property Taxes 1 .3876* .6423* .3751» .4897* .1287 Table 6. SIMPLE CORRELATION COEFFICIENTS OF SANITATION EXPENDITURES AND INDEPENDENT VARIABLES. 43 OBSERVATION GROUPS CITIESa CITIES5 CITIES: SINGLEa SUBURBANr VARIABLES )10.000 > 50.000 > 100.000 CITIES CITIES 1 . . 33:13:31” .0007 .0012 .0009 .0011 .0004 r H h 1d 330,683,300 -.1262 ,-.1203 .1795 -.5131 -.1717 (6.17) (6.54) (6.90) (5.75) Total Retail .0000 .0000 .0000 .0000 .0000 Sales Household Buying .0001 .0004 --.0000 .0001 .0000 Income. - Median Home .0800 .1196 ..:.1599 .5525 .0056 Values (5.39) (5-75) (5.09) at of Negro Head .0500 .2616 .2796 .1895 -.169.0 a of Household (6.06) (6119) (6.16) (8.98) Intergovernmen~ .0001 .0001 . tal Revenue 00901 .0001 -.0004 Pr°Perty T353“ .0001 .0001 .0001 .0001 .0001 fi— e.58 cities in total. b.18 cities in total. 0.8 cities in totalenn Arbor,Dearborn,Detroit,Flint,Gr.Rapids, Lensing,Livonia,Warren. d.28 cities in total,exclude all Detroit suburbs and some depenp dent cities such as E.Lansing,Muskegon Heights,etc. e.50 cities in total,include all Detroit suburban cities and some dependent cities mentioned above. Table 7. SIMPLE REGRESSION COEFFICIENTS OF SANITATION EXPENDITURES AND INDEPENDENT VARIABLES. (Standard errors of estimate are in parentheses where coefficients are significant) 44 Parks and recreation expenditures The parks and recreation expenditures include Ope- rating expenditures for cultural-recreational activities, organized recreation, swimming pools and bathing beaches, municipal parks, and special recreational facilities such as sport arenas, recreation piers, skating rings, golf courses, playgrounds and yacht harbors. Expenditures for cemeteries are excluded because the irregularities of the sizes and expenditures as well as special functions such as auditoriums, museums and state park maintenance. The most common variable has been total retail sales for three observations. The other most significant vari- ables are: median home values, intergovernmental revenue, property taxes and percentage of household income less than $5000. Median home values and expenditures have a cor- relation coefficient of .4444 and a regression coefficient .6219 with a standard error of estimate 4.89 in in single or isolated cities. The same is not true for other Obser- vations. The consistency of data is limited in this func- tion therefore there are not much correlations in the egg— regate data. Sales or volume of commerCial activity may lead to an acquisition and maintenance Of more Open space. Quality differentials in maintenance Could be explained by the correlation of median home values and expenditures for single (isolated) cities. 45 OBSERVATION GROUPS SUBURBANE CITIESa CITIEST’: CITIE? SINGLEa VARIABLES )10 .000 ) 50 .000 > 100 .000 CITIES CITIES Population . Density -.l510 -.2051 -.5790* .2523 -.3384 % of Household income (35.000 -.0010 .1555 .0156 -.3737* -.O920 Total Retail ° .- Sales .5550* .2905 .1126 .5868* .5074* Household Buying - Income. .0452 .1904 .3591 .1640 .1594 Median Home Values .1146 .1701 .2119 .4444* .1792 % of Negro Head .-. of Household .1376 .2125 -.0558 .1825 -.1911 Intergovernmen- tal Revenue .2557* .2197 .0815 .5265 .4929* Pro ert Taxes P y .3050* .2556 .1199 .3271 .6462* a.58 cities in total. b.18 cities in total. 0.8 cities in tota1:Ann Arbor,Dearborn,Detroit,Flint,Gr.Rapids, Lansing,Livonia,Warren. d.28 cities in total,exclude all Detroit suburbs and some depen- dent cities such es E.Lansing,Muskegon Heights,etc. e.50 cities in total,include all Detroit suburban cities and some dependent cities mentioned above. * Denotes significance at the .05 level: Table 8. SIMPLE CORRELATION COEFFICIENTS OF PARKS AND RECREATION EXPENDITURES AND INDEPENDENT VARIABLES. 46 OBSERVATION GROUPS CITIESa CITIESb CITIES‘3 SINGLEa SUBURBANe VARIABLES )10.000 >50.000 - >100. 000 CITIES CITIES Population Density ‘00003 ‘00006 "00012 00096 ‘00009 % of Household ' -.0008 . .216 .02 O -.6811 -.0 88 Total Retail Sales .0000 .0000 .0000 .0000 .0000 3°“Beh°1d Buyind .0001 .0007 .0017 .0005 .0003 Income - _ Median Home 1023 .2059 '.5965 .6219 .1681 Values (6.06) (8.01) (9.52) (4.89) . (5.98) % 0f N°8r° Head .0200 .1428 -.0334 .0818 -.1050 of Household (11.57) Intergovernmenr .0001 .0000 .0000 .0000 .0029 tal Revenue -' Property Taxes .0001 .0000 .0000 .0000 .0010 4 1 a.58 cities in total. b. 18 cities in total. 0. 8 cities in total: Ann Arbor, Dearborn,Detroit, Flint ,Gr.Rapids, Lansing,Livonia,Werren. d. 28 cities in total, exclude all Detroit suburbs and some depen- dent cities such as E. Lansing,Muskegon Heights, etc. 6.50 cities in total,include all Detroit suburban cities and some dependent cities mentioned above. fi—_.. Table 9. EXPENDITURES AND INDEPENDENT VARIABLES. (STANDARD ERRORS OF ESTIMATE ARE IN PARANTHESES WHERE COEFFICIENTS ARE SIGNIFICANT)‘ SIMPLE REGRESSION COEFFICIENTS OF PARKS AND RECREATION 47 Highways and streets Per capita highway expenditures include Operating expenses for major or local streets, maintenance and other auxiliary services such as: guard rails and posts, sweeping and flushing, tree trim and remowal, traffic signals and signs, pavement markings, snow and ice removal? The per capita highway expenditure in 55 cities are distributed about a mean $12.24 with a standard deviation 5.96. The significant correlations between highway expen- ditures and socio-economic variables have been