University of Southern California Dissertations and Theses

Contributions To The Theory Of Industrial Location: Urban Decentralization

(USC Thesis Other)

Contributions To The Theory Of Industrial Location: Urban Decentralization

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted. The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction. 1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity. 2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame. 3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete. 4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced. 5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received. Xerox University Microfilms 300 North Zeeb Road Ann Arbor, Michigan 48106 s MacREYNOLDS, M1111am Kenneth, 1944- I CONTRIBUTIONS TO THE THEORY OF INDUSTRIAL f LOCATION: URBAN DECENTRALIZATION. University of Southern California, Ph.D., 1974 | Economics, theory ! Xerox University Microfilms, A n n A r b o r , M i c h i g a n 48io e THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED. CONTRIBUTIONS TO THE THEORY OF INDUSTRIAL LOCATION URBAN DECENTRALIZATION by William Kenneth MacReynolds A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Economics) June 1974 U N IV E R S IT Y O F S O U T H E R N C A LIF O R N IA THE GRADUATE SCHOOL U NIVER SITY PARK LOS ANGELES. C A LIFO R N IA 9 0 0 0 7 This dissertation, written by William Kenneth MacReynolds under the direction of h.ik?... Dissertation Com mittee, and approved by a ll its members, has been presented to and accepted by The Graduate School, in partial fulfillm ent of requirements of the degree of D O C T O R D O C T O R O F P H IL O S O P H Y <L Dean D a t e . . . . A ^ls L}hJ.9 Jl TABLE OF CONTENTS LIST OF TABLES ............................... LIST OF ILLUSTRATIONS ......................... CHAPTER I. INTRODUCTION ........................... The Scope and Purpose The Spatial Analytical Framework Transportation Epochs Urban Decentralization and Growth jhe Flow of Analysis II. THE BACKGROUND OF INDUSTRIAL LOCATION THEORY ................................. Theories of Industrial Location The Spatial Structure of the Nineteenth Century City A Concept of Urban Decentralization III. A MODEL OF SPATIAL EQUILIBRIUM ........ Assumptions of the Analysis Equilibrium in the Land Market Equilibrium in the Labor Market Spatial Equilibrium of the Firm Adjustments to Complete Equilibrium The Impact of Distance on Production Differential Labor Skill Gradients IV. EVIDENCE OF DECENTRALIZATION IN THE TECHNOLOGICAL EPOCHS ................... Changes in the Mode of Transportation Evidence of Decentralization Comparison of Results in the Two Periods V. AN EMPIRICAL TEST OP DECENTRALIZATION..........66 The Characteristics of Decentralization in Los Angeles, 1949-1964 The Test The Empirical Results VI. SUMMARY AND CONCLUSIONS........................ 84 Summary of Analysis Summary of Empirical Results Concluding Remarks BIBLIOGRAPHY ........................................ 97 LIST OF TABLES 1. Sample Breakdown for Both Periods..................55 2. Distribution of Origin and Destinations, 1908-1920 ........................................ 56 3. Distribution of Origin and Destinations, 1949-1964 ........................................ 59 4. Regressions on Destination Distance of Manufacturing Firms in Los Angeles, 1949-1964 . . 78 iv LIST OF ILLUSTRATIONS 1. Land Rents at Equilibrium................... 28 2. Wage Rates with Constant Commuting Costs .... 30 3. Money and Time Commuting C o s t s .............32 4. Wage Rates with Increasing Commuter Costs .... 33 5. Shift in Land Rent Gradient.................38 6. Shift in Wage Rate Gradient.................39 7. Wage Rate Gradients for High and Low Skill Labor........................................ 46 8. Shift in Land Rent Gradient with Major Growth at Core............................................ 91 9. Shift in Land Rent Gradient with Major Growth Away from C o r e ....................................92 v CHAPTER I INTRODUCTION The Scope and Purpose This study focuses on the peculiar pattern of indus trial relocation that occurs as cities experience growth. Urban growth can be characterized by increases in the urban population, the extension of geographical boundaries, increases in the urban per capita income, and increases in employment and industrial activity. Whatever measure is used and whatever source is consulted, the data clearly show that the city of Los Angeles, California has experi enced continuous urban growth since the latter part of the nineteenth century. Urban growth and the causes of this growth are out side the purview of this study. Yet it is important to recognize that growth, particularly in population and in come, creates increased demand for goods and services. When urban areas are also growing in geographical size the friction of space becomes an important parameter. Changes in transportation technology, including the building of new roads and highways, and the addition of automobile travel to railroad travel tend to reduce the friction of space by 1 reducing the unit cost of travel. The improvements in transportation technology improve the movement of goods and people and facilitates growth. Changes in transportation technology tend to occur at distinct moments in time. Thereafter gradual adjustments to the altered situation ensue. When transportation costs are reduced, areas not central to business activity become more accessible. Greater accessibility of outlying areas implies an increasing demand for these areas as potential sites for residential and business activity. The movement of manufacturing plants away from the central business area of a city— hereafter called the core— represents an impor tant phenomenon of urban growth. This study examines the phenomenon of urban decentralization in the light of epochal changes in transportation technology. The Spatial Analytical Framework Industrial activity in large United States cities during the eighteenth, nineteenth, and early twentieth centuries tended to concentrate in a central area. Trans portation links to other cities and to the surrounding residential areas emanated from the core area. Shipments from other cities were delivered to the core area at a central goods handling location. This was necessitated by the large scale economies of loading and unloading proc esses, particularly for rail and water shipments. Manufacturing industries, receiving raw materials and intermediate goods for use in production, congregated around these depots to save on transport costs. Other manufacturing industries which shipped their product local ly to these manufacturers congregated near them to also save on transportation costs. Service industries, such as banking, insurance, retail, and other professionals, etc., also tended to locate centrally to serve a large market and to economize on communication and transportation costs with their clients. These interindustry linkages accompanied by the desire to economize on transportation and communication costs caused the great degree of spatial agglomeration ob served in U.S. cities. Further, the agglomeration of in dustrial activity creates signals to potential producers to set up business in or near the core area in order to ex ploit existing business and spatial arrangements. That is, agglomeration is additionally encouraged by external economies emanating from existing spatial arrangements. Households tended to congregate around the core area near the major transportation links with the core. Since households did not carry on business directly with other cities on a regular basis, the central goods handling facility had less attraction. Instead, local work trips and shopping trips were the important activities of the city's residential population. These activities also led to a desire to economize on transportation costs incurred during trips to and from the core area. The pattern that emerged was a city dominated by a central core area containing most of the industrial activ ity of the city with the supporting services surrounded by households. As distance from the core increased, land use became less intensive. Residences became more diffused as accessibility to the core was reduced with distance. The analysis of a core-dominated city dates back to the work of von Thunen.l Recent adaptations of this work to urban land markets have been made by numerous scholars. The most prominent studies include those by Alonso,2 Muth,3 and Wingo.^ The principal focus of these models has been to derive the equilibrium distribution of house holds and industrial firms from which emanates the equilib rium price of land. The price of land, land rent, is shown to fall with distance from the core area at equilib rium. These scholars also argue that the intensity of land use for housing also falls with distance from the core at equilibrium. These static land use models, while fairly elaborate, do not allow for variations in the price of labor in an urban area. M o s e s ^ extends the traditional core-dominated city analysis to include variations in labor wage rates. He argues that workers would be willing to accept lower wages for home employment due to the reduction in commuting costs. The location of jobs and industries need not always be at the core. Thus, the equilibrium distribution of firms becomes an important issue in urban spatial arrange ment. The spatial analytical framework of this study begins with the notion of a core-dominated city and the associated equilibrium conditions. A change in the transportation technology represents a disturbance to this equilibrium. The conditions of the initial equilibrium suggest movement away from the core area when transportation costs are re duced. The process of urban decentralization is character ized by plant relocation at greater distances from the core area. Transportation Epochs This study parallels the recent work by Moses and Williamson^ although there are many significant differ ences in the empirical results. There are some differ ences in the analytical framework which represent an im provement in exposition over their work. A strong similar ity is imposed by the choice of test periods. In both studies data are chosen from two periods, 1908 to 1920 and 1949 to 1964. While their data were gathered for the city of Chicago, these periods represent equivalent transporta tion technological epochs for Los Angeles. In the early period, the introduction of the truck into intracity travel represents the primary technical change just prior to 1908. The duration of the period allows enough firm relocation to establish a pattern of movement. In the later period, the building of intercity highways and intracity freeways represents the important technical change. Again, a pattern of firm movement emerges for the duration of this period. Both this study and the study of Chicago find a pattern of urban decentralization occuring in both periods. Both studies look at the migration of manufacturing firms. According to Vernon^ manufacturing firms have experienced their greatest growth at noncore areas of a city even though the transport bill represents a large portion of a manufacturing firm's costs. It is the break in the trans portation tie to the core area that, in part, facilitates the relocation of manufacturing plants at greater distances from the core. The major contributions of this study are an opera tional definition of urban decentralization, the empirical isolation of the impact of wage rates on decentralization, the empirical isolation of wage rate gradients for two distinct labor skills, and the theoretical and empirical significance of firm size and factor intensity for firm relocation. In the process of presenting these findings the analysis of a core-dominated city becomes more general ized to include spatial arrangements that exhibit more uniform land use intensity as well as multiple uses at various locations within the city. Urban Decentralization and Growth The decentralization of economic activity in major metropolitan areas is a familiar phenomenon. In the typical process of development a city emerges as a small, core-dominated area with a large portion of the goods con sumed being produced in other areas. As these cities attract more people and newer business the physical size of the core area enlarges and the supporting population spreads out around the core. The city becomes many times the size of the original city and will likely become a net exporter of goods rather than an importer. It has also been observed that cities develop multiple core areas as growth occurs. It does so through the portioning of econom ic activities into noncore areas which have the potential to rival the original core in activity. Of course, a certain amount of this development is planned and antici pated by local governmental bodies. Yet a great deal of this redistribution is spontaneous and unplanned. In Los Angeles, the dominance of the Downtown area has receded in recent years. This is due, in part, to the limits imposed by past capital accumulations of buildings and surface travel which limit the growth of this area. The valleys nearby and the flatland to the west are, how ever, tied by transportation routes to the Downtown area. The pattern of traffic flows during rush hours gives evidence that this area is still dominant to a degree. The Wilshire area to the west, the San Fernando Valley to the northwest, the many cities to the east, and the areas to the south as far as Long Beach rival the Downtown area for economic activity. The evidence presented here for the period 1949 to 1964 indicates that the Downtown area still retains dominance and that the net flow of economic activ ity is away from this core. It is not clear from this pattern of movement that any other area is as dominant. Vernon® maintains that the dominance of a core per sists over time yet changes in character. Service indus tries replace manufacturing industries in the core area and these displaced manufacturing industries enlarge scale of plant at noncore locations. Manufacturing firms switch from highly interdependent production to self-contained processes by routinizing their process and standardizing their product. This results in a shift in the demand for labor at core and noncore locations. The core continues to hire and attract specialized resources while the other areas shift to more standardized production processes with less demand for diverse labor skills. That decentralization occurs in manufacturing is not meant to indicate that decentralization occurs in all in dustries. For those service industries where close and cheap communication is important, the desire to agglomerate will tend to continue. Yet these industries are somewhat uninteresting from the standpoint of transportation. The product they "produce" is disseminated at very low uniform cost throughout the urban area. Their major influence on transportation is through commuting. Thus the location of services and all other nonmanufacturing industries will be considered given and remain outside the influence of the analysis. The Flow of Analysis Chapter II presents the important variables that impact the process of urban decentralization and the impor tant contributions of other authors that bear heavily on this study. Chapter III presents the formal model of urban spatial equilibrium used in this study. The model is, at first glance, a von Thunen type model of a core-dominated city. Yet it dissolves into considerations of comparative cost a la Weber. Externalities of proximity, or agglomera tion effects, are assumed to exist only at the core. Their influence does not spread to any other area and, thus, does not influence noncore location decisions. Three types of transport costs are considered to have significant impacts at equilibrium. They are the costs of transporting final product from the core, the costs of transporting inter mediate goods from the core, and the costs of transporting people on a round-trip work basis. These costs completely determine the location of households, the location of firms, and the choice of job location by workers. What emerges are two price-distance relationships for land and labor, the land rent gradient and the wage rate gradient, respectively. The analysis shows that under the given assumptions these factor gradients slope downward in all directions away from the core. That is, land and labor are cheaper at greater distances from the core. Chapter IV presents data on the movement of manufac turing firms in Los Angeles for the period 1908 to 1920. This corresponds to the time period chosen by Moses and Williamson^ for Chicago for which directly comparable results are presented. It is an important time since the introduction of the truck had major implications for the cost of moving goods within the urban area. The length of the period is assumed to be long enough for the majority of reactions to this technological change to take place. It is not long enough, however, to allow complete struc tural changes in the urban area. The period of this test represents an attempt to isolate the relocation and ex pansion of firms assuming that other spatial arrangements remain fairly constant. The data shows that a net de centralization of manufacturing activity took place over this period. That is, manufacturing firms moved away from the core area with those firms closer in being set loose by the technological change. 