in population density, median home values and percentage of household in- come less than $5000 in two Observations: cities bigger than 10000 and single(isolated) cities. These negative correlations may mean either (l)that higher densities ref- lect lower ability to pay, which result in lower per capi- ta expenditures on local roads and streets, or (2)that higher densities reduce the physical mileage per person that must be maintained and therefore per resident expendi- tures are lower. The negative relation with median home values variable and its high regression coefficient (.6515 with standard error 8.44) defines it as a second determi- nant. The latter indicates that residents of higher in— come cities both demand and can afford a higher level of highway services but these services are offered in places where home values and income are low. The average of me- dian home value in 28 single (isolated) cities 815.55 is 48 the lowest average. This result suggests that highway maintenance expenditures are more intense in isolated cities than the suburban ones and consequently higher level of Operating expenditures are spent for streets and highways of single, isolated cities. 49 OBSERVATION GROUPS CITIESa CITIESB: CITIESE * SINGLEa SUBURBAN:- VARIABLES )l0.000 > 50.000 > 100.000 CITIES CITIES §2£§I$§1°n s.4712* -.2455 -.o75O 1,5333. -.l786 InggngESEICCC .3363* .3349 .0697 .0174 .0091 SZI2: Retail .0974 -.o47o -.l585 -.2509 .1744 111333;?“ Buying «2323* -.1814 .2332 -.2586 .1070 I $23322'3°m° -.5558* -.2953 4.1241 -.3874* -.o555 §f°§o§§§§31§ead .0251 .1275 -.0813 -.5447 .2747 —- igfegsszgfigmen' -.lOO6 -.o947 -.1732 -.2241 -.0356 Pr°perty Taxes -.O706 -.0218 -.1193 -.2021 .3161 I a.58 cities in total. b.18 cities in total. 0.8 cities in total: Ann Arbor, Dearborn,Detroit, Flint ,Gr.Rapids, Lansing,Livonia,Warren. d. 28 cities in tOtal, exclude all Detroit suburbs and some depen~ dent cities such as E. Lansing,Muskegon Heights, etc. 8.50 cities in total, include all Detroit suburban cities and some dependent cities mentioned above. * Denotes significance at the .05 level; Table 10 0 AND INDEPENDENT VARIABLES. SIMPLE CORRELATION COEFFICIENT 0F HIGHWAY EXPENDITURES 50 OBSERVATION GROUPS CITIESEl CITIESB: CITIES“ SINGLEat SUBURBAN? VARIABLES )10.000 > 50.000 )100.000 CITIES CITIES Population ._ ' ‘ , .. .0001 Density .0011 .0005 0003 0016 g of Household .3240 .2331 .0105 .0369 .0713 income (33.000 (6.92) (9.5.8) 15.31 Retail .0000 .0000 .0000 .0000 .0000 Sales ' Household. Buyine .0004 .0004 .0001 .0010 .0007 Income. - “ Indian Home _.3257 -.2502 .0234 -.6313 -.1485 Values ~ (8.37) (7.19) (8.44) (7.06) it of Negro Head .0126. .0600 .1149 -.1799 .0320 of Household (10.3) (14.2) Intergovernmenp .0000 .0000 .0001 .0000 .0000 tal Revenue 1 Property Taxes .0000 .0000 .0004 .0000 .0000 a.58 cities in total. b. 18 cities in total. 0.8 cities in totelenn Arbor, Dearborn,Detroit, Flint ,Gr.Rapids, Lansing,Livonia,Warren. d.28 cities in total,exclude all Detroit suburbs and some depen- dent cities such es E.Lansing,Muskegon Heights,etc. 0.50 cities in total,include all Detroit suburban cities and Table 11. some dependent cities mentioned above. SIMPLE REGRESSION COEFFICIENTS OF HIGHWAY EXPENDITURES AND INDEPENDENT VARIABLES. (Standard errors of estimate are in parentheses where coefficients are significant) 51 Effects of SociO-economic Variables Social Variables Population density and the percentage of negro head of household are the social variables we adopted in our study. Brazer found that population density was a highly significant factor in measuring the variation in per capi- ta Operating expenditures among 462 cities of more than 25000 and achieved similar results when analyzing the for- ty largest cities.6 The results of a two variable correlation analysis indicate that higher pOpulation densities are significant- ly associated with higher per capita expenditures for three expenditure categories. Parks and recreation and fire pro- tection expenditurres did not have any significant corre- lation except the former service has a correlation in ci- ties of 100000. Most correlations are with single cities where density.variable follows a certain pattern of Varia- tion whereas in suburban cities irregularities in the V89 riation of density do not lead to a correlation. Second variable, the percentage of negro head of household is related to police eXpenditures and sanitas tion expenditures. The strongest relation (.9026) is found in cities 100000 or more; this result is consistent with the actual location of negro population in the Uni— ted States cities. Larger cities have more percentage 52 of negro population than the medium size and small cities and their police protection expenditures are relatively higher than others. Glenn W. Fisher, in his analysis, used the education variable, percentage of population over 25 years of age with less than 5 years schooling, and found significant relations with police and sanitation expenditures of 50 " states with 1960 data? In the present study the variable of percentage of negro head of household indicates the samei relations with the dependent variables. Assuming the fact8 that negro population in large cities are less educated than the majority our results for Michigan cities follow the general trend found in country-wide studies mentioned above. Simple regression coefficients indicating the degree of predictability show the same pattern of relation of these social variables with the dependent variables. Wealth variables Wealth variables consist of (l)household buying in- come,(2)percentage of households with an income less than $5000, (5)median home values, (4)total retail sales. These variables explain the expenditures with higher correlation coefficients than social variables. Buying income variable effects fire protection expen- ditures the most; it has a relation to highway expendi- tures in cities of 10000 or more. Second variable which 53 is a determinant for the poverty level has effects on fire protection eXpenditures again and police protection. In spite of the negative relationships of buying income and fire expenditures this second variable has positive corre— lations. The third variable, median home values, follows the same pattern of variation as the previous variables, it relates to fire and police protection expenditures with a negative coefficient. The fourth variable represents the commercial capacity of the city, its strong relation(in three Observations) is with police expenditures. Total re- tail sales affect also sanitation and parks and recreation expenditures. The capacity of commercial activity is very high in larger cities; we obtained the highest correlation coefficient in the expenditures of large cities. Another consistent result is with sanitation expenditures; the cor- relation is with the expenditures of all cities except su- burban where the level of service is higher and there is less commercial activity than in others. Capacity to finance variables Intergovernmental revenue and property taxes are the finance variables that we adopted. Intergovernmen- tal revenue has frequently been found to be an important determinant of local government expenditures. However, one point should be stressed that in these statewide studies, grants to large cities are the backbone of re- venue; in our study because the small number of large 54 cities we excluded the grants made from federal and speci- al funds e.g., highway, bridge construction, state recre- ational parks. We included only items mentioned in local auditors‘ report under "Revenues from other governmental agencies" heading which are: state income tax, sales tax, liquor licence tax, motor vehicle operator tax, intangible tax, and dog licence tax. These revenues and property taxes represent in average about 60% of the total revenue of cities smaller than 100000. Intergovernmental revenue variable is related posi- tively to police protection, sanitation and parks and recre- ation expenditures. As sales tax is part of this revenue variable, it has the same effects on expenditures as the total retail sales variable. The second variable, proper- ty taxes has a high correlation coefficient again with the three public functions stated above. We can infer from this result that wealthy communities pay more for parks and recreation services and central cities with greater commercial activity spend more on police protection and saw nitation services. 55 Effects of Suburbs on Central City Expenditures IntrametrOpolitan disparities that we have mentio- ned in the first chapter may impose costs to the central ci- ties. Suburbs as part of the Standard metrOpolitan statis- tical areas may use central city facilities and yet may not contribute to the supply of these services? In our analysis which consists of eight different suburban characteristics and corresponding central cities' expenditures we did not find any correlation between ex- penditures and variables but one, percentage of households with income less than $5000. This variable is significant- ly correlated (at the .05 significance level) with police, fire protection and sanitation expenditures. The negative correlation with police and sanitation expenditures may mean that the lowest is the percentage of poor people in the suburbs the highest will be the level of expenditure in central city. As you will see in the next pages the regression coefficients corresponding to these correlations are quite high (-l.l6; -l.l5) thus, the variable is a good determinant for the expenditures studied. Fire expenditures are positively related and the regression coefficients are relatively low for this relation. Comparing the results of this part of analysis to other urban-suburban exploitation studies we could say that ours is limited. This may be due to the limitations of the sample. 56 Scale Economies In spite of the limitations that we have mentioned in the first chapter we can assume that these limitations tend to be minimum when dealing with expenditures of one state. Quality differentials and socio-political varies tions are less meaningful than comparing interstate varies tions. Per capita expenditures in the next page present some economies of scale if not many. Police expenditures seem to be minimum at medium sized cities. Fire expendi- tures variations are not much, however larger cities have least cost per capita. Sanitation expenditures are gene- rally high in large cities but the maximum cost is at the cities 20000-50000 population bracket. Parks and recreation expenditures do not offer any significant economies of scale; per capita expenditures are rather inconsistent. In general, small cities have less cost than mid- size or larger cities. This is reasonable since small ci- ties need to a lesser degree parks and recreation facili- ties. The demand in small places is also less than larger cities. Highway expenditures are diminishing as the size of the city grows. Thus, as you might Observe in the fol- lowing tahua,scale economies appear at larger sizes. 