11 In Chapter V more extensive sampling of manufacturing firms' movement is taken for the period 1949 to 1964. A similar pattern of decentralization is observed. In an attempt to isolate the savings on land and labor costs a regression test is performed to explain the distance desti nation of these movements. Two directions are partitioned as possessing qualitative differences in movement. Also, small firms are separated from large firms in another partition. The results show the significant impact of the factor savings and the effect of economies of scale on firm movement. Chapter VI conveys some of the conclusions readily drawn from the regression analysis. Of particular interest is the behavior of firms that moved toward the core rather than away. The regression tests indicate that, in part, these firms moved closer to the core for some of the same reasons that other firms moved away, i.e., savings in labor and land costs. The central hypothesis of this study is that firm movement occurs in response to economic variables of profit maximization and utility maximization. If this is the case, firms relocate and expand in response to savings in costs and increases in profits. This is hardly a unique notion yet is still difficult to observe. The movement of manufacturing firms in the periods studied is in response to factor savings and the possibilities to exploit new 12 processes of production. The hypothesis that movement is spurious and unresponsive to economic variables is thereby rejected as the tests will indicate. 13 FOOTNOTES 1. Johann Heinrich von Thunen, Von Thunen's Isolated State, trans. Carla N. Wartenberg, ed. Peter Hall (Edinburgh: Pergamon Press, 1966). 2. William Alonso, Location and Land Use: Toward a General Theory of Land Rent (Cambridge: Harvard University Press, 1964) . 3. Richard F. Muth, Cities and Housing (Chicago: Univer sity of Chicago Press'^ 1969) . 4. Lowdon Wingo, Transportation and Urban Land (Balti more: Johns Hopkins Press, 1961). 5. Leon N. Moses, "Towards a Theory of Intra-Urban Wage Differentials and Their Influence on Travel Patterns," Regional Science Association Papers,9(1962):53-63. 6. Leon N. Moses and Harold F. Williamson, "The Location of Economic Activity in Cities," American Economic Review# 18 (May 1968): 211-222. 7. Raymond Vernon, "Production and Distribution in the Large Metropolis," The Annals of the American Association of Political and Social Science, 314(November 1$57):15-29. 8. Ibid. 9. Moses and Williamson, "The Location of Economic Activ ity in Cities," 211-222. CHAPTER II THE BACKGROUND OF INDUSTRIAL LOCATION THEORY Theories of Industrial Location Two prominent approaches to industrial location theory stem from the work of von Thunen^ and Weber.2 A comparison of the implications of the two approaches was studied by Fales and Moses.2 The von Thunen approach em phasizes the impact of a central production zone on settle ment surrounding this area. Most cities are at one stage or another characterized by a dominant production and dis tribution zone. The surrounding areas are assumed to be exclusively housing. Thus the study of the location of housing in an urban area has important spatial meaning. Muth^ concluded that the rate of decline in land rentals with distance from the core varies directly with the rate of decline in housing prices and inversely with changes in land's share in the value of housing output. Similar con clusions emanate from the studies by Alonso,2 Muth,6 and Wingo.^ The implications for industrial location are that industries locating away from the core face the same type of declining rent with distance. 14 Models of industrial location in the Weberian style are profit maximization or cost minimization models. These models generally ignore the housing location problem and assume their spatial distribution to be given. The im portant parameters are the location of raw materials, the use of capital, labor, and land in production, the location of the market, the costs of transporting goods, and the prices of goods and factors of production. Goldberg0 ex tends this analysis to examine the role of the size of the firm in intraurban location decisions. The confluence of intraurban industrial location theory with the choice of household location comes when one considers the problem of workers in households choosing job locations in response to commuting costs. Moses^ presents this analysis emphasizing the differential wage rates that emerge in the urban area. This study attempts to integrate the basic theories of industrial and household location with that of job location. The implications of the analysis are very strong when considering what happens in the process of urban de centralization. Before turning to the analytical model to be developed in this study it might be useful to consider some results comparing the usefulness of the von Thunen and Weberian approaches. The model to be presented in Chapter III combines the two approaches in an analytical model. 16 The Spatial Structure of the Nineteenth Century City Fales and Moses-1 -® analyzed the location activity in Chicago just after the great fire in the early 1870's. Re building occurred in a short period of time so that changes in construction technology and transportation technology did not play crucial roles. The fire destroyed most of the fixed capital and eliminated much of the capital com mittment to the past. Their study highlights some of the important variables to consider in intraurban location. The von Thunen type model explains land use by the competitive process of bid rent curves. Households and industries bid for various locations with the realization that increased distances from the central core area incur greater transport costs. All bid rent curves peak at the center. The intersections of various bid rent curves establish boundaries between uses and represent the optimum allocation of space. Each segment is occupied by the use with the highest bid rent curve there. The overall gradient is the envelope of the individual bid curves. The derived land pattern is a series of concentric zones, one ofr each type of use. There is no intermixing of uses within any given concentric zone.H Implicit in this analysis are the assumptions that firms within an industry exhibit constant returns to scale in production and that firms are not linked in buying and selling arrangements. Thus, firms of different size are not found in the same zone and locational decisions are 17 based on the notion that intermediate goods move at zero cost. In testing the von Thunen-type model for Chicago Fales and Moses looked at the broad implications of the von Thunen model. One of the implications of Thunen-type models is that the intensity of land use declines with distance from the center of a city. This is im plied by the negatively sloped rent gradient. A second implication is that this decline in inten sity is uniform in all directions from the center. This follows from the assumption that transport costs are radially identical. The final implica tion is the exclusive zone concept noted a b o v e .12 TJsing data on employment densities and population densities in separate tests they were found to fall with distance. They did find that the effect of direction on distance from the core was significant. Size of firm was found to be an important explanatory variable when the data on employment densities was partitioned by direction. The final test on exclusive zones showed that population was more dispersed than employment. Yet there was a signi ficant amount of employment in each area of the city. The latter result implies the importance of linked firms in location decisions. While they accepted the declining land use intensity the Fales and Moses tests pointed to the importance of transport costs on intermediate goods and linked firms, channelization of transport routes, and scale economies within individual industries. The model developed in 18 Chapter III and the tests that follow include recognition of these qualifications to the core-dominated theory of von Thunen. Weberian location theory considers such things as the cost of transmitting information, the material intensity of production, and the costs of transporting people and goods. Optimal location of a firm could be determined by finding the place where the total cost of bringing together raw materials and delivering products to the market was at a minimum. The structure of transport costs are thus impor tant in determining the broad location of firm production. The weight of localized materials to final product weight was thought to answer whether a product would tend to be market or materials oriented. It was also possible for labor to vary in cost between locations. Firms could move from cheap transport areas to cheap labor areas if the savings in labor exceeded the extra transport costs. The second part of Weberian theory was concerned with the agglomeration or deglomeration effects of industry within a given region. Agglomeration forces included econ omies of scale, linkages between firms, more effective marketing situations where many firms are located, and re ductions in the costs of overhead items such as streets with large industry concentration. Weber believed that all deglomerative factors followed from the rise in land values which was caused by concentrations of industry. 19 Fales and Moses found that the Weberian approach better explained concentrations of industry, particularly around central goods handling facilities. Economies of scale in this operation played a prominent agglomeration role. The concentration of traffic along relatively few routes and at relatively few terminals permitted these economies of scale and the resulting low shipping rates. Scale of goods movement within cities was small in compar ison to intercity shipments. The results of the Fales and Moses tests show that the extended Weberian theory explains more fully the pattern of industrial location. They reject the exclusive zone concept in favor of a comparative cost approach with intermixing of land uses. A Concept of Urban Decentralization Decentralization of economic activity was analyzed by Moses and Williamson.^ They argued that the tie to the core is transportation within the urban area and from dis- tai.; locations. Implicit in this reasoning is the impor tance of agglomeration effects at the core. Deglomeration effects or decentralization effects are the savings in factor costs on land and labor. It is the reduction in transportation costs that alter location patterns. The consideration of von Thunen and Weberian location theories indicates the important variables in intraurban location decisions. Decentralization occurs as 20 transport costs are reduced and is characterized by more dispersed employment densities and mixed uses at various locations. It can best be analyzed in a comparative static manner utilizing the received theories of industrial location. 21 FOOTNOTES 1. Johann Heinrich von Thunen, Von Thunen's Isolated State, trans. Carla N. WartenbergT ed. Peter Hall (Edinburgh: Pergamon Press, 1966). 2. Alfred Weber, Theory of the Location of Industries, trans. Carl J. Friedrich (Chicago: University of Chicago Press, 1929). 3. Raymond L. Fales and Leon N. Moses, "Land-Use Theory and the Spatial Structure of the Nineteenth-Century City," Regional Science Association Papers 28 (1972):49-80. 4. Richard F. Muth, Cities and Housing (Chicago: Univer sity of Chicago Press, 19(59) . 5. William Alonso, Location and Land Use: Toward a General Theory of Land Rent (Cambridge: Harvard University Press, 1964). 6. Richard F. Muth, "Economic Change and Rural-Urban Land Conversions," Econometrica. 29(January 1961):1-23. 7. Lowdon Wingo, Transportation and Urban Land (Balti more: Johns Hopkins Press, 1961). 8. Michael Goldberg, Intrametropolitan Industrial Loca tion: Plant Size and the Theory of Production (Berkeley: Institute of Urban and Regional Development,University of California, 1969). 9. Leon N. Moses, "Towards a Theory of Intra-Urban Wage Differentials and Their Influence on Travel Patterns," Regional Science Association Papers, 9 (1962):53-63. 10. Fales and Moses, "Land-Use Theory," pp. 49-80. 11. Ibid., p. 52. 12. Ibid., p. 53. 13. Leon N. Moses and Harold F. Williamson, "The’Location of Economic Activity in Cities," American Economic Review, 18 (May 1968):211-222. CHAPTER III A MODEL OF SPATIAL EQUILIBRIUM Assumptions of the Analysis This chapter will present a formal model of a core dominated city. This type of model is used most often to generate housing patterns as in Muth^ and Alonso.2 The concept was first used by von Thunen^ to determine land use and rents with particular emphasis on agricultural uses. In this analysis land use is restricted to indus trial and housing activities within an urban area. No attempt will be made to introduce agricultural uses. The theory to be presented is quite general and provides useful insights into the economic factors that influence urban decentralization. The first part of the analysis on land rent gradients draws heavily on the scheme of analysis utilized by Muth.^ The analysis of wage rate gradients closely follows the presentation by Moses.^ It will be argued that land rent gradients emerge due to the cost of transporting goods in the urban area and that wage rate gradients are derived from the cost of moving people for work trips. The equilibrium location of a firm is argued to be related to 22 23 the land rent and wage rate gradients and the cost of moving intermediate goods. Capital costs are assumed invariant with distance and, as such, do not directly affect the location decision of the firm. The model assumes the form of the von Thunen hypothet ical city where the city is divided into two subregions, a centralized production and distribution zone surrounded by a satellite area in which workers and their families live. All goods for local consumption and export are produced in the core area. Some goods for local consumption are pro vided indirectly by the core. These goods are produced elsewhere and channeled to households and firms through the transport and service industries at the core. All inter- urban transport flows into the core and all intraurban transport is tied to the core. Industry is perfectly competitive and the prices of goods at the core can be treated as given. Firms utilize the services of labor, land, capital, intermediate goods, and raw materials. Intermediate goods are available at the core at zero transport cost from other firms located there or from the interurban transport facil ities. Raw materials are assumed to be everywhere avail able, ubiquitous, and incur no transport costs. Capital, in the form of buildings and machines, is assumed to be available if credit service industries are present. Credit is extended at the same rate to all locations in the urban 24 area. In order to pay for goods households provide labor services to the core area producers. All laborers are treated as identical in skills and attributes and form a single competitive market in the urban area. Land in the urban area is assumed to possess homogene ous physical characteristics in all directions from the core. Goods produced or imported in the core area are dis tributed from the core to the outlying market areas. The transport of these final consumption goods incurs the fric tions of space. The transport cost curve is the envelope of delivery costs to all distances from the core. It is assumed that transport costs increase at a decreasing rate as distance from the core increases. Since direction has no effect on transport costs the result is a transport plane. Equilibrium in the Land Market All households are assumed to be identical. As such they have identical utility functions for all goods (in cluding housing) and time (including leisure time). Since the set of prices on goods is given, for households to be in equilibrium the resulting level of utility must be the same for each household. For households to be in full, spatial equilibrium, utility at all distances from the core must be equal. This implies that three conditions must be satisfied: 25 (A) households are in traditional price equilibrium; (B) households are in spatial equilibrium? (C) and the land market is in equilibrium. In traditional price equilibrium ratios of marginal utilities to prices for all goods must be equal. Alterna tively, the ratio of the marginal rate of substitution for any two goods is equal to the ratio of prices for the same two goods. Since the income of all workers is the same and the utility functions are identical, these conditions imply that every household is on the same indifference curve at equilibrium. Goods are not available at all locations at the same price as at the core. As distance from the core increases the price of goods rise due to transport costs on these goods. To equalize net incomes and, hence, overall utility at all distances from the core, the price of housing can vary with distance. Households are in spatial equilibrium such that at their optimal locations the saving in housing costs from a small change in distance must exactly equal the change in transport costs. This condition is repre sented as, (1) -qp’(d) = T'(d) where q is the household’s consumption of housing, p is the price per unit of housing, T is the transport cost of goods, 26 and d is the distance from the core area. Both p and T are a function of d. Housing, according to Muth,6 is "the bundle of con sumer services supplied both by structures and by the land on which they located." Since T1(d)>0 and q is always positive, except at the core, p'(d)<0. Since housing is a normal good, at equilibrium, the per capita consumption of housing, q,increases with distance, i.e., Sq/6d>0. The change per unit distance in q depends upon the real-income- constant price elasticity of demand for housing and the change in price per unit distance. It can be shown that the price of housing, p, declines at a decreasing rate with distance. According to Muth, the condition that the net saving on housing and transport costs per unit increase in distance not increase with dis tance implies that, (2) p"(d) > -(1/q) [p* (d) (5q/6d) +T"(d)] Since q is positive, p1 (d)<0, and <Sq/6d>0, then if T1 ' (d)<0 the results must be that p1 ' (d)>0. Equilibrium in the land market can be derived by looking at the market for housing. Producers of housing are assumed to have identical production functions using land and physical capital as factors of production. The price per unit of the composite bundle of capital is assumed to be the same at all locations. In a competitive market, 27 the first order profit maximizing conditions for housing producers show that the unit price of land, r(d) is deter mined such that, (3) r' (d) = (q/R) p'(d) 7 where R is the amount of land used in housing. Since q/R>0 and p'(d)<0, r'(d)<0. Thus if the price of housing falls then the rent per unit of land must fall with distance from the core. Given the decline in land rents with distance from the core, the output of housing per unit of land declines with distance as firms substitute land for non-land factors so as to maximize their incomes. The change in output per unit of land with distance depends upon the elasticity of substitution of land for non-land factors and the rate of change in rent per unit distance.8 By differentiating equation (3), the rent on land falls such that r''(d) is positive at equilibrium. That is, land rents, like the price of housing, decline at a decreasing rate with distance. Thus an equilibrium land rent gradient emerges as in Figure 1. For households to be in full spatial equilibrium the model implies that the price per unit of housing, rent per unit of land, and output of housing per unit of land all decline with distance from the core. The consumption of housing increases with distance from the core. The impli cation is that the density of settlement, households per unit of land, declines as distance from the core increases. 