57 Cities Police Fire Sanit. Parks&R Hwy. 10-20000 22.22 16.19 4.74 4.61 15.59 (11 cities) (7.57) (4.56) (4.09) (2.55) (5.25) 20-30000 23.19 16.35 9.09 8.19 12.41 (14 cities) (6.76) (5.83) (4.9) (4.6) (6.9) 30-40000 24.44 15.30 578 12.98 9.49 (7 cities) (5.8) (8.8) (4. 5) (4.1) (5.1) 40-60000 19.37 13.95 4. 24 6.13 9.80 (10 cities) (4.7) (7 1) (3 .5) (2.3) (7.1) 60-100000 21.21 15.00 7. 23 8.42 9.89 (10 cities) (4.6) (4.1) (4 .3) (3.4) (4.4) 100000- 24.34 14.81 7. 25 15.31 11.33 (5 cities) (6.2) (4.2) (5 .4) (8.2) (5.2) (a, 1.7". . i j . \.1 Table 12.§!PER CAPITA EXPENDITURES OF PUBLIC SERVICES IN MICHIGAN MUNICIPALITIES. 58 If we analyze standard deviations we could find the maximum reliability in the police expenditures. We should mention that by reliability term we mean higher coefficient of variation which, itself is equal to: giandard deviation expenditure average Fire protection expenditures is the second to offer low coefficient of variation. Sanitation, parks and recre— ation and highway expenditures have higher coefficients of variation. Min. coef. of variation for expenditures police protection .2215 fire protection .2778 sanitation .6003 parks and recreation .3870 highways .3359 3. 4. 5. 59 FOOTNOTES W. Hirsch, "Cost functions of an Urban Government Service," Review of Economics and Statistics, (February,l965) pp.87-93. Compendium of City Government Finances in 1960, Bureau of the census U.S. Dept. of Commerce, (1962) R.W. Bahl,(1969) p.73; Opus Citation see p.21. Ibid., p.120. For complete list of maintenance items see State Audi- tors' Reports for Michigan municipalities. Dept. of Municipal Finance, Lansing Michigan. H. Brazer, City Expenditures in the U.S., Opus Citation see p.13. Fisher, "Determinants of State and Local Government Expenditures," Opus Citation, see p.22. Theodore H. Sizer,"The schools in the City," The Metropo— litan Enigma, Anchor pub. (1967) pp.360—362. See p.9 of this study for further explanation. CHAPTER V CONCLUSIONS Michigan Cities and Urban Public Services (:In small cities, e.g., within 10-30000 pOpulation bracket, economic variables have greater influences than social variables on expenditures. Total retail sales, me— dian income, home values are significantly affecting the- level of eXpenditures in all functions except parks and recreation service. There is no significant relation bet- ween expenditures and socio-economic variables. This re- sult may be explained that in small cities parks and rec- reation need is less than other cities. Another observap tion is that these cities are generally located in rural areas and not in the proximity of metropolitan settlements. Middlesize cities1 in Michigan offer stable variar tions in expenditures. Within the independent variables economic characteristics are the most significant ones. Population density becomes significant variable yet percen- tage of Negro head of household affects to a lesser degree the expenditures. Pr0perty taxes are related to parks and recreation and highway expenditures; they are more signifi- cant in larger cities since commercial activities become more intense, property taxes augment. 6O 61 A major source of analytical difficulty in the met- ropolitan area arises as a consequence of differences among local communities in the characteristics of their pOpulan tions. As the Advisory Commission on Intergovernmental Relations noted recently: "Population is tending to be in- creasingly distributed within metropolitan areas along eco- nomic and racial lines. Unless present trends are altered, the central city may become increasingly the place of resi- dence of new arrivals in the metrOpolitan areas, of non- whites, lower income workers, younger couples, and the el- derly.fiéjThe justification of this statement for Michigan cities is apparent,in the table on page 35, by the correla~ tion of percentage of negro head of household with police protection, sanitation expenditures of larger cities and ~‘ single cities. / ~~-~ _.._._ ...i (4 The Detroit Area Study's findings on the income expe- rience of whites and nonwhites and residents of the suburbs and the central city, for the period 1951-59 reveal some contrasts. Median family income rose by 9% in the central city but 47% in the suburbs. At the same time, the median income of white families increased by 55% for the area as a whole compared to only 8% for nonwhites. The movement of white, higher income families to Detroit's suburbs, coup- led with their replacement in the central city by lowh income newcomers? Similarly sharp contrasts, emphasizing the diversity among municipalities in structure may be seen in the Detroit area. In 1958, the assessed value of 62 residential prOperty in thirty-four cities, villages, and townships comprised 42% of total assessed valuation in these communities. For the city of Detroit the ratio was 40%, whereas for such industrial enclaves as River Rouge, Trenton, Hamtramck, Highland Park, and Warren it was less than 20 percent; while at the same time, in the Grosse Pointe communities and in Dearborn township, the ratio was 85 percent or higher. Such extreme inequalities as those in the distri- bution within metropolitan areas of socio-economic groups of population and the prOperty tax base give rise to wide differences in expenditures and tax rates. Tax rates and per capita expenditures both tend to be highest in central cities, but ranks with respect to tax rates and expenditures diverge for communities outside the central city:::] Data presented in the Appendix A indicate.that there are substantial differences between the central city and the rest of the metropolitan area in the amounts spend per capita in total and for the separate major services. Part of such differences stems from the fact that the area outside the central city is less fully urbanized, but a large part is undoubtedly attributable to the differences in demographic and other characteristics outlined above. Highway expenditures tend to be inversely associated with population density (see page 49), so that we should we expect them to be higher outside the central city. 63 Implications for Planning Policies One great shortcoming in the problem-solving ef- forts of metropolitan areas is the planning of the allo- cation of physical and financial resources. The urban finance problem, perhaps the most complex of all urban problems, presents a need for the coordination of fiscal and physical planning. For example, an understanding of the implications of a given longrange land-use plan for the spatial distribution of expenditure requirements within the standard metropolitan statistical area or a! mong independent cities is an essential prerequisite for orderly and efficient urban growth. For this latter goal, coordination of the efforts of planner and fiscal econo- mist is crucial. The planner is primarily interested in designing the integrated city. Structural expenditure analysis aids by facilitating the designation of problem areas in relation to certain expenditure levels. Bahl mentions that: "The~ effectiveness of the planner's contributions to a coordi- nated approach to the urban problem may be greatly enhan- ced if he can recognize and anticipate these problem areas and their longrange implications for efficient metropolitan government."4 If a large labor-intensive plant is being considered for a central city location, the planner must anticipate the possible problems created by this particular location. Probably, greater sanitation expenses, the development of 64 high density housing, increased traffic congestion will evolve as the main problem areas. The planner must anticipate also the general ef- fects of the metropolitan land-use plan on the city budget. A plan that does discourage residential migrap tion to suburbs is almost certain to enhance the fiscal resources of the central city, but is not as certain to reduce expenditure requirements. Depending on the extent of anticipated residential dispersion and the degree to which the SMSA is politically fragmented, physical and fiscal planners might look far ahead to potential adequacy of various kinds of non pro- perty taxes. 0n the other hand, many long range land-use plans provide not only the outward movement of central city re- sidents but also, by industrial parks in the urban fringe and an adequate transportation network, the diffusion of commercial and industrial activity within the standard metropolitan statistical area. Of course, suburbanization has a lot more social implications than fiscal, but the cost of suburbanization in terms of expenditure is very important either from social or physical point of view and should be taken into consideration in the land-use planning stage. 65 Summary and Recommendations for Further Study Recognizing that the analysis presented here is confined to data for a limited number of cities, extended generalizations are not justifiable. Certain results and trends have emerged from the study, however. The following comments must be taken as being relevant only for fifty-six cities which were analyzed and only for the time period, e.g. 1970, which was considered. Conclusions that are relevant enough to be mentio- ned here are the following: 1) For police protection and sanitation expenditures the determinants, starting by the most important, are: total retail sales, percentage of negro head of households, pro- perty taxes, intergovernmental revenue, population density. For fire protection expenditures, buying income, median home values and percentage of household with an income less than 33000 are the main determinants. For parks and recrea- tion expenditures, total retail sales, property taxes; and for highway expenditures, population density and median home values are the significant determinants. 2) Expenditures are varying by the type of service or structure of the cities. Suburban cities are differing from single (isolated) cities in their allocation and Spending for urban public services. 3) For better exploration of economies of scale, quality and service level measurements should be developed. 