28 Land Rent Distance from Core Figure 1. Land rents at equilibrium That is, more land per unit housing is used at greater distances. Equilibrium in the Labor Market The analysis so far has only considered the cost of moving goods from the core to the household markets. Households, however, also interact with the core in another market, that of labor services. For labor, the market is initially at the center area of the city. All workers in cur some cost of travel to commute to their work place. Since all workers are identical in their utility for lei sure, as well as other goods, the equilibrium conditions must allow for time costs of travel. Once the equilibrium distribution of households and 29 the resulting land rent gradient are established, the wage rate gradient can be determined. Labor services, unlike the services of land, are not perfectly immobile. "Labor can be drawn to any given point from all other points in the city, depending on wage rate differentials, and the time and money cost of t r a v e l . it is assumed that all workers are identical in skills and work a fixed and equal number of hours with an equal number of work trips to the core area. All workers earn an identical gross income at the core where the wage rate is given. Under conditions of an equilibrium spatial distribution of households, the price of residential space varies in such a way as to precisely compensate for variations in net income, i.e., levels of house hold satisfaction do not vary spatially. Neverthe less it is true that a worker who could reduce his travel expenditure by changing the location of his job, residence held fixed, would improve his level of satisfaction. Net income, which is gross income less transportation expenditures, must also be adjusted for work trips. Net income depends upon the structure of intraurban transport costs for goods and people and the distance separating residence from core and residence from work place. In this case, the work place is the core. The entire urban area is a single labor market and any individual can be drawn from any location in the urban area for work at the core. Yet it is possible to induce a worker to accept a job at some lower wage, between his residence and the core area. 30 W w Wage Rate O W Distance from Core Figure 2. Wage rates with constant commuting costs Two different transport costs for moving people can be described yielding two distinct wage rate patterns. Suppose money commuting costs are a positive constant over the urban area. This case resembles the pricing structure in public transportation where waiting time and travel time are ignored. In Figure 2, OW is the core area wage rate and WW' is the constant transport cost. Every individual is willing to work at home for OW' for which he incurs zero transport cost. Ke will accept nothing less than OW at any other location. The effective wage rate gradient is a horizontal line from W where work place is not at the resi dence. Even for the case of reverse commuting, going to work away from the core, this same wage rate gradient is in 31 effect. This gradient assumes that the utility of leisure or the disutility of travel time is zero. Otherwise the wage rate gradient will have some downward slope toward the point of residence. Since the value of time is likely to be significant this wage rate gradient can be accepted if work hours include the time spent in travel. Workers com muting from longer distances thus work less time than others once they reach the core area. A second transport cost structure combines the notion of money costs and time costs where work hours are held fixed among all workers. Suppose money costs of travel are linear with distance. The use of automobiles traveling at a constant speed or with small variations in average speed has this type of cost structure. Also, suppose that time costs are a function of distance and the speed per unit of distance. Increasing distances increase the time cost as does decreasing speed per unit of distance. Since the density of settlement declines with distance from the core, the density of commuters increases as the core is approached. That is, increasing congestion is incurred thereby lowering the velocity per mile and increasing the unit cost of travel time as the core is approached. As suming that the reduction in speed has no effect on the money costs of travel this represents an increase in the time cost of travel as the core is approached. 32 Commuting Cost P 0 Distance from Core Figure 3. Money and time commuting costs The situation is represented by the transport cost segments AB, CD, EF, etc., in Figure 3. Mosesll calls the cost of a round-trip to the core— commuting costs— for a worker whose residence is at A, twice the sum of all of the segments. The commuting cost for residence at C is ob tained by ignoring transport costs incurred by households further from the core, such as segment AB. The total cost of commuting from each location is the curve OM which in creases at a decreasing rate as distance from the core increases. A second wage rate gradient is obtained by sub tracting the commuting cost, OM, from the core area wage rate, OW, in Figure 4. The commuting cost is represented by the vertical distance between WW and WW1. WW' is the 33 Wage Rate A O E C J G L Distance from Core Figure 4. Wage rates with increasing commuter costs new wage rate gradient. The curve indicates the wage rates at which a worker would be indifferent as between employ ment at home, at the core, and at any other point between the residence and the core. The implication is that an employer who establishes a plant outside the core area can offer a lower wage. If the wage offer follows along the wage rate gradient only labor from more distant locations will be attracted. The move ment of a single firm would not affect the wage rate gradi ent. If insufficient labor is forthcoming from further distances a premium wage rate above the gradient wage would have to be offered to attract workers from areas closer to the core. In Figure 4, for a firm locating at A, to at tract workers from E the offer must be at least AT. For 34 wage rate AT at location A and OW at the core the worker at E would be indifferent between the two locations for em ployment . Reverse commuting gradients emanate from the wage rate gradient, WW1, as indicated by segments G'S and E'T in Figure 4. Each segment increases at a decreasing rate since less congestion is encountered as distance from the core increases. Workers at A who also live at A receive a premium or economic rent if the wage rate exceeds AW'. For simplicity, at equilibrium all premiums of this type are removed and, hence, reverse commuting is assumed to be nonexistent. Optimal location of work place for the household is achieved when for small changes in distance the change in wage rate offers exactly equals the change in commuting costs such that, (3) -hw'(d) = C'(d) where h is the number of hours worked and w is the wage offer. C, the commuting costs are a function of distance as is the wage offer. w(d), at equilibrium, is the wage rate gradient. As with the land rent gradient, this new wage rate gradient decreases at a decreasing rate with distance from the core. The remaining factor price gradient for liquid capital, the interest rate gradient, can be assumed to be 35 constant throughout the entire urban area. Capital, in the form of credit, is assumed to be perfectly mobile with in the urban area so that the interest rate is everywhere the same, a locational constant. Summarizing to this point, utility is maximized by adjustments in the price of land and labor such that util ity of households, consumers, and workers is a locational constant. The adjustments take the form of differentials from core prices. The differentials increase at a de creasing rate with distance from the core. Spatial Equilibrium of the Firm The locational equilibrium of the firm is achieved when profits of firms are maximized. Up to this point, it has been assumed that firms locate centrally to capture the economies of agglomeration and minimize costs of transpor tation. That this is an optimal location has not yet been determined. Assume that agglomeration economies fall off abruptly with distance from the core such that they are zero for any positive distance from the core. Assume also that these economies are constant for all firms at the core and are figured into product price. Product prices at the core are figured in the land rent gradient and determine the intercept, OR, in Figure 1. Thus agglomera tion economies can be ignored in determining optimal firm location in noncore areas. Since raw materials are 36 assumed to be ubiquitous and the transportation costs on finished goods are priced away in the land market, the only additional costs that firms incur at locations other than the core are transportation costs on intermediate goods. These costs occur since firms must transport the intermediate goods from the core to the production location. Assume that the transport costs on intermediate goods are similar to the transport costs for finished goods and that the receiving firm pays for these costs. S(d), the one-way transport cost for intermediate goods increases at a decreasing rate with distance from the core. Since firms will use land and labor at a new prospective location, savings in the price per unit of land and labor will accrue to the firm with movement away from the core. With output and product price held fixed, profit maximiza tion for the firm becomes a problem of cost minimization. Since only the labor, land, and transport costs (for inter mediate goods) vary with distance, cost minimization deter mines the production location when, (4) -Rr'(d) - hw'(d) = S'(d) where R and h represent the services of land and labor, respectively, and r(d) and w(d) are the land rent and wage rate gradients. For distances greater than the equilibrium distance from the core, the addition to transport costs 37 exceeds the savings in factor costs. S(d) is similar to T(d) in slope. Adjustments to Complete Equilibrium Household location, work place, and firm location are completely determined by the costs of transportation in the model presented. A reduction in the marginal costs of transporting goods and people will disturb the equilib rium. Assume that the reduction in transport costs initially affects only the movement of goods. That is, no effect on the wage rate gradient occurs and households re ceive the full amount of the cost reduction in augmented real income; In this manner, households receive greater amounts of goods at the old prices. Firms, however, re spond to the transport cost reduction by expanding output. But the savings in factor costs at the initial firm loca tion now exceeds the marginal transport costs. Expansion occurs at distances further from the core and ceases when equation (4) is again satisfied. The change in location through expansion and relocation does not occur for all firms. For some firms the reduction in transport costs is not sufficient to break the tie with the core area. That is, the economies of agglomeration still dominate the firm's profits and location decision. Instead, a gradual shifting of firm location occurs which, for the moment, has no impact on the prices of land and labor at 38 Land Rent Distance from Core Figure 5. Shift in land rent gradient increasing distances from the core. This is the type of urban decentralization that is of interest in this study. Firms moving away from the core do so, in part, to save in factor costs. Notice that the land rent gradient in Figure 1 and the wage rate gradient in Figure 4 have their greatest slopes nearest the core. Firms initially located in the core or adjacent to it have the greatest opportunity for factor cost savings. That is to say, the inducement to expand and relocate declines as distance increases. As firms expand and relocate further from the core, the use of factors at more distance locations increases. If enough firm movements occur the prices of land and labor will be affected. In particular, the prices of land and 39 Wage Rate 0 Distance from Core Figure 6. Shift in wage rate gradient labor will be bid up at more distant locations. The result is a decline in the factor price differentials to the core. This change is represented in Figures 5 and 6. Keeping the core area land rent and wage rate constant the new gradi ents become RP1 and WY, respectively. The upward shift of the land rent gradient implies that population becomes more evenly dispersed throughout the urban area. The upward shift in the wage rate gradient implies that congestion costs are less significant around the core area and are spread out more evenly in the urban area. Keeping the land area constant the net change in distance traveled to work should decline as firms move closer to the sources of labor. Some of the reduction in transportation costs is distributed as increased leisure 40 time. Since population is more dispersed, the amount of goods shipped to the furthest distance, A, will increase. It is not clear, however, whether goods are transported over greater distances. What is evident is that the unit cost of moving these goods has declined. The Impact of Distance on Production The upward shifts of these gradients, due to the re duction in transport costs and the resultant relocation of firms, represents a decline in the dominance of the core area. For some firms the benefits of agglomeration econ omies in core area locations become insignificant when transport costs are reduced. In a dynamic sense, reduced impact of core area agglomeration effects is an important part of urban decentralization. Firms that expand at new locations may internalize some economies by adding to their production processes. By engaging in new activities that were previously performed by other firms at the core firms tend to enlarge at noncore locations. Examples of this are adding a legal staff, new processing of raw materials, materials storage, etc. The most significant change in the production process, however, is likely to be the exploitation of economies of scale. Not only is there the opportunity to exploit lower cost factors, but the opportunity to change production technique. The shift to the production of standardized 41 goods with a simple production process is a definite break with the elaborate, small-scaled, unstandardized product process that V e r n o n - ^ claims characterizes the core area. Core area production contains small firms which subcontract a portion of the production process to specialists. The specialists, in turn, demand proximity to other small firms to gain a sufficient amount of business. Large firms in ternalize these functions and are less tied to central location. The change from "batch" production in small plants to continuous production in large scale plants has led to a shift in the use of single level manufacturing plants. The preference is for open sites where horizontal direction is relatively unimpeded. The shift to large scale plant, in general, increases the significance of materials handling and repetitive operations. In the preceding model the price of capital is constant throughout. If the wage rate and land rents in crease relative to the interest rate, at greater distances from the core, then ordinary economics implies that the production process will become more capital intensive as outward expansion ensues. The shift to a higher capital- labor ratio in response to changes in relative prices is an empirical problem which will be assessed in following chapters. The expansion of firms to new locations could shift the potential agglomeration economies away from the core. 42 This might occur if a large producer of intermediate goods exploits economies of scale away from the core. Users of this product might then tend to agglomerate around this source of input. This might bring other suppliers of in puts to each firm away from the core. This indicates that the traditional attraction of core location does not dimin ish per se but is dissipated throughout the urban area via the process of urban decentralization. Gannon^ presents an urban location model with two firms engaged in bilateral trade. The degree of agglomera tion of linked firms depends upon the nature of their pro duction functions. Widely disparate capital-output and land-output elasticities tend to reduce economies of prox imity and linked firms will tend to remain spatially separated. In particular, as the wage rate relative to the interest rate declines with distance, firms with greater labor-output elasticities will tend to locate further from the core. Hence, complete decentralization, i.e., the equalization of all factor prices at all distances from the core, is limited by differences in production functions. What may still emerge, however, are cities with multiple centers each acting as a mini core area containing firms with similar production functions. In the tests to follow, production functions are not observed. Instead firms will be separated by capital and labor production intensities. It is assumed that capital 43 intensive production processes have larger capital-output elasticities and that these firms tend to locate closer to the core. In terms of the hypotheses to be tested, firms that move greater distances are likely to be labor- intensive in production techniques. Alternatively, accord ing to Gannon's model, firms in the process of urban decen tralization become more labor-intensive at greater dis tances from the core. This simple hypothesis emanates from a static model in which relative prices are unchanged. In the process of decentralization wage rates are bid up at greater dis tances. If the price of capital remains unchanged, the relative price of capital and labor has declined. Hence firms are likely to become more capital-intensive at greater distances from the core. This hypothesis contra dicts the previous one so that empirical tests will not reject both hypotheses. In some sense, land, capital, and labor are comple mentary factors of production. Casual observation sug gests that economies of scale, in particular those that are captured in single level plants, tend to be capital- intensive. The hypothesis that economies of scale at greater distances are more prevalent suggests also that capital-intensive methods are utilized. The tests in later chapters attempt to separate scale and intensity effects 44 and, hence, shed some light on their postulated relation ship. Differential Labor Skill Gradients A simple distinction can be made between different types of labor in the urban area. Labor, more than any other resource, has multitudinous distinguishing character istics. Labor can acquire new skills, more productivity through experience, better health, etc. It is useful, therefore, to regroup labor into qualitatively different groups. For the sake of analysis, consider two broad groups, high skill and low skill labor. High skill labor is characterized by possessing greater experience, by being more productive, and by being more specialized than low skill labor. High skill labor is dominant in the profes sions, specialized services, and manufacturing industries such as electronics and chemicals. For high skill labor the ability to make a living is greatly enhanced by urban ization. His skills are more efficiently and extensively utilized with a high degree of agglomeration and elabora tion in production. Low skill labor possesses qualities that are not too important for urbanization. That is, these skills can be used just as efficiently in agriculture. Likely employ ment of low skill labor is in standardized manufacturing plants where the functions are simple and repetitive. 45 These skills improve when working with capital equipment, but are limited in the alternative work this labor can do. Certain types of maintenance, repair, and service jobs fit this category. Yet it could be argued that the techni cal potential to substitute capital for low skill labor is great compared to the ease of substitutability for high skill labor. A different wage rate gradient for each skill cate gory can be derived. Assume that high skill labor commands a higher wage rate at every location. The principal cost difference between the two skill groups is the cost of com muting. Assuming they each use the same mode of travel with the same money cost-distance function the commuting costs will differ due to the time costs of travel. Two laborers, differing in skill, living at the same location will be impacted differently by congestion. For the low skill worker the opportunity cost of time is less than that for high skill worker due to the competitive difference in skills and income. The same congestion yields a higher cost of travel to the high skill worker. This results in two wage rate gradients as in Figure 7. HH1 and LL' are the wage rate gradients for the high and low skill labor, respectively. At any location the rate of increase in the slope of HH' is greater than that for LL' reflecting the different time costs of commuting. This implies that at some point beyond E, in Figure 7, where the slopes of 46 Wage Rate O E Distance from Core Figure 7. Wage rate gradients for high and low skill labor HH' and LL' are equal, the price of high skill labor rises relative to that for low skill labor. The nature of the differential impact of labor skills on commuting costs determines how close the point E is to the core. The tests to follow will indicate whether this differentiation in skills is significant and whether the point E is fairly close to the core. 47 FOOTNOTES 1. Richard F. Muth, Cities and Housing (Chicago: Univer sity of Chicago Press, 1969). 2. William Alonso, Location and Land Use: Toward a General Theory of Land Rent (Cambridge: Harvard University Press, 1964). 3. Johann Heinrich von Thunen, Von Thunen1s Isolated State, trans. Carla N. Wartenberg, ed. Peter Hall (Edinburgh: Pergamon Press, 1966). 4. Richard F. Muth, "The Spatial Structure of the Housing Market," Regional Science Association Papers, 7(1961):208-220. 5. Leon N. Moses, "Towards a Theory of Intra-Urban Wage Differentials and Their Influence on Travel Patterns," Regional Science Association Papers, 9 (1962):53-63. 6. Muth, "Spatial Structure," p. 210. 7. Housing producers maximize profits only with the price of housing allowed to vary with distance. This is different from the parameters faced by producers of other goods, as shown below. 8. Muth, "Spatial Structure," p. 211. 9. Moses, "Theory of Intra-Urban Wage Differentials," p. 54. 10. Ibid., p. 55.' 11. Ibid., p. 57. 12. Raymond Vernon, "Production and Distribution in the Large Metropolis," The Annals of the American Association of Political and Social Science, 314(November 1957):15-29. 13. Colin A. Gannon, "Intra-Urban Industrial Location Theory and Inter-Establishment Linkages," Geographical Analysis, 5(July 1973):214-244. CHAPTER IV EVIDENCE OF DECENTRALIZATION IN THE TECHNOLOGICAL EPOCHS Changes in the Mode of Transportation The decentralization of economic activity in major metropolitan areas is generally considered to have begun in the early nineteenth century. Single, core-dominated cities evolved into suburbanized, multi-centered metropol itan areas in the present century. Development of areas like Los Angeles have been characterized by low and rather uniform population densities. Industry, banking, and other services that were core oriented in other cities in the past tend to be broadly dispersed in Los Angeles. This im plies that the von Thunen type analysis is less useful in explaining patterns of urban land use in the current cen tury. Fales and Moses,^ in analyzing the land use pattern in Chicago after the great fire, utilize a modified von Thunen-Weberian approach. They reject the exclusive zone concept and apply a model based largely on the structure of transportation, agglomeration and deglbmeration effects, and economies of scale. The model presented in Chapter III incorporates their approach with implicit emphasis on each 48 49 of these variables. Yet transportation costs appear as the key variable affecting the nature of urban decentralization. The use of automobiles is generally considered to have caused the low, uniform density development in this century. It has been argued that automobiles facilitated the spreading of population by improving the movement of people. Yet in Los Angeles, prior to 1920, significant clustering of housing in and around the central area occurred. PegrumJ notes that economic activity clustered around the interurban transport links and the major intra urban railroad links with the core area. Single family residences with fairly even-sized lots gave the effect of uniform density in the early twentieth century. Multi family dwellings are a more recent phenomenon in the Los Angeles area. The orientation of economic activity in the core area and its major transport arteries suggests that potentially significant decentralization might have occurred in Los Angeles in this century. The hypotheses offered by Moses and Williamson to explain nineteenth century core dominance might also be applied to twentieth century cities. In the nineteenth century, they argued, "the cost of moving goods within cities was: (1) high relative to the cost of moving people within cities; and (2) high relative to the cost of moving goods between cities."^ The method of moving goods in Los Angeles around the 50 turn of the century was by rail and horse and wagon. People were transported to and from jobs principally on rail trolley cars. Truck haulings replaced the horse and wagon for intraurban goods movement and later replaced railroads for intercity goods movement. As the city grew in size the use of the automobile and the truck facilitated the disporportionate growth of the Cutlying areas. Prior to these changes, central location of economic activity in the goods handling areas dominated the activity pattern in Los Angeles. The Decline in Importance of Central Location Moses and W i l l i a m s o n , 5 P e g r u m , 6 and Fales and Moses^ point out that central location was important when a con siderable amount of goods were received from, and shipped to, other regions at a central location. "The economies of scale in rail transport were such that the receiving and sending of such shipments were concentrated at one or at most a few large, centrally located freight terminals."8 Vernon^ observes that during urban growth the break with the center resulted in the local market being the chief source of manufacturing output. Less reliance was placed on interurban transport of inputs. The shift to truck in intracity travel allows the firm to move closer to the market and be less locationally reliant on sources of in puts. As cities grow, they become less dependent on trade 51 with other areas and are able to exploit economies of scale in local markets. Scale economies are more fully exploited in satellite areas. Hence, the large urban areas evolve to self-sufficiency. If the relative cost relationship played a crucial role in the dominance of the core, only after technological changes occur in transportation will the attraction of non core areas be felt. The major change was the introduction of the truck which reduced the cost of moving goods within cities. Its effect on the spatial structure of cities can, roughly, be divided into two phases. During the first, the motor truck was introduced and became the dominant form of intraurban carriage, but interurban carriage was still done by railroads. In this period — the first two decades of this century— firms could leave the core but were still tied to it for ship ments to and from other regions. This tie was weak ened during the second phase when improvements in the truck and in the interregional highway system meant this mode could be used for long-distance transport. The full impact of this change was prob ably not felt until the revival of a strong peace time economy after World War II. The attractiveness of the satellite area in this period was increased by the automobile which allowed firms to draw labor from a broad area.10 Other changes might also account for the attractive ness of noncore areas, such as improvements in data proc essing and communications. Meyer, Kain, and WohlH have pointed out that these changes meant that firms could re main "near" other firms though they had moved miles away. Evidence of Decentralization Two periods, roughly corresponding to the timing of major changes in transportation technology, were selected 52 for analysis. Data on the movement of manufacturing firms within Los Angeles County was accumulated for the periods 1908 to 1920 and 1949 to 1964. At the beginning of the first period the major transportation change was the intro duction of the truck into intracity travel. At the begin ning of the second period and continuing throughout this period the major transportation change in Los Angeles was the construction of intraurban freeways. Considerable effort also went into the construction of intercity high ways connecting Los Angeles with other major cities in the western United States. Implicit in the selection of the time periods of ob servation and the length of these periods is the notion that there are lags in the reaction to changes in trans portation technology. Capital investments in new plants and facilities are generally required for most plants to relocate. The investment plans require some period of evaluation and study before even finance arrangements are made. Further, the basic model in Chapter III indicates that firm relocation is not accomplished all at once. As firms move the nature of agglomeration effects and inter firm relationships change suggesting that firm relocation to specific areas begins slowly, accelerates, and then tapers off as externalities become diffused amongst com peting areas. The length of the two periods allows for considerable adjustment to the transport cost changes. 53 In both periods only two observations are made, the location of the firm at the beginning date and its location at the end date. As such, the speed and frequency of relocation over the time periods cannot be observed. It can be argued that firms relocating in 1909 are not facing the same factor price conditions as in 1919. That is, it is not possible to gauge the intensity of relocation activ ity in any year of the period. It is possible that reloca tion responses come to a climax some time prior to the end of the period. Since it is a very costly process to time the movements of firms the assumptions of adjustment will have to suffice. The analysis of relocation for the manufacturing sector is made since it is the largest employer in most metropolitan areas and has had the largest central city decline in employment. Furthermore, manufacturing is subject to economies of scale where specialized output and elaborate production processes are less dominant. Service industries, such as banking and insurance, fashion, and retail trades are not as standardized in output nor as simple in production. Also, manufacturing processes re quire a consistent set of specialized input over time. That is, the set of externalities associated with the production of manufactured goods is consistent. 54 Decentralization in the Early Period The early period and the later period correspond to the Moses and Williamson test periods. The two sources of evidence should be directly comparable. In the early period the evidence is fragmentary on the proposition that the introduction of the truck reduced the cost of moving goods relative to the cost of moving people. The cost of moving people seemed fairly constant throughout this period. Data up to World War I from the American Transit Association^ indicates the passenger far was "almost universally five cents in this period."14 The evidence on truck versus horse and wagon costs indicates that there appeared to be a lower transport cost for trucks. A 1918 Department of Agriculture survey found that the horse and wagon costs were $.33 per ton-mile and truck costs were $.15 per ton-mile.15 Another survey in 1920 found average hours per ton-mile were one-third to one-half as great for motor truck as for horse and wagon.16 The truck appears to have had a time as well as money cost advantage, partic ularly as the distance traveled becomes longer. The higher average speed of truck travel enabled trucks to haul more goods in less time than the horse and wagon. The average cost per ton-mile for trucks declines as distance increases while the cost per ton-mile increases rather dramatically for the horse and wagon. There is no record of truck registrations in Los Angeles during this period. However, 55 TABLE 1 SAMPLE BREAKDOWN FOR BOTH PERIODS 1908-1920 1949-1964 Number of Firms Percent age Number of Firms Percent age Total sample 135 100 440 100 Movement 38 28 130 30 No movement 97 72 310 70 Movement 38 100 130 100 Outward 27 71 86 66 Inward 9 24 44 34 No change 2 5 0 0 the use of trucks seems to have increased and the use of horse and wagon decreased. Moses and Williamson^? found this to be the case for Chicago during this period. The evidence of decentralization from 1908 to 1920 for manufacturing firms in Los Angeles is based on a sampling of 135 firms. Names of manufacturing firms and their addresses were compared for the years 1908 and 1920.18 Of the 135 firms sampled, 38 firms moved during this period. The movement was either a relocation or an expansion to a new location. Each firm location was plotted on a street map and the point to point difference between the core area point and the addresses were 56 TABLE 2 DISTRIBUTION OF ORIGINS AND DESTINATIONS, 1908-1920 Origins Destinations Net Change Distance from Core (in miles) Number (1) Percent of 1908 Firms (2) Number (3) Percent of 1920 Firms (4) Column (3) minus Column (1) (5) 0.0 to 1.5 12 31 6 19 -6 1.6 to 2.5 7 27 5 21 -2 2.6 to 3.5 5 25 5 25 0 3.6 to 4.5 5 28 6 32 1 4.6 to 5.5 4 28 5 33 1 5.6 to 6.5 2 25 3 33 1 6.6 to 7.5 2 25 3 33 1 7.6 to 8.5 1 33 5 71 4 calculated. The core area was defined as the center of the Downtown Los Angeles District, the intersection of 1st and Main Streets. For each firm, the 1908 location was measured in terms of the distance from the