66 4) Determinants of Urban public service expendi- tures help policy makers for a better anticipation of the future cost of urban areas to the municipalities. Limiting the present study to cities within the State of Michigan had both advantages and disadvantages. The governmental and financial structure was similar for all cities included in the study. Disadvantages were noted, however, in the size of the samples and in the distribution of cities within the size categories selected. Another limitation of the study as it stands is that it is based on data provided for only one year. A study would be more representative if data were averaged over a peri- od of many years. After a review of the results of this study, some general recommendations can be made for future research. First, there should be more variables to determine the le- vel of expenditures, more social and political variables. Secondly, and probably most important of all, the ques- tion of what quality and quantity of service can be bought for a given price needs to be answered. Statistical ana- lysis would be more accurate using multiple regression analysis and partial correlation coefficients would eXplore to a greater degree the accuracy of the determinants. 67 FOOTNOTES By middle size cities we mean cities having a pOpulation of 30000-60000. Government Structure and Planning i2 Metropolitan Area , a report by Advisory Commission On Intergovernmental Reéations, Government Printing office, Washington D.C. 19 l. C. Doxiadis, Emer ence and Growth of an Urban Region, Vol.1, Detroit Edison Company,(l966}77 Roy W. Bahl. Metro olitan Cit Expenditures, University of Kentucky Press (I969) p. I51. BIBLIOGRAPHY BIBLIOGRAPHY Adams, R.F. "Determinants of Local Government Expenditures," Review of Economics and Statistics, No.47 (November, 1965) ppT308-313. Advisory Commission on Intergovernmental Relations. Perfor- mance of Urban Functions: Local and areawide, Govern- ment Printing office, Washington D.C., 1963. Bahl, R.W., Saunder, R.J. "Determinants of Changes in State“and Local government Expenditures," National Tax Journal, (March 1965) pp.50-57. Bahl, R.W. "Studies on Determinants of Expenditures," Functional Federalism; edited by S. Mushkin. George Washington University Press, Washington 1968. . Metr0politan City Expenditures, University of Kentucky Press, Lexington 1969. Brazer, Harvey E. "The role of Major Metropolitan Centers in State and Local Finance," American Economic Review (May 1958) pp. 305-316. . "City Expenditures in the U.S.," Qccasiohal Paper No.66, National Bureau of Economic Research, NoYo 9590 _________. "Some fiscal Implications of Metropolitanism," City and Suburb, edited by B. Chinitz, Prentice-Hall, New Jersey I964. . E88 8 in Finance, Institute of Public Adminis- tration, Ann Arbor I967. Breton, Albert."Scale effects in Local and Metropolitan Government Expenditures," Land Economics, (November 1965) PP. 370-372. Crecine, John P. Financin the Metro olis, Sage publications Beverly Hills 1970. 69 Davis, 0.A. and G.H. Harris, Jr. "A political approach to the theory of Public Expenditures: the Case of Munici- palities," National Tax Journal, XIX (September 1966) pp. 259-275. ' .Duncan, 0. Dudley. "The optimum Size of Cities," Reader in Urban Sociolog%,-edited by Hatt and Reiss, Free Press 9 0 pp. - 450 Fabricant, Solomon. The Trend of Government activity_ingthe U.S. since‘1900, National—Bureau of Economic Research Inc., New York 1952. pp.112-l39. Fisher, G. W. "Determinants of State and Local Government Expenditures: a Prealimina analysis,” National Tax Journal, XIV (December 1961 pp. 57-74. . "Interstate Variation in State and Local Govern- ment Expenditures," National Tax Journal, {March 1964) pp,.10;-105. Gabler, L.R. "Economies and Diseconomies of Scale in Urban Public Sectors," Landgficonomics, (November 1969) pp.425-431. Hansen, Alvin H. and H.S. Perloff. State and Local Finance, in the National Economy, W.W. Norton & Co. Inc., New York i941. Hansen, Niles M. "The Structure and Determinants of Local Public Investment Expenditures " Review of Economics and Statistics, No.2 (May 1965) pp. 150-162. Hawley, Amos H. "Metropolitan P0pulation and Municipal Government Expenditures in Central Cities," Journal of Social Issues, (June 1951). Hirsch, Werner Z. "Expenditure Implications of Metropoli- tan Growth and Consolidation " Review of Economics and Statistics, (August 1956) pp.232-241. . "Determinants of Public Education Expenditures," National Tax Journal, (March 1960) pp.29-40. . "Factors affecting Local Government Expenditures," Exploring the Metropolitan Community, edited by J.C. Bollens, University of California Press, Berkeley 1961. . "Cost function of an Urban Government Service: Refuse collection," Review of Economics and Statistics, XLVII (February 1965) PP. 87-95. 7O Hirsch, Werner Z. "The Supply of Urban Public Services," Issues inggrban Economics, edited by Perloff & Wingo, J.H0pkins Press, Baltimore 1968. Howard, William A. "City Size and its Relationship to Muni- cipal Efficiency: Some observations and Questions," Ekistics, V01.28 (November 1969) pp.312-316. Isard, Walter and Robert Coughlin. Municipal Costs and Re- ggnues resulting from Community Growth, Chandler-Davis Publications 1957. Kee, Woo Sik. "Central City Expenditures and Metropolitan Areas," National Tax Journgl, XVIII (December 1965) pp 0 337-354 0 . "City Suburban Differentials in Local Govern- ment Fiscal Effort," National Tax Journal,(June 1968) pp ' 183-189 0 Morss, Elliott R. "Some Thoughts on the Determinants of State and Local Expenditures," Nationaligax Journal, XIX (March 1966) pp. 95-103. Musgrave, Richard A. Theoryof Public Finance, Mc-Graijill, New York 1959. Phillips, Hugh. "Municipal Efficiency and Town Size," Journal of Town Planningjlnstitute, (May-June 1942) PD 0 1:3‘9-148 o Pidot, G. Jr,. "A Principal component Analysis of the Deterb minants of Local Government Fiscal Patterns," Review ofBEconomics and Statistics, Vol. 51 (May 1969) pp. 176- 18 . Ridley, Clarence E. and H.A.Simon. Measuring Munici a1 Activities, International City Managers Assn. C cago, I958. Sacks, Seymour and Robert Harris. "The Determinants of State . and Local Government Expenditures and Intergovernmental Flows and Funds," National Tax Journal, XVII (March 1964) PP. 75-85. Samuelson, P.A. "A Pure Theory of Public Expenditures," Review of Economics and Statistics, (November 1954). . "Diagrammatic Exposition of a Theory of Public Expenditures," Review of Economics and Statistics, (November 1955). "ASpects of Public Expenditure Theories,” Review of Economics and Statistics, Vol.24 (November 1958). 71 Schaller, Howard G. ed. Public Expenditure Decisions in the Urban Community, John Hopkins Press, Baltimore 1963. Schmandt, H.J. and G. Ross Stephens. "Measuring Municipal Output," National Tax Journal, (December 1960) pp.369— 375. . "Local Government Expenditures in the U.S.," Land Economics, (November 1963) pp. 397-406. Shapiro, Harvey. "Measuring local Government Output: a Com- ment," National Tax Journal, No.9 (December 1961) PP. 394-397. . "Economies of Scale and Local Government Finance," Land Economigp, (May 1963) pp. 175-186. Siegel, Barry N. "0n the Positiveiheony of State and Local Expenditures," Public Finance and Welfare, P.L. Kleinsorge, editor, New York 1965. Scott, Stanley and E.L. Feder. Factors Associated with Variations in Municipal Expenditure Levels, Berkeley I957. Tiebout, Charles M. "A Pure theory of Local Expenditures," Journal of Political Econom , (October 1956) pp.416-424. Walker, Mabel L. Municipal Expenditures, John Hopkins Press, Baltimore 1930. Wilensky, Gail. "Determinants of Local Government Expendi- tures," Financin the Metr olis, edited by J.P.Crecine, Sage PubIIcations, CEIifornia I970. pp.197-219. Will, Robert E. "Scalar Economies and Urban Service Require- ments," Yale Economic Essayg, Vol.5 (Spring 1965). APPENDICES, 72 APPENDIX A Table 13. 1970 EXPENDITURES IN MICHIGAN CITIES OF 10000 OR MORE.(D011ar amounts in thousands) Source: State Auditors' Reports, Bureau of Municipal Finance State of Michigan, Lansing. Michigan (1970). Qipies Police Eype Sanit. Parks&R. Hwy. ADRIAN 373 316 300 118 316 ALBION 281 292 146 129 229 ALLEN PARK 838 413 369 234 135 ALPENA 267 286 - 96 367 ANN ARBOR 2273 1289 917 1698 477 BATTLE CREEK 1344 1155 361 616 356 BAY CITY 1022' 1272 - 182 902 BENTON HARBOR 619 357 152 82 229 BERKLEY 384 245 285 64 217 BEVERLY HILLS 400 - y 202 ll , 333 BIG RAPIDS 153 143 84 23 190 BIRMINGHAM 787 660 74 248 439 CENTER LINE 375 191 128 107 78 CLAWSON 322 42 166 86 254 DEARBORN 2362 1263 1171 2840 2411 DEARBORN HEIG. 1057 633 584 202 750 DETROIT 55877 23410 28854 25766 13170 E.DETROIT 1105 413 - 288 503 E.GRAND RAPIDS 213' 167 79 92 185 E. LANSING 633 523 240 114 304 ECORCE 800 444 537 352 689 ESCANABA 270 301 94 101 200 FARMINGTON 174 100 79 35 143 FERNDALE 893 742 399 142 419 FLINT 5665 3804 1401 1827 2764 FRASER 384 92 130 26 32 GARDEN CITY 596 312 330 449 127 GRAND HAVEN 259 198 - 107 297 GRAND RAPIDS 4639 3565 717 2021 2534 GRANDVILLE 146 43 l2 19 38 GROSSE PT.FARMS 485 316 - 118 - GROSSE PT.FARK 445 292 146 129 229 GROSSE PT.WOODS 412 178 213 140 172 HAMTRAMCK 810 694 327 - 130 HARPER WOODS 552 347 264 133 272 HAZEL PARK 524 314 93 124 196 HIGHLAND PARK not available HOLLAND 601 356 - 259 516 INKSTER 912 320 - 234 399 JACKSON 1303 1121 235 216 1082 KALAMAZOO KENTWOOD LANSING LINCOLN PARK LIVONIA MADISON HE. MARQUETT MELVINDALE MENOMINEE MIDLAND MONROE MT. CLEMENS MT. PLEASANT MUSKEGON MUSKEGON HE. NILES NORTON SHORES OAK PARK OWOSSO PLYMOUTH PONTIAC PORTAGE PORT HURON RIVER ROUGE RIVERWIEW ROSEVILLE ROYAL OAK SAGINAW ST. CLAIR SHORES ST. JOSEPH SAULT ST. MARIE SOUTHFIELD SOUTHGATE STERLING HE. TAYLOR TRAVERSE CITY TRENTON TROY WALKER WARREN WAYNE WESTLAND WYANDOTTE WYOMING YPSILANTI Appendix A (continued) 1840 183 3262 927 1600 not 283 372 117 549 571 605 206 968 407 316 173~ not 334 228 2448 467 753 881 350 991 1453 2017 1314 331 302 1514 688 1167 977 256 777 1067 148 3402 634 1131 830 725 810 73 1657 106 2782 444 1062 available' 307 201 146 499 497 360 126 961 210 207 113 available 391 125 1738 225 686 567 180 530 1089 1539 775 225 228 1146 314 573 503 258 592 100 50 2403 347 713 658 310 494 60 13 470 373 610 75 154 50 247 308 325 96 357 36 3 7 76 1393 95 9O 74 306 953 593 469 158 6 460 41 154 285 327 l 1386 207 546 17 185 850 23 3265 2506 340 13 27 465 315 119 77 330 23 34 22 85 30 921 250 178 211 424 1277 399 142 42 641 165 97 276 55 416 211 841 216 262 386 313’ 311 1151 268 .1013 325 1722 631 151 94 631 423 120 277 683’ 343‘ 218 85 268 160 1709 444 320 186 65 453 532 645 799 240 338 708 511 319 210 296 257 57 1489 727 616 167 74 APPENDIX B Table 14. SOCIO—ECONOMIC CHARACTERISTICS OF CITIES CH.1:§qu1ation Densiiy per sq. mile; source:1970 U.S.Census Ch.2:Po ulation Number(by thousands); source:1970 U.S.Census Ch.3: ercentage of Households with an Income of less than 3;,000; source:Sales Management Magazine,‘19705Survey 0 Buying Power. Ch.4:2pta1 Retail Sales(by millions); source:Sales Manage» ment Magazine, 970 estimates. Ch.5:Effective Buying Income per Househ01d(Dollars by Thousands); source:Sales Management Magazine, 1970 Survey of Buying Power. Ch.6:Median Value of Houses(Dollar value by thousands); source:I97O U.S.Census Reports General Housing Charac- teristics. Ch.7:Percentage of Negro Head of Household; source:1970 U.S. Census Reports General Housing Characteristics. Ch.8:;ptergovernmenta1 Revenue(Dollar value by thousands); sources:1970 U.S. Census Reports of City Finances, and Official Reports of Municipal Finance Bureau,State of Michigan, Lansing 1970. Ch.9:§r0perty Taxes; source31970 U.S.Census Reports of City Finances, and Official Reports of Bureau of Municipal Finance, State of Michigan Lansing. Cities Ch.1 Ch.2 Ch.3 Ch.5 Ch.6 .Ch34 ADRIAN , 3514 20.3 21.3 9.9 14.9 69.6 ALLEN PARK 5506 40.7 18.1 10.1 22.1 93.9 ALPENA 1866 13.8 17.5 9.8 11.9 53.2 ANN ARBOR 4578 99.7 15.9 14.0 27.7 244.3 BATTLE CREEK 3299 38.9 22.5 9.2 11.4 144.4 BAY CITY 4945 49.4 17.3 9.5 12.7 137.7 BENTON HARBOR 4578 16.4 2313‘ 8.6 11.3 98.6 BIRMINGHAM 5816 26.1 5.7 21.0 31.9 137.6 CENTER LINE 6105 10.4 6.3 12.6 19.8 35.3 DEARBORN 4253‘ 104.2 6.4 14.6 22.8 383.1 DETROIT 10953 1511.5 13.5 11.0 15.6 2400.1 E.DETROIT 9004 45.9 3.9 13.1 20.4 97.3 E.LANSING 5282 47.5 1338 17.9 29.3 47.8 ESCANABA 1220 15.4 19.6 -8.6 12.7 49.9 FARMINGTON 5130 13.3 3.7 16.5 31.8 89.9 FERNDALE 7910 30.9 6.6 12.1 16.6 91.7 75 APPENDIX B (continued) Cities Ch.l Ch.2 Ch.3 Ch.4 Ch.5 Ch.6 FLINT 5894 193.3 13.6 489.9 10.5 14.6 GARDEN CITY 6541 41.8 2.4 62.9 11.9 19.6 GRAND RAPIDS 4402 197.6 17.2 458.4 13.9 14.8 GROSSE PT.WOODS 6630 21.9 3.8 23.3 22.2 34.5 HAMTRAMCK 12974 27.2 15.0 53.6 10.6 10.1 HARPER WOODS 7764 20.2 4.7 121.9 13.7 22.3 HAZEL PARK 8494 23.8, 7.5 38.1 11.6 15.5 HOLLAND 1908 26.3 13.3 76.2 11.2 14.9 INKSTER 6126 38.6 8.1 46.0 9.8 17. 6 JACKSON 4251 45.5 17.5 160.8 13. 3 11.8 KALAMAZOO 3492 85.5 16.7 259.5 11.6 14 3 LANSING 3939 131.5 14.3 383.7 10.4 16. 3 LINCOLN PARK 8831 53.0 4.8 170.7 12.1 17.4 LIVONIA 3050 110.1 3.4 302.1 14.7 27.1 MARQUETT 1997 22.0 17.7 49.2 11.4 16.1 MIDLAND 1413' 35.2 12.8 100.3 14.2 22.1 MONROE 2507 23.9 14.2 77.7 11.8 17.8 MT.CLEMENS 5250 20.5 13.9 94.3 12.2 18.3 MT.PLEASANT 4020 20.5 18.9 48.6 14.5 17.3 MUSKEGON 3433 44.6 18.0 122.2 9.1 11.3 MUSKEGON HEIGHTS 5244 17.3 16.4 28.4 8.5 9. 8 NILES ' 2498 13.0 16.3 56.0 10.9 13. 1 OWOSSO 3655 17.1 18.3 60.8 ' 9.9 13. 9 PLYMOUTH 5112 11.7 7.4 57.4 13.2 23.8 PONTIAC 4329 85.2 11.5 217.6 10.6 15.4 PORT HURON 4773 35.8 19.6 95.8 9.7 12.3 ROSEVILLE 6176 60.5 5.7 153.6 11.6 18.9 ROYAL OAK 7308 85.5 5.4 192.8 14.2 21.3 SAGINAW 5309 91.8 16.4 226.0 9.8 13.7 ST.CLAIR SHORES 7403' 88.1 4.2 110.1 13.3 22. 5 SAULT ST.MARIE 964 15.1 19.8 37.9 8.6 8. 7 SOUTHFIELD 2501 69.3 4.5 343.7 16.4 36.0 SOUTHGATE 4710 33.9 3.9 128.1 12.1 19.4 TRAVERSE CITY 2314 18.0 18.8 103.9 11.7 15. 4 TRENTON 3260 24.1 73.8 35.3 14.4 26.0 TROY 1177 34.9 5.0 88.4 14.2 29. 8 WARREN 5242 179.2 5.0 321.9 11.8 23.4 WAYNE 3509 21.0 6.6 72.5 12.2 18.2 WYANDOTTE 7896 41.6 7.6 78.4 11.5 16.9 WYOMING 2318 56.5 7.7 160.4 10.7 14.4 YPSILANTI 7204 29.5 15.9 90.9 13.3 20.7 76 APPENDIX 0 SIMPLE CORRELATION AND REGRESSION ANALYSIS Source: N.M. Downie and R.W. Heath. Basic Statistical Methods, Harper & Row, New York (1970): Chapter 7‘and 9. The size of the Pearson product-moment correlation coefficient (r) used in the study varies from +1 to -1. Most correlation coefficients tell us two things. First we have an indication of the magnitude of the relationship. A correlation of -.88 is the same as one of +.88. The sign does give only information about the direction of the rela- tionship. When two variables are positively related, as one increases, the other increases, too. In everyday usage an r of .80 and above is considered a high coefficient, an r around .50 is considered moderate; and an r of .30 and below is considered a low coefficient. It should be stated that a Pearson r is not a measure of causality, although in some cases causal relationships may exist between the two variables. The formula used in the study is: r 3 :XY -|(zx)(;Y)/N1 \szi in )‘ /N]H 2Y‘- (me/NIT Y: 1970 expenditures for public services I: Socio-economic characteristics of cities where: The term"regression analysis" refers to the methods by which estimates are made of the values of a variable from a knowledge of the values of one or more other variables, and to the measurement of the errors involved in this estimation process; although correlation analysis refers to methods for measuring the strength of the association among variables. Linear regression means that an equation of a straight line of the form Y:=a* bx , where a and b are numbers, is 77 used to describe the average relationship that exists bet- ween the two variable and to carry out the estimation pro— cess. The factor whose values we wish to estimate is re- ferred to as the dependent variable and is denoted by the symbol Y,representing urban public service expenditures in our study. In other terms, values of expenditures are de- pendent upon the values of X, socio-economic data of Michi- gan cities. In the analysis chapter of the thesis, coeffi- cients b for each relation are given in tables followed by their standard error of estimation. Both analyses are programmed for computer, CDS 6500, by: Eva Clark . Correlation and Regression analysis, Michi- gan State University, Computer laboratory. Sept. 25, 1961. CO-OP ID: 02 UCSD BIMD in FORTRAN. HICHIGRN STATE UNIV. llll 8 LIBRRRIES 312931007 9761