core, called the origin distance. For the firms that moved, the new address distance was measured also from the core. The new location was called the destination distance. Of the 38 relocating firms, 27 moved away from the core— destination distance exceeded origin distance— and nine 57 moved inward. Two firms that moved did not change their distance from the core. The first two columns of Table 1 give the sample breakdown for the early period. If factor savings are an important determinant of movement and if the land rent and wage rate differentials increase at a decreasing rate with distance, the hypothesis consistent with these assertions is that firms closer to the core are affected more by the transport cost shifts. The hypothesis is consistent with the notion that the costs of expansion are higher for firms closer to the core. Thus, firms initially closer to the core have a greater in ducement to move. The data in Table 2 for the early period shows roughly, that, as distance from the core increased the net change in the movement of firms increased. For distances close to the core the number of origins exceeded the number of destinations. Fairly uniform net decentral ization has occurred in this period as indicated in the column 5 of Table 2. Column 1, Table 2 shows that the number of origins declines with distance. Yet it cannot be concluded that the propensity to move is greater the closer to the core a firm is located. If the number of origins is divided by the total number of firms originally found at varying distances, this ratio is fairly constant as indicated in Column 2, Table 2. That is, although the movements emanating from distances near the core are greater, the 58 number of firms located there are also greater. The notion that firms at closer distances have a higher pro pensity to move must be rejected. In a similar fashion, the constancy of the number of destinations at each distance does not imply that these distances possess equally desirable characteristics. As a proportion of the firms located at these distances in 1920, this measure increases as shown in column 4 of Table 2. The evidence supports the notion that relocation is not spurious, but, rather, systemic influences are at work. In the light of this data, it is argued that relocation patterns emerge from the savings in factor costs at greater distances from the core. A comparison was made between the average origin distance of moving and nonmoving firms. The former firms moved, on the average, from a distance of 2.9 miles from the core. The latter firms that remained in their loca tions were located, on the average, 3.8 miles from the core. The difference between the two distances is signifi cant at the .01 level. This evidence supports the factor savings hypothesis that firms closer to the core are in duced to move. Hence, for decentralization of the type expected to take place, the net change in the number of relocating firms must increase with distance and the mean origin distance must be lower than that for nonmoving firms. 59 TABLE 3 DISTRIBUTION OF ORIGINS AND DESTINATIONS, 1949-1964 Origins Destinations Net Change Distance from Core (in miles) Number (1) Percent of 1949 Firms (2) Number (3) Percent of 1964 Firms (4) Column (3) minus Column (1) (5) 0.0 to 1.5 29 29 13 16 -16 1.6 to 2.5 26 29 16 21 -10 2.6 to 3.5 24 30 15 21 -9 3.6 to 4.5 19 28 18 28 -1 4.6 to 5.5 15 31 17 33 2 5.6 to 6.5 9 30 18 46 9 6.6 to 7.5 5 31 19 59 14 7.6 to 8.5 3 27 14 64 11 Decentralization in the Later Period To seek additional evidence on this type of decen tralization, data was generated for the period 1949 to 1964. Addresses of 440 manufacturing firms were compared in a similar fashion as in the early p e r i o d .19 The core area point was the same as in the early period. Of the 440 firms, 30% or 130 moved over the period, which repre sents a frequency of movement of only 1.8% greater than in 60 the earlier period. The percentage of moves that were outward declined to 66%. That fewer movements were ob served in a direction away from the core area suggests that some of the dominance of the traditional core had receded by this time. That more firms moved toward the center might indicate they are likely moving away from other small centers that appeared between the two sample periods. Whatever this motivation was, the incidence of outward movement still dominates. The sample breakdown is shown in the last two columns of Table 1. Table 3 indicates the distribution of origins and destinations for the 1949-1964 period. Net change in the number of firms tends to increase with distance as indi cated in column 5 of Table 3. In the later period, significant net decentralization also occurred. As in the early period the ratio of the number of origins to the total number of firms remained fairly constant while the ratio of the number of destinations to the number of firms increased with distance. The mean origin distance of nonmoving firms again was significantly greater than that for moving firms. The difference was 4.4 to 3.5 miles. In the later period, net decentralization of the type hypothesized did, indeed, occur. 61 Comparison of Results in the Two Periods These two period results compare favorably with the results for Chicago presented by Moses and Williamson.20 In the early period they found that the average origin dis tance was less for moving firms. In the later period they found a similar distribution of origins to the one pre sented here. Also the ratios of origins to number of firms remained fairly constant with distance in their study. While no data is presented by these authors for distance moved from the core, the average distance moved from the core in this study was 2.1 miles and 3.6 miles for the early and later periods, respectively. This latter result bears on the whole matter of growth and technological change in transport technology. In the later period the impact of intraurban freeways had considerable effects throughout the urban area. Areas at greater distances from the core became more accessible to commuters and shoppers. Both the costs of moving goods and people were reduced. This facilitated the tremendous urban sprawl characteristic of the Los Angeles area. This result also indicates that decentralization of manufacturing was a continuing process in the interim span of time. While the noncore area factor price differentials had been reduced through decentralization, the attraction of more distant locations was enhanced by the dispersion of economic activity. A greater distance of relocation 62 represented less of a break associated with the core area attractiveness and more opportunity to exploit scale econo mies. That is, even though the marginal savings in factor costs declined at each distance, so did the marginal cost of transportation. The evidence suggests that the latter reduction was greater than the former. Markets for goods became more dispersed and the re duction in transport costs allowed firms to shift location as the orientation of the market shifted. The orientation toward local markets, as Los Angeles grew in population and size, induced firms to move greater distances.21 The difference between origin distance for the two periods suggests a long term pattern of decentralization. In both periods the firms closer to the core tended to move. The firms that remained stationary at their location in the early period might have been induced to move in the later period. To establish this leaping pattern, many more observations using different time periods must be taken. The direction of manufacturing movements differed significantly in the two sample periods. The dominant movement away from the core in the early period was to the west. This move was in the same direction as the major rail tie to the central city and also with the major surface streets. Some southward movement occurred along the major rail and surface ties to the south. The net 63 change, however, was away from the railroad routes toward major surface s t r e e t s . In the later period movements occurred in all direc tions from the core. Industries moved to the southeast and the northwest as well as to the south and southwest. These moves occurred predominantly along the new and proposed freeway construction paths. The larger manufacturing plants moved to the south, southeast, and northwest where the newest populations came to reside.^3 The impact of the automobile has often been empha sized as the primary cause of urban sprawl. Yet the evi dence from these two epochs suggests that other plausible stimuli to decentralization has occurred. Automobiles do not directly explain the movement of manufacturing firms except insofar as they affect wage rates. There are several key elements in reducing the tie to the core area. Automobiles is but one parameter. 64 FOOTNOTES 1. Raymond L. Fales and Leon N. Moses, "Land-Use Theory and the Spatial Structure of the Nineteenth-Century City, Regional Science Association Papers, 28(1972):49-80. 2. Beverly Duncan, Georges Sabagh, and Maurice D. Van Arsdol, "Patterns of City Growth," The American Journal of Sociology, 67(January 1962):423. 3. Dudley F. Pegrum, Transportation: Economics and Public Policy (Homewood: Richard D. Irwin, Inc., 1963), p. 560. 4. Leon N. Moses and Harold F. Williamson, "The Location of Economic Activity in Cities," American Economic Review, 18(May 1968):212. 5. Ibid. 6. Pegrum, Transportation, pp. 555-568. 7. Fales and Moses, "Land-Use Theory," p. 61. 8. Moses and Williamson, "The Location of Economic Activity," p. 213. 9. Raymond Vernon, "Production and Distribution in the Large Metropolis," The Annals of the American Association of Political and Social Science, 314(November 1957):15-29. 10. Moses and Williamson, "The Location of Economic Activity, p. 213. 11. John Meyer, John Kain, and Martin Wohl, The Urban Transportation Problem (Cambridge: Harvard University Press, 1965). 12. Vernon, "Production and Distribution," p. 19. 13. American Transit Association, Fare Structures in the Transit Industry (New York:n.p., 1933). 14. Moses and Williamson, "The Location of Economic Activity," p. 214. 15. George W. Grupp, Economics of Motor Transportation (New York: Appleton and Company, 1924). 65 16. U.S. Department of Agriculture, Bulletin No. 910 (Washington, D.C.: Government Printing Office). 17. Moses and Williamson, "The Location of Economic Activity, p. 214. 18. Directory of Manufacturing, 25 vols. (Los Angeles: Chamber of Commerce, 1904-1934), 3(1908) and 12(1920). 19. California Manufacturers Register (Los Angeles: California Manufacturers Association), 1949 and 1964. 20. Moses and Williamson, "The Location of Economic Activity," pp. 215-222. 21. See Duncan, Sabagh, and Van Arsdol, "Patterns of City Growth," pp. 418-428 and Arthur Grey, "Los Angeles: Urban Prototype," Land Economics, 35(August 1959):232-242. 22. Pegrum, Transportation, p. 560. 23. Fred Case and James Gillies, "Some Aspects of Land Planning: The San Fernando Valley Case," The Appraisal Journal, 23(January 1955):15-41. CHAPTER V AN EMPIRICAL TEST OF DECENTRALIZATION The Characteristics of Decentralization in Los Angeles, 1949-1964 During the 1949-1964 period the Los Angeles metropol itan area! underwent substantial changes. Estimates of pop ulation growth show an annual change in population at about 12% in the 1950's.^ The suburbanization that began in the early twentieth century accelerated during this period. Productive agricultural land to the west, northwest, and southeast of the central core area was converted into hous ing. The post World War II housing boom was characterized by uniform construction as to square feet of interior space and lot size. It is not surprising that Muth observed fairly uniform population densities for Los Angeles 1950 3 census data. The population growth and the nature of new housing construction contributed to urban sprawl in the Los Angeles area. The uniformity of new housing extended, as might be expected, to uniform selling prices in the newly developed areas. This, in part, induced the concentration of popula tions which were homogeneous in other respects. The new 66 67 housing, located at greater distances from the core than old housing, was inhabited by people of fairly uniform family size, income, age composition, and occupation.^ These new communities were only occasionally infiltrated by nonhousing construction.® Since job location, at the outset of the later period, was still concentrated in traditional areas® the demand for new transportation routes connecting residential to industrial areas rose considerably. Prior to and during World War II, industrial activity in the area to the west and south of Downtown Los Angeles grew in response to the war effort. The airplane industry and its associated machine parts, tools, and electronics industries concentra ted production near the coastal and airport areas away from existing residential areas. Much of the postwar growth of Los Angeles is associated with these dominant industries. New Intraurban and Intercity Highway Construction The spatial expansion of housing and the growth in population created a large demand for increased transporta tion facilities. The earliest intraurban freeway construc tion connected the core area to areas in the south, east, and northeast. Commuting travel from the west was still dominated by railroad access and automobile travel along rail lines until the late 40's and early 50's. The ex pansion of intraurban freeways during this period 68 concentrated in connecting the core area to areas in the south, west, and northwest. Intercity and interregional freeway construction connected Los Angeles to major cities in the north and east. These state and national construction projects were so vast that a dramatic increase in intercity truck haul- ings occurred from 1944 to 1968. In 1944, truck hauls were a little over 50,000 million ton-miles. By 1968, this number increased to over 400,000 million ton-miles.^ During the same period, intercity railroad hauls increased from around 650,000 million ton-miles to just over 700,000 million ton-miles.® The ratio of truck to rail haulings increased from about 1 to 7 to 4 to 7. The greatest in crease in this ratio occurred in the western U.S. By the middle 1960's the pattern of intraurban free way routes connected all locations north, east, south, and west of the traditional core. Some routes enabled com muters to bypass the core area entirely. Major employment centers were observed to the southwest, west, and northwest of the core. By the late 1960's several areas had become major employment areas. Yet this pattern gradually evolved. The 1949-1964 period represented a major transi tion phase in the growth of Los Angeles. During this time the central railroad yards handled less and less goods. Truck hauls to and from other cities and between areas in Los Angeles increased dramatically. 69 The Test Movement away from the core in all outward directions is hypothesized to be induced by savings in factor costs. The data in Chapter IV indicate that the inducement to move from any location is fairly constant throughout the urban area. That is, the number of origins as a percentage of the total number of firms is relatively constant for all locations. Although the propensity to move from any location is fairly constant, the ratio of the number of destinations to the total number of firms tends to increase with distance from the core. The straight line distance moved from origin to destination tends to fall off sharply. The average dis tance moved was about 3.5 miles with a low variance. The range of actual straight line distance moved is fairly small. This suggests that the distinguishing characteris tic of firm movement is the destination distance from the core. Thus the pattern of decentralization is due largely to the pattern of firms' destinations. The basic model to be tested is (1) DD = f [w(d) , r (d) ] where DD is the destination distance from the core and w(d) and r(d) are the wage rate and land rent gradients, respectively. As d increases both w(d) and r(d) are expected to decline. 70 Two properties of the factor gradients are of inter est for the test. First, the height of the factor gradient declines with distance from the core. Destination dis tance, DD, is expected to be negatively related to wage rates and land rents at the destination points. Second, the slope of the factor gradients becomes flatter as dis tance from the core increases. Suppose the shift in trans portation costs lowers the marginal cost of travel for goods and people by the same amount at each location. Ad justment by firms to a new equilibrium implies that firms originally closer in will move shorter distances to capture the same savings in factor costs. That is, the marginal savings in factor costs for unit changes in distance is greater at locations nearer the core. Thus, small absolute changes in wage rates and land rents from origin location to destination location are associated with greater destin ation distance. By the late 1940's, several small core areas emerged in all directions from the traditional core. Movements away from the traditional core might reflect movements toward these emerging core areas. The diffusion of exter nalities to other areas may account for the attraction of noncore areas rather than savings in factor costs. A strict interpretation of the core-dominated city theory does not allow for multiple centers. Yet, over time, some centers may become more important as their wage offer 71 increases, as their land rents increase, and as the set of externalities emanating from these centers increase. Thus a smooth reduction in factor costs may not be observed for some directions and distances from the traditional core. The basic test relationship might be modified to include the effects of multiple centers. The incidence of inward movement in the later period^ suggests that multiple centers may, indeed, play a crucial role in firm relocation. In the regression test that follows movement direction variable, MD, takes on the value of 1 if movement is away from the core and -1 value if movement is toward the core. If a center is at some dis tance, say three miles from the core, and movement occurs toward this area, then the coefficient on MD will likely be negative. A negative coefficient implies that movements toward the core and movements away from the core do not cross. Destination converges at some distance from the core. On the other hand, if the coefficient is positive the movements tend to cross, which means that movements in ward represent closer destination distances than movements outward. A second addition to the basic equation is implied by the direction of movement from the core. The majority of relocation movement occurred to the south of the core area. The character of the area to the south of the core changed rapidly during the later period. From a white, 72 middle income community this area changed into a predomi nantly black, lower income community. The production of new housing was sparse as other newly developing areas dominated the growth scene in Los Angeles during this period. Another variable, SN, was added to the basic equation. SN takes on the value 1 when destinations are to the south of the core and zero when the destinations are in nonsouth directions. An examination of movements on a map of Los Angeles clearly shows that this demarcation is operationally meaningful. Finally, the character of production is hypothesized to be sensitive to distance from the core. The exploita tion of scale economies is expected to be greater at in creasing distances from the core. A third variable, S, takes on the value of 1 if certain types of manufacturing firms are subject to economies of scale. A study on Philadelphia is used to classify firms by scale effects.^® S takes on the value of zero in the absence of scale effects. The exploitation of economies of scale at greater distances from the core implies that land use changes in character to single level large plant facilities. The low cost of land is expected to induce this shift in technique and plant configuration. Another production technique change may evolve from the change in relative prices of the factors. In particu lar, capital and labor substitution depends upon the 73 relative factor price change. During a period of growth, it has been argued that the wage rate gradient may shift upward. The interest rate gradient might also shift upward in response to real growth. Thus, it is not a priori obvious which change dominates as decentralization takes place. A fourth variable, FI, takes on the value of 1 if firms are characterized by capital intensive techniques and zero if characterized by labor intensive techniques.H A positive coefficient for this variable implies that firms engaged in capital intensive production are induced by relative factor price changes to set down at greater dis tances from the core. The basic test equation is modified to include the above effects, such that, (2) DD = f[w(d), r (d), MD, SN, S, FI] where MD distinguishes inward and outward moves, SN dis tinguishes south from nonsouth moves, S acknowledges scale effects, and FI distinguishes between capital intensive and labor intensive production processes. Measurement of the Factor Gradients Surrogate measures must be used for land rents and wage rates. Land rents, particularly from multiple uses, are difficult to observe and calculate. The most widely used surrogate measure for land rent is population density. 74 The use of this measure follows from the theoretical notions presented earlier. As distance from the core in creases, population per unit area is hypothesized to de cline in the core-dominated city model. Population density was measured by census tract d a t a . ^-2 M u t h ^ - 3 found that for a sample of census tract areas in Los Angeles in 1950 popu lation declined but not at a decreasing rate from the core. This suggests a rather flat rent gradient existed for part of the later period. Employment data by census tract was not available for this period nor was it possible to calculate for the sample of firms. A possible surrogate measure is labor availabil ity. Yet this is the same as population density and is unacceptable as a distinctive measure. Moses and William s o n ^ used labor availability (population density) as a surrogate for the wage rate gradient. As such, they did not isolate the land rent from the wage rate gradient. The theory presented earlier does not suggest a good surrogate measure as in the case of the land rent. A useful surrogate measure for the wage rate gradient is suggested by some rather complex market phenomena in the urban area. Areas further away from the core tend to have a higher growth rate in population. Newer housing is gen erally concentrated at more distant locations. Housing in or near the central core area is typically depreciated at a lower rate and maintains higher occupancy rates. This is 75 a result of the process of growth and filtering.^ In this theory, older houses justify higher maintenance costs and longer lives by commanding higher rents and higher oc cupancy rates than new housing when adjusted for age of structure. That is, older, more central housing maintains competitive market value at higher occupancy rates. Laborers in adjusting their location to equilibrium commuting costs bid up the occupancy rate at the older, more central, lower growth rate areas. The market main tains a higher inventory of housing, lower occupancy rates, at newer, less central, higher growth rate areas. As transportation costs are reduced, a queue develops at greater distances and a reduction in the occupancy rate differential occurs. This encourages more uniform dis tribution of new home construction throughout the urban area. Thus, the occupancy rate of dwellings is expected to be similar to the distribution of wages in an urban area. The nature of land rents, derived from the prices of hous ing, suggest that occupancy rates decline at a decreasing rate from the core. The occupancy rate plus the vacancy rate for dwellings in each area equals unity. Occupancy rates were used as a surrogate for the wage rate gradient. Occupancy rates and population densities were mea sured at origin location and destination location. The 76 absolute value change in each measure from origin to destination was also measured. This latter measure indi cates the slope changes with changes in destination dis tance. Slope changes in the gradients are expected to be smaller at greater distances. Population density is repre sented by variable PD and occupancy rate by OR. AD, an 0, and a C preceding the variables names refer to destination, origin, and change from origin to destination, respectively. The Empirical Results Data on occupancy rates and population densities were 17 taken from 1960 U.S. Census data. The year 1960 is near the latter part of the time period. The choice of a single point in time near the end of the period was arbitrary. Since the movements were not times the 1960 data represent a single observation for the whole period. That is, some stability in the land rent and wage rate gradients over the entire period is assumed. The regression results in Table 4 show fairly high coefficients of determination, r^, which means that a high percentage of the variation in the dependent variable, destination distance, is explained. The distribution of errors for these regressions were not skewed as the Durbin- Watson statistic range from 1.85 to 2.2. It can be con cluded that the errors are randomly distributed, indicating that there are little systematic errors, like multiple core 77 areas, absent from the estimating equations. Equation (1) in Table 4 shows the results for the sample of 130 relocating and expanding manufacturing firms. The coefficients on the independent variables have the ex pected signs. The effect of the wage rate gradient is captured by the origin occupancy rate, OOR, and the change 18 in occupancy rate from origin to destination, COR. The negative sign for the OOR coefficient means the higher the origin occupancy rate the lower the destination distance. This relationship captures, in part, the hypothesized negative relationship between the wage rate and distance and the observed small range of movement from origin to destination. The negative sign on coefficient for COR indicates the hypothesized slope effects. For small changes in occupancy rate the destination distance is ex pected to be greater since movement occurs over a portion of the wage rate gradient at some distance from the core. The larger the change in slope the lower the destination distance is a direct implication of the negative sign. This result is consistent with expectations. Similar reasoning applies to the sign of the coeffi cients for DPD and CPD. These variables capture the height and slope effects of the land rent gradient. The coeffi cient on CPD is less statistically reliable as might be ex pected by the Muth test on 1950 Los Angeles data. TABLE 4 REGRESSIONS ON DESTINATION DISTANCE OF MANUFACTURING FIRMS IN LOS ANGELES, 1949-19643 No. of Independent Variables Observa tions OOR COR DPD CPD MD SN F FI r2 1. Full sample 130 -6.4 i to • t —* -.79 t . 26b 1.6 -.4 .631 2. Outward move 86 -9.6 -4.2 -.79 -.09° .586 3. Inward move 44 3.9 1.1 -.72 • o -o o .483 4. South move 69 in • 00 I 1 O • ID -.72 -. 42b l.lb .567 5. Nonsouth move 61 -3.6 -3.8 -.80 . 06c 2.6 .601 6. Scale effects and factor intensity 90 -6.8 -1.8 -.64 o 0 CM • 1 1.2 -.7 1.6 .6b .672 aAll coefficients are significant at the .01 level unless otherwise indicated. ^Coefficient is significant at the .05 level. Coefficient is insignificant. 0 0 79 The significant coefficients for MD and SN suggest that the partitioning of the sample by inward and outward movement and south and nonsouth movement might yield inter esting differences in the effects of the factor gradients. The positive coefficient for MD suggests that destination distance is greater for outward moving firms and that in the process of relocating inward moving firms cross the outward moving firms. The negative coefficient for SN im plies that firms moving south of the core did not end up as far from the core as nonsouth movers. The partition tests are shown in equations (2) through (5) in Table 4. The results in equation (2) show the more pronounced effect of the wage rate gradient on outward movement when distinguished from equation (1). In equation (3) the posi tive coefficients for OOR and COR can be interpreted to indicate some significant movement away from a distant core location to again save on labor costs. That is, high oc cupancy rates are associated with a core area. The posi tive relationship between OOR and DD suggests a core area is encountered at some distance from the traditional core. The positive sign for COR indicates that movement occurred away from the distant core location toward the central traditional core area since large changes— occurring near the more distant core— means greater destination distance. The result suggests that firms moved inward for much the same reason as they moved outward, to save on factor costs. An examination of destinations to the south of the core area are shown in equation (4). In comparison to equation (1) south moves were not differently impacted by the land rent gradient. However, the wage rate gradient was significantly different. The same can be said in com paring equations (1) and (5) and equations (4) and (5). The latter comparison shows that firms moved further in distance in the south direction than in the nonsouth direc tion. That is, the difference in coefficients for OOR and COR indicate that firms moved a shorter distance in the nonsouth direction to capture the same savings in factor costs. The wage rate gradient is thus found to be steeper in the nonsouth direction than the south direction for the range of moves sampled. This is not an unexpected result. The available labor supply to the south of the core can be characterized as low income, black and Mexican-American workers with low skill and little alternative employment possibilities. While to the west and northwest labor is generally white, middle to upper income, and specialized skill. These workers can be classified as high skill workers with many alternative employment possibilities. The analysis in Chapter III suggested that the wage rate gradient for high skill labor is more steeply sloped than that for low skill labor. Equations (4) and (5) confirm this expectation. Further partitioning of equations (2) through (5) did not 81 yield any interesting results. In equation (6) out of the sample of 130 moving firms, 90 could be classified by scale and factor intensity effects. In the presence of scale effects the positive coefficient on S suggests that firms exploiting economies of scale relocated at greater distances from the core. This result was anticipated by the theory. The positive coefficient for FI suggests that firms subject to capital intensive production processes tended to relocate at greater distances from the core. This suggests that the price of labor increased relative to that of capi tal at greater distances over this period of analysis. Factor intensity and scale effects thus play important roles in the firm relocation. The other coefficients had the same signs and magnitudes as in equation (1). Parti tioning equation (6) did not yield any more useful results. 82 FOOTNOTES 1. Beverly Duncan, Georges Sabagh, and Maurice D. Van Arsdol, "Patterns of City Growth," The American Journal of Sociology, 67(January 1962):424. 2. Arthur Grey, "Los Angeles: Urban Prototype," Land Economics, 35(August 1959):239. 3. Richard F. Muth, "The Spatial Structure of the Housing Market," Regional Science Association Papers, 7(1961):208- 220. 4. Grey, "Los Angeles: Urban Prototype," p. 237. 5. Fred Case and James Gillies, "Some Aspects of Land Planning: The San Fernando Valley Case," The Appraisal Journal, 23(January 1955):15-41. 6. Hadley E. Smith, "A Review of Plant Location Theory and a Questionnaire Study of Plant Location in Los Angeles County," (M.A. thesis, University of Southern California, 1954). 7. E. S. Lee etal., An Introduction to Urban Decentral- ization ResearcE (Oak Ridge: Oak Ridge National Laboratory, 1971). 8. Ibid. 9. The number of inward moves was 44 out of 130 relocat ing firms. 10. Richard Epps, "Strategy for Industrial Development," Federal Reserve Bank of Philadelphia Business Review, (November 1966):3-l7. 11. Ibid. 12. U.S. Department of Commerce, Census of Population and Housing: Census Tracts, Vol. 82 (Washington, D.C.: Govern- ment Printing Office, 1960). 13. Muth, "Spatial Structure," p. 218. 14. Leon N. Moses and Harold F. Williamson, "The Location of Economic Activity in Cities," American Economic Review, 18(May 1968):211-222. 83 15. Ira Lowry, "Filtering and Housing Standards: A Con ceptual Analysis," Land Economics, 36(November 1960):362- 370. 16. Department of Commerce, Census. 17. Ibid. 18. The decision to use origin occupancy rate rather than the destination occupancy rate was made since OOR entered more equations with significant coefficients and had higher coefficients of determination. CHAPTER VI SUMMARY AND CONCLUSIONS Summary of Analysis The model presented is an analysis of a core dominated city. It serves as an initial starting point in which to view the process of urban decentralization. In this model consumers or households maximize utility and firms maximize profits. Laborers are shown to adjust work location and/or household location in response to differen tial wage rates. At equilibrium, land rents and wage rates and the location of households, firms, and workplace are completely determined by the structure of transportation costs. Goods flow away from the core and people move to ward the core. The cost of moving goods away from the core is subject to economies of scale in quantity shipped and distance traveled. The unit cost of moving goods away from the core declines with distance and quantity of shipment. The unit cost of moving people toward the core increases due to increasing congestion costs as the core is ap proached. Consequently, at equilibrium land rents and wage rates decline with distance from the core at a decreasing rate. 84 85 A reduction in transportation costs for moving people and goods is the major disturbance to equilibrium. Firms adjust equilibrium location by relocating away from the core since at their old location the savings in factor costs exceed the cost increase in transportation for a small change in location. As some firms expand and re locate away from the core, factor prices, at more distant locations, are bid up. At a new equilibrium, firms are more dispersed and factor price differentials between the core and outlying areas are reduced. The major hypothesis of this analysis is that the process of urban decentralization can be initiated by a reduction in transportation costs which induces firms to move away from the core area in order to save on factor costs. Due to the hypothesized shape of the two factor gradients firms originally located close to the core can achieve greater savings in factor costs for a unit move in location. This leads to a second important hypothesis which says that firms closer in to the core will have greater inducement to relocate than firms further away from the core. The model of urban location is simple, yet yields the same implications about changes in residential density and employment density as more elaborate models. Niedercorn^ shows that transport cost reductions imply a shorter journey to work and "causes land rents and employment 86 densities to fall less rapidly as distance increases." The focus.in this study is on the locational response of firms to reductions in transportation costs on goods. Yet it is possible that reductions can occur in commuting costs at the same time. This means that workers closer to the core are earning economic rent at their original loca tions. This increase in real income enables workers to buy more living space which is available at cheaper costs at greater distances from the core. Land rents are bid up at more distant locations and residential densities increase at further distances. As firms offer higher wages to more distant laborers the journey to work distance increases. Similar results were found for the more elaborate model used by Niedercorn.^ A third hypothesis regarding firm size emanates from the work of Goldberg^ and Vernon.^ Land is argued to be more available at cheaper costs at greater distances from the core. As a city grows in income per capita, popula tion, and geographical boundaries, it changes from a net importer to a net exporter of goods. In the process of growth, however, the major market for manufactured goods is the local market. As the local market expands firms at tempt to exploit economies of large scale. They do so by switching to single-level plants requiring a great deal of horizontal expanse. The production process becomes more routinized and less elaborate. The specialized services 87 from other firms found at the core are internalized into the large plant production. Large firms become more self- sufficient in the specialized services and thus have less utility for central location. As transport facilities ex pand and bypass the core area, firms located at more dis tant locations are less dependent on the core for shipments of intermediate goods. Hence, in the process of urban de centralization firms are expected to exploit economies of scale at more distant locations. A fourth hypothesis emanates from the theory of neo classical production. As the relative price of capital to labor changes the use of quantities of capital and labor changes. In the process of decentralization the wage rate increases at more distant locations. If the price of capital, the interest rate, remains fixed, firms are ex pected to shift production to more capital intensive tech niques. However, growth tends to increase the interest rate as well. It is not clear from the analysis by what degree the relative price of labor and capital will change over time with growth and decentralization. This hypothesis is to be distinguished from the static hypothesis which shows the price of labor declining relative to the price of capital at more distant locations at equilibrium. As distance increases firms in the same industry are expected to use more labor intensive processes due to the relative factor price change. What is being 88 examined is the relative price change at each location over time during the process of adjusting to a new equilib rium. If wage rates are bid up further at more distant locations and if the interest rate is bid up the same at each location firms using capital intensive methods are expected to be induced to move greater distances. The analysis shows that labor of different quality would exhibit distinct wage rate gradients. High skill labor would have a gradient more steeply sloped due to the effect of higher time costs of travel. Firms utilizing specialized, high skill labor are expected to move shorter distances in order to capture the same savings in labor costs. Summary of Empirical Results Two technological epochs in transportation were considered. The first period, 1908-1920, corresponded to the period immediately following the introduction of the truck into intraurban travel. The second period, 1949- 1964, included the introduction of intraurban freeways, intercity highways, and the increased use of truck hauls between cities. Observations of manufacturing firm move ments were made for the Los Angeles area in both periods. Decentralization of the type expected occurred in both periods. The origin distance for moving firms was found to be significantly less than that for nonmoving 89 firms. More firms moved outward than inward in each period and that pattern of movement indicated that firms became more dispersed and, hence, employment density became more dispersed. As distance from the core increased the difference between the number of origins and the number of destinations at each location increased. Thus, net urban decentralization of manufacturing firms occurred during both periods. A regression model was tested for the later period highlighting the major hypotheses concerning wage rates, land rents, scale effects, factor intensities, and differ ent labor quality. The results show falling wage rates and land rents with distance from the core. Firms subject to scale economies are shown to relocate at greater dis tances from the core. Capital intensive firms tend to re locate at further distances, indicating that the hypothesis of a locationally constant interest rate and a locationally variant wage rate yield the proper implications for the dynamic changes during the process of urban decentraliza tion. Different wage rate gradients are observed for the area to the south of the core area and for areas nonsouth of the core. Low skill, low income, racially concentrated labor to the south is found to have a flatter wage rate gradient. Firm relocation to the south is at closer des tination distances than to the nonsouth direction. This result captures at least two effects. First, firms origi nally located to the south of the core which also moved southward were initially at small origin distances. Second, the major reductions in transportation costs oc curred to the west, northwest, and east. The firms moving in nonsouth directions from the core were thus faced with much greater reductions in transportation costs. The test results lend evidence toward acceptance of the major hypothesis. It was also possible to identify movement away from distant core locations for the savings in labor costs as well. Cities characterized by multiple centers exhibit less steeply sloped factor gradients as shown in equation (3) of Table 4. There emerges a trough in the factor gradient between centers which is of greater height than in the absence of multiple centers. The com petitive bidding of multiple centers raises the wage rate at all locations and reduces the differential at all non core locations. Concluding Remarks The test by Moses and Williamson^ failed to isolate the effects of the land rent and wage rate gradients. The test in this study separates the two factor gradients and includes the effects of scale and factor intensity shifts in production. The latter is of particular importance in evaluating the overall degree of urban decentralization. 91 P R Rent 0 Distance from Core Figure 8. Shift in land rent gradient with major growth at core Vernon? observed that the character of the central area changes over time. Service industries and retail trades tend to replace the manufacturing activity at the core. More intensive use of land in the core area might result from this replacement as highrise office buildings replace the lower manufacturing plants. That is, although decen tralization of manufacturing is observed centralization of other industries might take place. The character of overall centralization might likely intensity the density of land use at core locations. Growth, then, might be distributed such that land rents in crease more nearer the core area locations as for the move ment from RR to PP in Figure 8. The implicit assumption thus far has been that the land rent and wage rate gradients 92 Land Rent Distance from Core Figure 9. Shift in land rent gradient with major growth away from core shift in the same manner with growth and decentralization. The empirical results suggest that wage rates increase more at more distant locations. If the land rent gradient shifts in the same manner, shifts should occur from R'R' to P'P1 as in Figure 9. A more complete analysis of differen tial changes in factor gradients would have to include the differential substitution effects in the production of services and manufactured goods and the relocation of these industries over time. A further qualification of the test results involves the nature of the sample observations. The movement of firms was not timed and, hence, the time frequency of re location is not known. Even with a fairly even number of 93 movements per year in the sample period, the factor gradients gradually shift upward over time as growth and decentralization occur. No attempt was made to consider the increase in real factor prices due to growth. Only the differential factor price shifts over space and the slope effects of the gradients were examined. The differential factor price shifts over space indicated that labor costs increased relative to capital costs at greater distances. This implies that the wage rate and possibly the land rent gradient shifted over time in response to decentralization activity. The method of observing factor gradients at one point in time does not capture these shifts. In particular, since the observa tions were taken near the latter part of the test period, the gradient slopes might understate the actual slopes at times prior to 1960. The results cannot be taken to measure before and after slope changes, rather, just the slopes at a single point in time. The simple theory of core-dominated cities yields hypotheses that are consistent with dynamic changes of growth and decentralization. While many other things oc cur during the test periods the changes in transportation technologies seem to induce decentralization of manufac turing activity. Transportation costs on goods are less important for service industries. Instead, transportation costs of commuting are crucial for these industries. The 94 simple model presented can be adjusted to explicitly treat differences in industry reaction by allowing differ ences in production functions. This makes the model more elaborate in detail. Its generality allows for such re finements. A simple refinement presented was the dis tinction between labor of different quality. A test for this distinction was made simpler by the concentration of housing by racial and income groups in the Los Angeles area. Such concentration occurred, in part, due to the imposed uniformity of housing construction after World War II. This induced neighborhoods of uniform incomes, ages, family size, and race. This may not be the case in other cities and this type of partitioning may not be useful for other areas. Central city areas have been losing their tax base for many years. Higher income people have moved to distant locations and contribute their property taxes to suburban areas. Many central city areas try to attract industry by building roads and giving special tax rates for central location. The building of roads reduces not only the costs of moving goods, but also the costs of moving people. The result of improved transportation has been a decentraliza tion of industrial activity and a decline in the tax base for core areas. Many central city mayors do not feel that their areas are adequately compensated for facilitating such mobility. An indirect benefit to central areas of 95 the rising costs of fuel might be a retardation of the decentralization process. Some households may choose to live nearer the major employment areas to save on commuting costs. The impact of rising energy costs in terms of in ducing greater spatial concentration can be handled by a model of the type presented in this study. 96 FOOTNOTES 1. John Niedercorn, "A Negative Exponential Model of Urban Land Use Densities and Its Implications for Metropolitan Development," Journal of Regional Science, 11, no. 3(1971): 317-326. 2. Ibid., p. 324. 3. Ibid. 4. Michael Goldberg, Intrametropolitan Industrial Loca tion: Plant Size and the Theory of Production (Berkeley: Institute of Urban and Regional Development, University of California, 1969). 5. Raymond Vernon, "Production and Distribution in the Large Metropolis," The Annals of the American Association of Political and Social Science, 314(November 1957):15-29. 6. Leon N. Moses and Harold F. Williamson, "The Location of Economic Activity in Cities," American Economic Review, 18(May 1968):211-222. 7. Vernon, "Production and Distribution," pp. 15-29. BIBLIOGRAPHY Adams, Gerard; Milgran, Grace; Green, Edward W.; and Mansfield, Christine. "Undeveloped Land Prices During Urbanization: A Micro-Empirical Study Over Time." Review of Economics and Statistics, 50 (May 1966):248-258. American Transit Association. Fare Structures in the Transit Industry. New Yorkrn.p., 1933. Beckmann, Martin. Location Theory. New York: Random House, 1968. Bradfield,- Michael. "A Note on Location and the Theory of Production." Journal of Regional Science, 11, No. 2 (1971):263-266“ Brigham, Eugene F. "The Determinants of Residential Land Values." Land Economics, 41 (November 1965):325-334. California Manufacturers Register. Los Angeles: California Manufacturers Association, 1949-1974. Case, Fred and Gillies, James. "Some Aspects of Land Planning: The San Fernando Valley Case." The Ap praisal Journal, 23 (January 1955):15-41. Directory of Manufacturing. 25 vols. Los Angeles: Los Angeles Chamber of Commerce, 1904-1934. Duncan, Beverly; Sabagh, Georges; and Van Arsdol, Maurice. "Patterns of City Growth." The American Journal of Sociology, 67 (January 1962):418-428. Epps, Richard. "Strategy for Industrial Development." Federal Reserve Bank of Philadelphia Business Re view (November 1966):3-17. Fales, Raymond L., and Moses, Leon N. "Land-Use Theory and the Spatial Structure of the Nineteenth-Century City." Regional Science Association Papers, 28(1972): 49-80. 97 98 Gannon, Colin A. "Intra-Urban Industrial Location Theory and Inter-Establishment Linkages." Geographical Analysis, 5(July 1973):214-244. Goldberg, Michael. Intrametropolitan Industrial Location: Plant Size and the Theory of Production. Berkeley: Institute of Urban and Regional Development, Univer sity of California, 1969. Greenhut, Melvin. Plant Location in Theory and Practice: The Economics of Space, chapel Hill: University o£ North Carolina Press, 1956. Greenhut, Melvin. "An Explanation of Industrial Develop ment in Underdeveloped Areas of the United States." Land Economics, 36(November 1960):371-379. Grey, Arthur. "Los Angeles: Urban Prototype." Land Economics, 35(August 1959):232-242. Grupp, George W. Economics of Motor Transportation. New York: Appleton and Company, 1924. Harris, R. N.; Tolley, G. S.; and Harrell, C. "The Resi dence Site Choice." Review of Economics and Statis tics, 50(May 1968):241-247. Hartwick, John M. "The Pricing of Goods and Agricultural Land in Multiregional General Equilibrium." South ern Economic Journal, 39(July 1972):31-45. Hoover, Edgar M. The Location of Economic Activity. New York: The McGraw-Hill Book Company, 1948. Isard, Walter. Location and Space-Economy. Cambridge: Massachusetts Institute of Technology Press, 1956. Kain, John F. "The Journey-to-Work as a Determinant of Residential Location." Regional Science Association Papers, 9 (1962):137-160. Lee, E. S.; Bresee, J. C.; Nelson, K. P.; and Patterson, D. A. An Introduction to Urban Decentralization Research"! Oak Ridge: Oak Ridge National Laboratory, T57T!-- Lowry, Ira. "Filtering and Housing Standards: A Conceptual Analysis." Land Economics, 36(November 1960):362- 370. 99 Meyer, John; Kain, John; and Wohl, Martin. The Urban Transportation Problem. Cambridge: Harvard Univer sity Press, 1'96$. Moses, Leon N. "Location and the Theory of Production." Quarterly Journal of Economics, 72(May 1958):259-272. Moses, Leon N. "Towards a Theory of Intra-Urban Wage Differentials and Their Influence on Travel Pat terns." Regional Science Association Papers, 9 (1962):53-63. Moses, Leon N., and Williamson, Harold F. "Value of Time, Choice of Mode, and the Subsidy Issue in Urban Trans portation." Journal of Political Economy, 71(June 1963):247-264. Moses, Leon N., and Williamson, Harold F. "The Location of Economic Activity in Cities." American Economic Re view, 18(May 1968):211-222. Muth, Richard F. "Economic Change and Rural-Urban Land Conversions." Econometrica, 29(January 1961):l-23. Muth, Richard F. "The Spatial Structure of Housing Market." Regional Science Association Papers, 7 (1961):208-526. Muth, Richard F. Cities and Housing. Chicago: University of Chicago Press, 1969. Niedercorn, John. "A Negative Exponential Model of Urban Land Use Densities and Its Implications for Metro politan Development." Journal of Regional Science, 11, no. 3 (1971):317-326. Pegrum, Dudley F. Transportation: Economics and Public Policy. Homewood: Richard D. Irwin, Inc., 1963. Schwartz, Lawrence E. "From Core to Ring: An Econometric Study of the Relocation of the Central Office and Its Workers." Ph.D. Dissertation, Harvard University, 1963. Smith, Hadley E. "A Review of Plant Location Theory and a Questionnaire Study of Plant Location in Los Angeles County." M.A. thesis, University of Southern Calif ornia, 1954. 100 U.S. Department of Agriculture. Bulletin No. 910. Washington, D.C.: Government Printing Office, 1920. U.S. Department of Commerce. Census of Population and Housing; Census Tracts. Vol. 5TI Washington, D.C.: Government Printing Office, 1960. Vernon, Raymond. "Production and Distribution in the Large Metropolis." The Annals of the American Association of Political and Social Science, 314(November 1957): 15-29. von Thunen, Johann Heinrich. Von Thunen's Isolated State. Translated by Carla N. Wartenberg. Edited by Peter Hall. Edinburgh: Pergamon Press, 1966. Weber, Alfred. Theory of the Location of Industries. Translated by Carl J. Friedrich. Chicago: University of Chicago Press, 1929. Wingo, Lowdon. "An Economic Model of the Utilization of Urban Land for Residential Purposes." Regional Science Association Papers, 7 (1961):191-205. Wingo, Lowdon. Transportation and Urban Land. Baltimore: Johns Hopkins Press, 1961.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

On The Theory Of Value And Market Syndicalism

PDF

Oskar Lange'S Contributions To The Theory Of Socialism, With Applicationsto Yugoslavia And Czechoslovakia

PDF

Socialist Price-Formation Models And Input-Output Analysis: A Study Of Hungary, 1959-1964

PDF

A Theory Of Regional Economic Growth: Growth Poles And Development Axes

PDF

Inter-Metropolitan Area Budget Cost Determinants: An Econometric Analysis

PDF

Some External Diseconomies Of Urban Growth And Crowding: Los Angeles

PDF

The Usefulness And Limitations Of Accounting Reports For Testing The Theory Of The Firm: An Appraisal

PDF

Pollution, Optimal Growth Paths, And Technical Change

PDF

The Development Of Western-Style Economic Planning: Theory And Practice

PDF

Regional Growth In The United States: A Production Function Approach

PDF

A Multidimensional Interpretation Of Comprehensive Developmental Planning

PDF

On The Dynamics Of Planned Economy: General Theory With Empirical Analysis Of The Hungarian Experience

PDF

The Theory Of Investment Programming: A Suggested Econometric Model

PDF

Cross-Sectional Analysis Of Manufacturing Production Functions In A Partitioned 'Urban-Rural' Sample Space

PDF

Differential Game Theory Approach To Modeling Dynamic Imperfect Market Processes

PDF

A Positive Approach To Advertising And Publicity As A Function Of Consumer Preference

PDF

A Generalized Economic Derivation Of The ''Gravity Law'' Of Spatial Interaction

PDF

A Constrained Price Discriminator: An Application To The U.S. Post Office

PDF

Appraisal Of Developmental Planning And Industrialization In Turkey

PDF

A Money Supply Model: Jordan

Asset Metadata

Creator Macreynolds, William Kenneth (author)

Core Title Contributions To The Theory Of Industrial Location: Urban Decentralization

Degree Doctor of Philosophy

Degree Program Economics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Economics, theory,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Gordon, Peter (committee chair), Brown, Alan A. (committee member), Niedercorn, John H. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-748242

Unique identifier UC11356698

Identifier 7501070.pdf (filename),usctheses-c18-748242 (legacy record id)

Legacy Identifier 7501070.pdf

Dmrecord 748242

Document Type Dissertation

Rights Macreynolds, William Kenneth

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA