TRIP LENGTHS AND TRAVEL COSTS: THEIR RELEVANCE TO RESIDENTIAL THEORIES AND MODELS; A CASE STUDY: LANSING, MICHIGAN Thesis for the Degree of M. U. P. MICHIGAN STATE UNIVERSITY JAMES M. Mc GRATH 3.970 .III!IIIIIIZIIIIIIILIIIIIIIIIIIIIIJIIIIIII f 7‘ ’5 W 41 meg: ; ' 75¢’l.-.wH-G -- i 0;; ”were: *3? £9":- ‘ twig; «Hf-ER“ Fifi! um RY ." )I‘ {3" . ’xf (Ii e ("fie-F 1 14m: 0; A r21“. N . t ABSTRACT TRIP LENGTHS AND TRAVEL COSTS: THEIR RELEVANCE TO RESIDENTIAL LOCATION THEORIES AND MODELS; A CASE STUDY: LANSING, MICHIGAN by James McGrath Residential location theory has been slowly evolving over a period of years to the point where certain elements of it have found their way into practice. Much of this has taken place while the planning profession was in its infant stages. It is at present still a young and volatile profession and consequently finds itself lacking established benchmarks or reliable predictive models on the residential location decision. In order to strengthen its reliability, the profession must consolidate, test, and prove its body of theory in regard to residential location. It is because there is such a need for research efforts in this field that this case study for Lansing, Michigan, was under- taken. In this work an attempt is made to examine the transport cost function as it relates to the decision of household location. There is an attempt to show the evolution of residential location theory and its bias towards the transport cost factor. The work then examines the latest and most pre-eminent of residential loca- tion theories, that of William Alonso. There follows an isolation of his main transport cost assumption and two correlary assumptions about lot sizes and land values and an outline for empirically testing the applicability of these to one city: Lansing, Michigan. Results are presented and analyzed and evaluated in the context of Alonso's theory. Finally other directions in residen- tial location theory are examined and evaluated in light of the test findings and some new directions for research are suggested. ACIQIOWLEDGMENE This being the appropriate place, I would like to take the opportunity to express my appreciation for the material aid and guid- ance so generously given to me by Professor Effat Mansour. Much of the data and the direction of research was supplied by Professor Man- sour as were innumerable hours of his valuable time. Without such aid, it is doubtful whether such a research undertaking would have been possible. The many hours that went into the typing and retyping of the first draft of this thesis aided its completion in no small way. For this I would like to extend my appreciation to Miss Carol'Morey. For the prodigious effort of proof-reading and final typing I owe thanks to Mass Judith Herron. Finally I would like to thank.those members of my thesis committee who, along with Professor Mansour, sat in on the final pre- sentation of this work. For this I thank.Professor Sanford Farness of the Department of Urban Planning and Professor James Wheeler of the Department of Geography. TRIP LENGTHS AND TRAVEL COSTS: THEIR RELEVANCE T0 RESIDENTIAL LOCATION THEORIES AND MODELS; A CASE STUDY: LANSING, MICHIGAN By James McGrath A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER IN URBAN PLANNING Department of Urban Planning and Landscape Architecture 1970 hapter LI L2 L3 LI L5 L6 prter TABLE OF CONTENTS Chapter One uuwwwwww o. mbuNNNNI—I 0 Introduction . . . . . . . . . . . . . . . . . . . Urban Land Allocation Theories - Theoretical Roots Recent Explanatory Theories. . . . . . . . . . . . 'Measuring the Cost of Travel - Choosing a Method . The Case Study Area. . . . . . . . . . . . . . . . . Summary. . . . . . . . . . . . . . . . . . . . . . . Two Introduction . . . . . . . . . . . Methods of Travel Costs. . . . . . . . . . . . . . . Delineating Trip Types . . . . . . . . . . . . . . Statistical Method . . . . . . . . . . . . .-. Multiple Centers of Urban Activity. . . . . Relationship of Travel Costs to Distance from the C. B. D. - Lansing Case Study Results . . . . . Relationship Between Residential Land Costs and Distance from the C.B.D.. . . . . . . . . . . Land Values - Their Effect on Lot Size . . . . . Test Method and Hypothesis Formulation for the . Examination of Lot Size . . . . . . . . . . . Test Results . . . . . . . . . . . . . . . . . . . Land Values - Ways of Measuring and the Formulation of a Test Hypothesis. . . . . . . . Test Results . . . . . . . . . . . . . Conclusions and Summary. . . . . . . . Three Recapitulation and Major Conclusions . . . . . . . . Analysis and Application of the Empirical Findings . Diagramatic Solution of Individual Equilibrium . . . Land Values as a Determinant of Residential Location Lot Size as a Determinant of Household Location. TheHousingPackagE. o e o o o o 0-0 s o s o o 0 New Directions in Residential Location Theory . Location Theory and the Practitioner . . . . . . 23 24 29 31 33 35 36 37 39 41 41 47 49 51 52 58 59 68 70 75 Al’s all; If.) ID ?I Appendix - Methods of Computations and Measures Traffic Zones . . . . . . . . . . . . . Trip Types . . . . . . . . . . . . . . Measuring Travel Costs . . . . . . . . Land Values . . . . . . . . . . . . . . Lot Sizes . . . . . . . Distance from.the C.B.D. . . . . . . . Decentralization . . . . . . . . . . . “@WDWNH Bibliography 78 78 79 80 80 80 81 82 ;_ r ~40 , 3 I. ‘23 ’1?) "1 Illlii: Timed mile: 5 r51a: ..her 1 faith j Ill 1": (fl CHAPTER ONE 1.1 Introduction Urban Planning, like most disciplines, finds its day-to- day practices rooted in a theoretical base. Therefore, theory- building, testing, and the synthesizing of empirical findings are viewed as essential activities for systematically advancing the knowledge of the phenomena under study, i.e. urban areas. Although a relatively young discipline, which has borrowed greatly from other fields of study, planning has little in the way of theory which it may truly call its own. This, however, is not extraor- dinary as Stuart Chapin points out: A relatively young field such as planning normally faces a long period of "finding itself" before it begins to develop schools of thought and acquire theoretical tradi- tions. Under the most favorable conditions, theory build- ing is a slow process, particularly so if it is to be sys- tematic and rigorous by contemporary standards of scientific inquiry. The time involved from.the first statement of theory, through the intermediate stages of empirical tests of its critical components, the successive reformulations and retests, to the stage when it becomes operational, is painfully long. It is so long that the components of theory often pass into operational use long before the full system can be made to work in practice. Chapin's words succinctly apply to the systematic develop- ment of residential location theory. The roots of present theory 1Chapin, F. Stuart, Urban Land Use Planning, Univ. of I11. Press, Urbana, Illinois, p. 69. a the pace :J‘IESE any 15:111- teed mu. ;0 1‘3: can be found as far back as 1826, when Johann von Thunen2 developed a theory of agricultural location based on distance from the market place. From this rather meager beginning, this chapter shall outline the development of residential location theory up to its present-day status. It will be possible to trace the influence of many early statements relative to urban spatial structure and view the manner in which these statements have carried down into the most recent formulations of residential location theory. This chapter shall also demonstrate that the majority of recent efforts in the field have tended toward the development and refinement of theories based on economic principles and shall consequently at- tempt to explain why economic thought has predominated. Having achieved this, it will explore those aspects of residential loca- tion theory which have passed into operational use as well as the nature of their impact. Finally, this chapter will delineate the need for inquiry and testing and point out the reasons for select- ing Lansing, Michigan, as the study laboratory. 1.2 Urban Land Allocation Theories - Theoretical Roots As mentioned in the introduction, von Thunen developed a theory of agricultural rent based on nearness to the market place. In brief, he viewed agricultural rent at any location as being 2von Thunen, Johann H., Der isolierte Staat in Beziehung auf Landwirtschaft und Nationalekonomde, lst vol. Hamburg, 1826; and new ed., Hamburg, 1863. ,_. <1 rm 33k: a—r H 5—: rn equal to the value of its product minus production and trans- portation costs. In early works, such as von Thunen's, very little mention is made of urban land. It was not until the turn of the century that any ac- tive thought within the field had sprung up in this country. In a study of New York in 1903, Richard M; Hurd3 was the first to utilize concepts drawn from land economics to treat the subject of urban land. His book entitled Principles of City Land Values outlined a theoretical framework for urban land values which closely approximated the principles espoused by von Thunen a half century earlier: As the city grows, more remote and hence inferior land must be utilized and the difference in desireability be- tween the two grades produces economic rent in locations of the first grade but not in those of the second. As land of a still remote and inferior grade comes into use, ground rent is forced still higher in land of the first grade, and so on. Any utility may compete for any location within the city and all land goes to the highest bidder.... Practically all land within the city earns some economic rent, though it may be small, the final contrast being the city's rentless and hence strictly speaking, valueless circulference.4 He summarizes by saying: Since land value depends on economic rent, and rent on location, and location on convenience, and convenience on nearness, we may eliminate the intermediate steps and say that value depends on nearness. 3Hurd, Richard'M., Principles of City Lgnd Values (New York: the record and guide), 1903, p. 1. 4Ibid., p. 11-12. 51bid., p. 13. Although he does give credence to other variables, the concept of accessability clearly stands out in his work: Underneath all economic laws, the final basis of human activities is psychological so that the last stage of analysis of the problems of the structure of cities turns on individual and collective tastes and preference, as shown in social habits and customs. He fails to include this important psychological factor in his final theoretical framework even though he did point out its influence. It is interesting to note, however, that during this period of time when little attention was given to individual social-psy- chological motivation, people such as Hurd, despite their failure to categorize these variables in a meaningful way, nevertheless perceived their importance to the individual decision making process of residential location. He turns the focus back to economics say- ing: Where residences contain more than one tennant, the basis of value is economic and conforms closely to the prin- ciples governing business property. Hurd felt that these economic principles lead to a spatial configuration which would approximate concentric rings around the city center, each succeeding one being of lesser density until a point is reached at which a valueless circumference occurs. 61bid., p. 11-12. 7Ibid,. p.78. Robert M; Haig8 took Hurd's principles of accessability and translated them into transportation costs which he then linked to location rents: Rent appears to be the charge which the owner of a rela- tively accessable site can impose because of the savings in transport cost which the use of the site makes possible.9 In a review of Haig's work, EffatMansour10 points out that the attempt to save on transport costs is viewed by Haig as a determi- nant, although not the only one, underlying the site location decision: Site rentals, according to Haig, are also important.... he as- serts that: 'While transportation overcomes friction, site ren- tals plus transport costs represent the social costs of what friction remains....the sum of the two items, the cost of fric- tion, is not constant....on the contrary it varies with the site. The theoretically perfect site for the activity is that which furnishes the desired degree of accessability at the low- est costs of friction.‘ From these insights Haig deduced that minimizing the costs of friction is the organizing principle which determines the layout of the metropolis. ‘He explains how these economic principles operate in the choice of a residence: '....one buys accessability precisely as one buys clothes or food. He considers how much he wants the contacts furnished by the central location weighing the "costs of friction" involved -- the various possible combinations of site rent, time value, and transportation costs; he compares this want with his other de- sires and resources, and he fits it into his scale of consumption, and buys.’ In other words Haig views residential location as a strictly economic problem.in which the selection of a place of residence is determined by two criteria: (a) accessability to the 8Haig, Robert M. , "Towards an Understanding of the Metropo- lis - The Assignment of Activities to Areas in Urban Regions", Quarterly Journal of Economics, 40, May, 1926, p. 402-434. 91bid., p. 420-421. 1oM’ansour, Effat, Towards an Inter-Urban Residential Location Theogy - An Inquiry into the Influence of Social Factors on the Location of urban Households, Early draft of an unpublished Dissertation for the degree of Doctor of Philosophy in City and Regional Planning, University of Pennsylvania, 1969. as. of Ru SOI 51. :he wei 31rd an coutrib " ' I mg 5 Si i1 te an center city and (b) minimized costs of friction in terms of less transportation costs and site rentals. Unlike iHurd, he does not mention the need for considering the social basis for residential location or, for instance, the possible impoiEance of other criteria such as the size of the site. Richard Ratcliff,12 writing in the late 1940's, lent the weight of his experience to furthering the views of both Hurd and Haig. His attempt to both systematize and clarify the contributions of Urban Land Economists included a restatement of Haig's concepts: Site rents and transportation costs are complimentary as illustrated by the fact that, as the distance to the cen- ter of the city increases, transportation costs increase and rent declines, though the sum.of the two is not con- stant... Each land use seeks the lowest total of"fent and transportation costs. The opportunity to minimize the total is limited for any one land use by the competi- tion of lower uses, which are willing to bid up to the savings in transpiration costs that the use of the site will permit them. 3 7’ h a.- Because of its highly accessable nature andaits convenience to the greatest number of people, Ratcliff felt that the central core becomes a very desirable location for various competing uses that ‘wish to capitalize on the locational advantages offered there. He apparently perceived the convenience offered byha site as the prime motivation in the decision to locate. Carrying this a bit further, 111b1d., Chap. 11,11. 4-5. 12Ratcliff, Richard U., Urban Land Economics,‘MoGrawiHill Co. , New York, N.Y. 13Ibid., p. 372. 7 he differentiated between forms of convenience which attract dif- ferent consumer types to various locations surrounding accessabi- lity hubs: For certain uses one type of convenience may be highly essential, another type a matter of indifference. In the case of residential use, the focus of space relation- ships is outward, for the convenience of the householder is the determinant of economic rent.1 Ratcliff's concept improves somewhat on Haig's simple economic analysis of the selection of a place of residence. ‘He recognizes the influence of social factors on residential location as he states: The disutilities of travel must be evaluated against the whole system of values of socail groups who are involved in travel. The differences among many groups may be 11- lustrated by the considerations that are controlling in the selection of a place of residence. Families with children seek proximity to schools and to healthful and wholesome surroundings often at the expense of convenience to the place of work. Some groups with strong religious ties tend to cluster about the church. Newly married couples without children prefer central locations that are convenient to the work place of both the husband and the wife and to amusement centers. Not only do systems of convenience desirability vary among groups, but also ... with time, for customs change and new activities take on new importance.1 Following the same pattern as those economists who have preceeded him, Ratcliff gave some recognition to social factors, but in the final analysis retreated to the more easily definable economic arguments. ‘He follows Haig's lead in emphasizing trans- portation costs as the major determinant of urban spatial structure, 14Ib1d., p. 375. 15Ibid., p. 373. is... .1 3 , L .. .... ...... . .9! 5.3%.: asserting that "the structure of the city is determined through the dollar evaluation of the importance of convenience."16 He does not, however, explain how the various factors identified may enter into this "dollar evaluation of convenience." Mansour points this out in his review: If the most convenient site is the city center, and the most convenient arrangement is based on minimized aggregate costs of friction in terms of land values and transportation costs to the center -- without any consideration for other factors such as accessability to social and physical amenities or site size -- then the resultant pattern of land uses will tend to be clustered around the center in extremely high density.17 1.3 Recent Explangtory Theory WilliamAlonso18 sees such an unrealistic pattern resulting from any of these general theories which we have discussed thus far. It is his contention that the economist's total disregard for considera- tion of site size is at fault and that this seriously weakens the premises of their theories. If one was to accept the framework of any one of these theories then he might conceivably find that all consumers would choose to locate around the center on very small lots which have mini- mal combinations of site rentals and transport costs. In the normal course of events, however, this is not empirically the case, and it would therefore seem in order that a revision of the content and extent of-these early general theories be undertaken. 16Ibid., p. 375. 17Mansour, Op. cit., p. 6-7, Chapter 2. 18Alonso, William, Location and Land Use, Harvard University Press, 1964. .vi .— .—_.—- 1.3 Recent and so too, refining ec filonso' s W theory as ' ! Ink has b {mtributi tion in pa 1.3 Recent Explanatogy Theogy Haig and Ratcliff elaborated on.Hurd's economic principles and so too, William.Alonso reconstructed their works while further refining economic theory for location and land use. In addition, Alonso's work draws heavily on von Thunen's agricultural rent theory as reconstructed by Edger S. Dunn and Walter Isard. Alonso's work has been widely aclaimed as one of the most important recent contributions in location theory in general, and residential loca- tion in particular. In his work, Alonso proffers a classical economic theory of market equilibrium which is achieved wherein the optimum location for each consumer is found. Alonso contends that this state is reached through the interaction of individual bid-price curves within the market mechanism. He then offers as postulates, three functional relationships which delineate the substitutability between the three locational goods that he sees consumers demanding in his theoretical land market. When these three functions are combined, they then form what he considers to be a preference, which he maps in the form of a three dimensional indifference surface. The indifference surface rep- resents the set of combinations for the three locational goods from which a given individual will be able to extract the same composite level of satisfaction regardless of changes in the quantity of any of the three variables. These variables are as follows: 1. The quantity of land (q): it is here that Alonso becomes first in the field to consider the size of the site as being material to the location decision of the household. 10 2. Distance from the city center (t): this is basically a proxy variable by which Alonso reflects the disutility of travel to the center. The cost of travel, however, is defined both in travel time and money costs as well as the physiological inconvenience of travel. 3. A composite good (2): this is a catch-all variable for all goods and services which the consumer may demand other than (2) and (t). According to Alonso an individual's equilibrium is reached when his locus of opportunities surface touches one of the indiffer- ence surfaces. It is where this point of contact occurs, that an individual's'bid-price curve may be found. This curve represents the set of prices for land that a given individual can afford given the fact that he must maintain a constant level of satisfaction. In his review Mansour defines in general terms the linkage between behavior of the individual consumer and the market economy in general: ...when firms and residents enter into a bidding process by which the most efficient users determine the use of the land as well as the question of who locates there. The most desirable location is the center of the city, and land is acquired for use by the consumer who has the steepest bid-price curve, i.e., the highest bidder per unit of land at a certain location. Using marginal sub- stitution among the three goods and the bidding process, each consumer finds an equilibrium.position for himself, and the market establishes an equilibrium among bidders which determines the structure of land values as well as the most efficient use of the land.19 19Mansour, Effat, Op. cit., p. 6-7, Chapter 2. 11 In the construction of his theory, Alonso has set forth a very restrictive set of assumptions which are general in nature, abstract, and largely unrealistic. These assumptions include cer- tain preconditions such as: featureless geography; free flowing transportation in all directions; a state of perfect competition; uniform.services and tax rates; full knowledge by all of market information at all times, and uniform desires for certain things such as: nearness to the C.B.D., more land, and a quantifiable composite good. I In addition to the distance and land quantity factors, a typical household would consider many other factors, but given the present state of the art in location theory, it wghld be pre- tentious to attempt to identify,let alone quantify, all of these other considerations. Two of Alonso's assumptions stand out as being particu- larly important: 1) the assumption relative to the price struc- ture of land, and 2) the assumption relative to the costs of travel which Alonso implies in his assumed disutility of distance from the city center. In his first assumption Alonso contends that the price per unit of land decreases in a steeply declining logarithmic func- tion. The shape of the function, shown in figure 1, implies sig- nificant decreases in price as distance from the city center increases. 12 FVt) O 't Figure l The assumed shape of the function is very critical to Alonso's theoretical framework. The shape is "a requirement for the existance of both individual and market equilibrium as well as essentially true for most cities."20 In his second assumption Alonso forwards the view that distance from the city center represents disutility which increases significantly with that distance. This relationship is expressed by the steeply rising function shown in Figure 2. Since Alonso associates the disutility of distance with travel costs, this would imply that as distance from the center city increased, travel costs would also increase significantly. K KIt) Figure 2 0 t 20Alonso, Op. cit., p. 20. 13 These assumptions, coupled with his assumption that indi- viduals prefer more land, represent a critical weakness in the ap- plication of Alonso's theory to reality. Mansour points this fact out: If both functions happen to be very flat, namely if land values do not vary or decrease significantly with distance from the center and if transport costs do not increase sig- nificantly with distance, then Alonso's "locus of opportuni- ties surface" will tend to flatten and so would the "prefer- ence surface" as individuals become indifferent between widely separated locations. The resulting market equili- brium, if such an equilibrium is still possible within the same framework, will protray unrealistically wide spread cities. 1 We note in this chapter that certain independent empiri- cal studies have shown weaknesses in Alonso's travel cost assump- tion, and we shall explore this further in a later section of this chapter. In spite of the many possible problems which may arise with the high degree of abstraction employed by Alonso to treat the subject of urban spatial structure, the framework he constructed is nevertheless an elegant and useful one. Alonso moves beyond his predecessors in economic location theory by attempting to in- ject the reality of a multi-factor economic world into a theoreti- cal framework, treating such things as purchasable social amenities, and goods with negative utility. Alonso's economic formulations, indicating that residents seek to maximize satisfactions in locating rather than to minimize 21Mansour, Effat, 0p. cit., p. 13, Chap. 2. 14 costs of friction, effectively enable him to disprove Haig's mini- mum.costs of friction hypothesis. Since the issue here centers around transportation costs, Alonso shows that technological im- provements in transportation mainly accomplish two things: 1) make commuting easier, and 2) make commuting cheaper. Alonso apparently feels that these technological innovations shift the individual bid- price curves in such a fashion as to reflect lower land prices at the center and higher land prices in the city frontier than would have existed previously. This is shown in Figure 2. He also con- tends that when transportation costs are lowered, an individual will seek more land to a point where the aggregate rent remains the same. Only when an individual does not desire more land does he then enjoy lower aggregate rents as a result of improved accessability. ‘Many ecological theorists explain the peripheral location of the wealthy in American cities by referring to a process of rapid growth and change, whereas Alonso feels that he can explain the peripheral location of the wealthy by: ...a strong appetite for land, so that the holdings of land vary greatly with income, the wealthier are affected rela- tively less by the costs of commuting because they spread the costs over larger sites. Consequently the rich are price-oriented, whereas the poor are location-oriented. Less accessability being bought with increasing income, accessability behaves as an inferior good. This is a possible explanation of the paradox encountered in Ameri- can cities, of the poor living on expensive central land and the rich on cheaper peripheral land.22 22Alonso, William, Location and Land Use, Harvard Univer- sity Press, 1964, p. 109. 15 He presses home his argument for individuals desiring more land as a possible disproof of the growth and change theories of the ecologists: In large American cities the higher income residents locate further away from the center of the city than those of lower incomes. The principle explanation of- fered has been that, as the city grows, the wealthier build their houses on the vacant land on the periphery, leaving their old houses near the center to the poorer. This explanation is based on the growth of the city's population. However, in a city which does not grow, the same pattern of peripheral rich and central poor obtains.... If the peripheral location of the rich is due to increases in population and to the aging and ob- solescence of more central housing, a governmental program of central redevelopment which provide downtown luxury apartments would lead, after some time, to centrally lo- cated rich and lower income people in the periphery, in the then aged suburban housing. If the view presented here is correct, and tastes do not change, the centrally located luxury apartments will satisfy only the limited demand of the rich with rapidly decreasing marginal rates of substi- tution and steep bid-price curves, and the pattern of cen- tral poor and peripheral rich will prove lasting.23 While Alonso's theory does have implications beyond those of an economic realm, it nevertheless does attempt to explain residential location mainly as being determined by travel costs and quantity of land. It is just that, the dependence on these two variables, that render the greatest weaknesses to his theory as Mansour points out: ...its major weakness in explaining residential location stems from its focus on treating it as determined by travel costs and quantity of land. Because his catch-all variable (2) has to be independent of any locational implications, 23Ibid., p. 109. 16 none of his other goods can have locational qualities. For instance (2) cannot account for the social and physical ameni- ties that are usually tied to specific locations. Also, his resulting land values are based on a simple price structure for land which does not account for the location within a prestige area, nor the type of house, services available, living space, etc., which are inextricably tied to the resi- dential lot. The treatment of land as the main focus of analysis, rather than just one of the factors, makes it dif- ficult to translate his theory into an operational form that could be tested. Likewise the treatment of the residential lot as a simple economic quantity, bleached from any social implications, mghes any explicit ties to ecological theories rather tenuous. As Stuart Chapin has pointed out, many portions of theory go into practice long before the entire theory ever becomes opera- tional, and so too it has been the case with residential location theory. Lowden Wingo, Jr.,25 combining the theories of land econo- mics with a theoretical analysis of traffic flows, has created an explicit mathematical model of the residential land market which has greatly influenced many transportation and land-use simulation models. To overcome the generalities of Alonso's theory and make his work more operational, Wingo explicitly limits himself to an economic treatment of residential location in which he introduces more specific assumptions such as: 1) less travel is preferred to more travel, and 2) more residential land is preferred to less 24Mansour, Op. cit., Chap. 2, p. 13. 25Wingo, Lowden, Jr., Transportation and Urban Land, Re- sources for the Future, Inc., Washington, D. C., 1964. 17 land regardless of location. The sole basis of transportation costs or savings in Wingo's model is the journey-to-work, the technological link between the labor force and the production process. His feeling is that, in terms of order and relative volume, the journey-to-work is the most signifi- cant class of person movement in any urban region. Although Wingo is concerned with the economics of location, his concentration on the journey-to-work results in his differentiat- ing socio-economic household groups as they locate with respect to the distance separating them from their place of work. Therefore, if one accepts Wingo's model as being representative of the great majority of urban development models, he can see that beneath their complex mathematics, many are based on a simple notion of accessa- bility that assumes new residential growth areas will be located as closely as possible to employment centers in order to minimize transportation costs. Michael Stegman26 feels that this logic arose because: Residential growth models are second generation offspring of transportation models which naturally view the major metropolitan daily traffic flow - journey-to-work - as the key to location decisions.2 The emphasis on the journey-to-work as the main determi- nant of residential location has prevailed in most urban land 26Stegman,Mithael, "Accessability Models and Residential Lo- cation," Journal of the American Institute of Planners, Vol. 35, #1, January , 1969 0 27Ibid., p. 22. I: I 18 simulation models, including those of Britton.Harris28 and Ira Lowrey. In Harris's work one finds as a basic assumption: ..residential locators in the urban land market seek to achieve an equilibrium.between their expenditures for space for residential accomodations, their ease of access to other opportunities, especially employment, and their ex- penditures on all other commodities. It is especially em- phasized that the utilities from space and accessability are a joint outcome of the locational decisions....the pos- session of land in the central locations confers a trans- portation advantage and consequently this land enjoys higher rents in direct relation to the savings on transportation costs and inconvenience.29 Using the "willingness to travel function" as the single determinant of location, Lowrey's model deploys the worker popula- tion into an existing housing stock, allocating them according to the most accessable zones around a given cluster of basic employ- ment. Given the amount of land available for residential develop- 'ment and specified density constraints, the model successively fills each "most accessable" zone and then allocates the overspills to the most accessable zones which have not yet reached their present den- sity limits. Since large metropolitan housing stocks generally form con- centric rings of residences in progressively decreasing densities outward from the core, the model can be calibrated or adjusted to reproduce an actual development pattern, but as Stegman points out: 28Harris, Britten, Basic Assumptions for the Simulation of the Urban Residential Land Market, Institute for Environmental Studies, University of Pennsylvania, Philadelphia, 1966. 291bid., p. 4-5. 19 ...it tells us very little about the locational decision process of housing consumers. As a matter of fact, one might infer from the model that since accessability is such a valued good, only the fortunate and affluent should gain control of the inlying residential land while the poor are pushed outward into the fringe. Since contrary empiri- cal evidence exists, the only other conclusion we might draw from the model is that the outlying worker populations choose to consume more land at a lower unit cost and pay a higher price in terms of greater costs and inconveniences of commuting to and from work and other activities. The discussion has now come full circle back to the basic premises of Alonso. Stegman expresses his opinion as to their real value: The theoretical principle involved more accurately reflects the existing pattern of residential development than the process by whigh housing consumers move about within the fixed pattern. 1 The points that have been discussed obviate the need for a thorough investigation of the basic transport cost assumption as it has applied to most residential location models up to this time. Examining this assumption and its relationship to the value and quantity of land assumptions requires that one should measure in some way travel costs and the manner in which they vary with distance from.the C.B.D. Having accomplished this, one might attempt to deter- mine what effects these have on land values and consequently the amount of land purchased as suggested by Alonso's theory. 1.4 Measuring the Cost of Travel - Choosing a Method Over a period of years various experts in the transporta- tion field have developed procedures for measuring the cost of travel. 3OStegman, Op. cit., p. 24-25. 311b1d, p. 24-25 20 Iskander I. Gabbour,32 in his study of trip length variations within urban areas, reviewed these methods in considerable detail and selected the bestvavailable measure. Essentially Gabbour was concerned with measuring the varia- tions in travel costs rather than examining the actual costs theme selves. 'This required an evaluation of the various indices of travel costs, and out of these, determining which best reflect the costs in- curred in getting from origin to destination. Gabbour's conclusion was that trip length, measured in travel time, is the most realistic measure of travel costs available. His actual study consisted of observations drawn from the transporta- tion studies done in five different cities, focusing on the varia- tions in travel cost with respect to different types of trips by purpose and mode of transport. Essentially Gabbour found that in small size cities, the average trip length increased as the distance between the place of origin and the center of the city increased. He analyzed trip lengths with regard to density and distance, showing that in the small size city, distance was the predominantly influential variable due to the centralization of activities in a dominant C.B.D. He also found a uniform density gradient prevailing as one moved farther from the C.B.D. Since the scope of Gabbour's study was limited to only five cities, it would seem imperative that further studies be undertaken to verify the credibility of its conclusions. Furthermore, since 32Gabbour, Iskander 1., Travel Cost Variations and the Size of Urban Areas, A dissertation for the degree of Doctor of Philosophy in City Planning, University of Pennsylvania, 1967. 21 Gabbour's findings pertinent to medium size cities were based on only two case studies, there is an obvious need for further empirical studies of similar size citieS. Essentially, Gabbour's tests were implicit tests of Alonso's transport cost assumption and therefore Lansing, Michigan, as a case study becomes doubly significant: one, as a further study of trip length behavior in medium sized urban areas; and two, as a means of providing a further test of the travel cost assumption put forth in many popular land allocation theories, the most pre-eminent of which has been espoused by William Alonso. 1.5 The Case Study Area The Lansing metropolitan area is a medium sized area of approximately 160,000 persons and is well suited to the type of study undertaken by Gabbour. As recently as 1965 the Tri-County Regional Planning Commission of Lansing, Michigan, undertook a transportation and land use study which served to provide the data necessary for this study. The Lansing metropolitan area also appears well suited for a real world examination of Alonso's transport cost assumption with its correlary land value and lot size assumptions. It fits the description of Alonso's theoretical city in that it has no major barriers geographically, is relatively flat in contour, and appears to have a characteristically strong Central Business District. With these facts established, the study proceeded along the conceptual lines described in Gabbour's study with one exception: the Lansing study concentrated solely on home origin trips and is therefore 22 oriented toward residential location theories and models, which are nevertheless based on the accessability concept. 1.6 Summapy In summary this chapter has traced the slow evolution of classical economic location theory from its roots in Germany over a century ago to its most recent form as forwarded by William.Alonso. Further, it has been noted that certain elements of resi- dential location theory have influenced planning practices, parti- cularly those concerned with transportation and urban land simulation models. In addition, this chapter briefly examined the concepts underlying two prototype simulation models and heard one critic suggest that the theoretical principle involved -- that of minimiz- ing transport costs, reflects the existing pattern of residential development rather than the process by which the housing consumer moves about within the established market. It was concluded that a systematic examination of Alonso's transport cost assumption was necessary as well as some examination of his correlary land value and lot size assumptions. Finally, the effort of one transportation expert to study travel cost variations in urban areas was briefly reviewed and found to be adequate in methodology for application to the Lansing metro- politan area. It was therefore adopted with some minor changes. ‘I'I‘H CHAPTER TWO 2.1 Introduction Building upon the background and foundations established in the preceding chapter, the second chapter finds as its purpose, the description of testing methods employed in the Lansing Case Study and the presentation of the empirical findings derived therefrom. As mentioned previously, Alonso's land value and lot size assumptions were tested as well as his travel cost assump- tion. The methodology for testing all three will be explained. 'Methods of measuring travel costs will be examined and related to previous tests of similar natures. The construction of test hypo- theses will be described along with the criteria for their accept- ance or rejection. Finally the test findings will be presented and evaluated in relation to the established criteria for hypo- thesis acceptance. Having accomplished this, conclusions will be drawn as to the applicability of Alonso's travel cost assumption to the Lansing Metropolitan Area and consequently his Residential Location Theory. A comparison will also be made between the find- ings of this study and those found in Gabbour's study. 23 ,l. C 3...“... .. n, . .... u. . .u I. do 24 2.2 Methods of Measuring Travel Costs Despite the fact that Alonso treats three variables in his Theory of Residential Location, he places greater emphasis on distance from the city center. This, he feels, is of considerable influence in any residential location decision. This bias, that households trade off accessability for cheaper land, can best be summarized by referring to Alonso's explanation: Given any combination of land and distance, a small increase in distance will produce dissatisfaction and will have to be compensated for by a small in- crease in the quantity of land for satisfaction to remain the same. A key determinant of residential location, according to Alonso, is the increase in commuting costs one must bear as dis- tance between the home and the C.B.D. increases. Because of these increasing commuting costs, Alonso con- tends that land values decrease with distance from the C.B.D., and that lot sizes increase as consumers strive to maximize their satisfactions by taking advantage of cheaper land prices at greater distances from the C.B.D. Essentially these are direct assumptions made by Alonso. It is his contention, as shown previously, that land values decrease 1Alonso, William, Location and Land Use, Harvard University Press, 1964, pp. 106, 109. 25 with distance from the C.B.D. It is also his contention that people desire more land and that as a consequence lot sizes will increase with distance. In order to test the validity of his primary assumption, namely, that distance brings with it greater costs of travel, one would not need to measure the costs of travel themselves, but rather to find out whether people do in fact make longer trips on the average, as distance between their place of residence and the city center increases. This in itself would provide a surro- gate measure of travel costs. This particular approach to the problem of measuring travel costs was developed by Iskander Gabbour in his study of trip length variations within urban areas. Essentially Gabbour was concerned with measuring the variations in travel costs rather than examining the costs themselves. Examining the nature and components of travel costs, Gabbour concluded that trip length measured in travel time was a valid surrogate measure of travel costs. His actual study consisted of observations drawn from transportation studies done in five dif- ferent cities. The focus of his study was on the variations in trip lengths, as indices of actual travel and mode of transport. Gabbour cites various ways in which people calculate their costs of travel. He explored these costs from.the point of view of the traveler rather than from an objective viewpoint, mainly because H 4 ~1- Ili‘, It? - “was. s . l‘ i w, . i. 1. ...HI. r .... 26 it is the way in which people perceive their costs of travel that will make them act in one way or another. One such measure is the out-of—pocket costs for transportation such as: gas, oil, and parking for the auto traveler; and direct fare paid for the mass transit rider. He reflects on other costs incidental to the auto traveler such as repairs, depreciation, and parts replacement, explaining that these are part of the incurred costs of auto travel, but that the average individual overlooks these in the calculation of his travel costs. Another measure of costs explored by Gabbour is the amount of time spent in transit. It is pointed out, however, that for any trip, there is a certain amount of time that is tolerated after which the value of time rises sharply. He suggests that the rela- tionship between value and time appears to be a non-linear one, and therefore may defy attempts to be translated into monetary terms. Speaking along these very lines, Lowden Wingo, Jr.,2 contends that time has a value relative to the income of the individual, and therefore each individual has a certain tolerance limit as to the amount of time that he is willing to spend commuting after which the value of the individual's time rises sharply. He suggests that time is a commodity that increases in value as income increases until it reaches a point at which it becomes a real constraint on all decisions 2Wingo, Lowden, Jr., Transporggtion and Urban Land, Resources for the Future, Inc., 1961. 27 to act. Gabbour acknowledges this fact, and the difficulty of assigning a dollar value to units of travel time, but neverthe- less emphasizes the importance of travel time by citing the fact that in work and school trips time has been stipulated as the ele- ment to which people have indicated the most sensitivity. Other elements to which travelers are sensitive were expressed as being of a more qualitative nature such as comfort and convenience. Gabbour concludes, therefore, that these fac- tors are at least as significant as the monetary costs involved in urban travel. However, the inability to quantify these factors, at least at the present time, precludes their use in the measure- ment of travel costs. In summary, Gabbour concluded that there are certain diffi- culties in estimating some of these travel costs, especially those of a more qualitative nature, and that such difficulties would be multiplied whenever an aggregation of costs is attempted. He there- fore felt that a partial avoidence of the complexities involved in estimating these costs could directly and satisfactorily reflect travel costs. Trip lengths was the chosen variable. Gabbour con- cluded, after examining its weaknesses, that these weaknesses are much smaller in comparison to those associated with the actual costs of travel. He therefore launched into an investigation of the various means for measuring trip lengths, examining their advantages and dis- advantages to determine how accurately these measures reflect trip length. 28 The first and most obvious measure that he examined was that of distance measured in miles from the point of origin to the point of destination. Such distance can be measured either by air- line distance or over the transportation network. Previous studies have indicated a stable relationship existing between the two means of measurement, and therefore one could be substituted for the other. Trip length measured in miles does not account for the effect of speed of travel from origin to destination, nor does it give any indication of the possible congestion evident in some loca- tions. A more informative measure of trip length, he felt, is the time spent in getting from the point of origin to the point of desti- nation. ‘Measurement by travel time does have problems of its own, namely, that it does not account for the time spent in waiting, as in the case of public transportation. It should be pointed out that many studies have recognized this problem and therefore add this amount onto the actual travel time. Another problem with reported travel time lies in the fact that its accuracy rests upon the accur- acy of the person reporting, and it has been frequently observed that there is a tendency to round off the travel time and report it in intervals of five, ten, or fifteen minutes. One further problem is that a trip, when made at different times of the day or by a dif- ferent mode of travel, will result in different travel times. This problem is directly related to problems of volume and congestion which influence the duration of the journey-to-work. u“ 29 Gabbour states the underlying assumption in most trans- portation studies: ...trips being made on an average day will accurately reflect the time when each trip has to be made as well as the combination of facilities and means of travel available to the traveler under the general conditions prevailing throughout the region.3 Given these two assumptions, the reported travel time could be con- sidered as the actual best or minimum travel time required for each trip. 2.2.1 Delineating_Trip Typgs Gabbour's study detailing his findings on trip length variations within five urban areas designated twelve different types of trips for analysis. However, only four of these were of home origin and therefore of concern to residential location theories. These four types are: 1. Total home-origin auto trips 2. Total home-origin transit trips 3. Home to work auto trips 4. Home to work transit trips 'Minor reflection on these four trip types obviates the special atten- tion given to the work trip. The journey to work receives a special designation for reasons best outlined by Gabbour in his study: ...early studies on trip length have focussed on the journey to work. The selection of that trip type for detailed analysis stems from the consideration of several factors that ascribe to work trips special importance among all trips undertaken in urban areas. First, work trips are highly ordered in the 3Gabbour, Iskander 1., Travel Cost Variations and the Size of Urban Areas, A.dissertation for thé’degree of Doctor of Philosophy in City Planning, University of Pennsylvania, 1967, p. 132. 3O sense that they are institutionalized in the community and that they are undertaken under specified time limits a factor that gives rise to their high concentrations at two peak periods during the day. 'Moreover, trips for work purposes range from one-sixth to one-third of all vehicular trips recorded in internal area surveys, with an average value or about one-fourth the total.4 He further cites the work of K. K. Liepmann in investigating the significance of the journey-toawork: Liepmann puts the relationships between the place of work and place of residence in a broad perspective by acknow- ledging the important role that these relationships play in the nature of community life. She concludes that: '...there are certain advantages in persons living near their work, and for many this will be a desirable arrangement. For others, however, daily traveling between the place of residence and work will be preferable not the least, in the interest of maintaining family life.... The aim.must be to enable the journey to fulfill its important functions while keeping the movement within such limits as to avoid excessive costs, un- reasonable hardships, and damage to community life.'5 Wingo adds to the emphasis given to the journey-to-work in his economic conceptualization of it: "For the worker, the time consumed by the 'work trip is a true cost - time is a valuable commodity which must be 'spent' if the trip is to be made."6 It logically follows then that in any transportation study, regardless of its underlying assumptions, there will be a major effort given over to the study of the work trip as one indicator of travel habits. 4Gabbour, 0p. cit., p. 132. 51bid, p. 132. 6Wingo, Op. cit., p. 52. 31 In treating other trips which originate from the home, Gabbour made no distinction between types, mainly because these are not so highly ordered in nature, nor are they made with the regu- 1arity or frequency that is so characteristic of the work trip. He therefore lumped them into an all-purpose category and treated them as a single variable. As mentioned previously, Gabbour found that in small size cities the average length of trips increased as the distance of the origin from the center of the city increased. Thus, a research ef- fort of this nature in addition to testing Alonso's assumption about travel costs will serve as another empirical case StUdY 0f trip length variations from which further conclusions may be drawn about the na- ture of these variations in small size cities. 2.2.2 Statistical Method To determine whether a relationship does exist between travel time and distance from the C.B.D. as well as the nature and strength of this possible relationship, Gabbour employed a statis- tical technique of correlation and regression analysis. Such a method enables the researcher to derive the coefficients of a speci- fied function, be it linear, parabolic, or otherwise, and relate a dependent variable (travel time) to an independent variable (distance from the C.B.D.) in order that the described function (the relation- ship between the two) would provide the best fit to the observed data on these variables. The Lansing study involves measuring trip length varia- tions as distance from.the center of the city increases. Trip lengths are measured in reported travel time,and distance is measured in miles e 32 from the C.B.D. The mode of testing is essentially the same as that employed by Gabbour, that of a correlation and regression analysis. In keeping with scientific method, we formulated a set of test hypotheses; a null hypothesis and the alternative hypothesis. They are as follows: 1. That no positive relationship exists between distance from the C.B.D., as measured in miles, and reported travel times. 2. That a positive relationship does exist between dis- tance from the C.B.D., as measured in miles, and re- ported travel times. The criteria for accepting hypothesis number one was that Beta, the slope of the line, be equal to or less than zero. The accept- ance of hypothesis two would be contingent upon the rejection of hypothesis number one. In the event that hypothesis number two should be accepted, it would be necessary to qualify our conclusion and determine whether the relationship is in fact a reasonably dependable one. This neces- sitated that we establish two contingent hypotheses which are as fol- lows: 1. That a relationship does exist between distance from the C.B.D., as measured in miles, and reported travel times, but that the relationship is not a reasonably dependable one. 2. That a relationship does exist between distance from the C.B.D., as measured in miles, and reported travel times and that this relationship is a reasonably dependable one. The criteria established for accepting hypothesis number one was that R2, the coefficient of determination, which indicates 33 the percentage of the observed variations in the (Y) dependent variable which can be attributed to observed changes in the indepen- dent variable, be equal to or less than 0.4999. The acceptance of hypothesis two was contingent upon the rejection of hypothesis number one . 2.2.3 Mpltiple Centers of Urbpn Activity Gabbour's tests assumed the existance of a single center ' of urban activity, and although this is consistent with Alonso's general theory, Alonso nevertheless treated the possibility of mul- tiple centers of urban activity in a later section of his book.7 In his section entitled "Applications and Outline for Research," Alonso reconsiders the concept of the single city center: The concept of the center of the city may be refined as well. Thus far we have assumed a singlg all-purpose cen- ter, but other forms may be considered. He then discusses the prospects of treating neighboring centers of equal and unequal sizes, pointing out that in both cases there will be a mutual competition by residential prospects of each center for that land which falls between the two centers as illustrated in Eigure 3. C5? AA HORIZONTAL VIEW VERTICAL VIEW Figpre 3 7For more detailed explanation see Alonso, Qp. cit., p. 130-142. 8Alonso, Op. cit., p. 134. 34 The possibility of just such a case in the Lansing area be- came abundently clear after some initial tests of decentrality were undertaken. We therefore deemed it necessary, before proceeding any further, to define as closely as possible the extent of the rela- tive influence of each center. Alonso suggested an approach utilizing bid rents for sites mapped in contour lines over the city.9 This approach, although a good one in theory, would require extensive data on the sale prices of land for the entire area for a given period of two or three years. This type of data for the Lansing area, as will be pointed out later, is not available in sufficient quantity to allow one to make generali- zations from it. This consideration prompted us to search for a method which could employ readily available data on an area-wide basis, and render a representative picture of the relative attraction of each center within the area. The distribution of employment opportunities in an urban area is a function of the distribution of economic activity and as such could be thought of as a measure of the potential drawing power or attraction of a particular center. Working under the assumption that any substantial center in the area would be a gravitational center for surrounding employ- Inent, we then mapped, in a contour fashion, the employment densities (number of jobs per census tract) for the metropolitan area which re- ‘vealed where the centers of employment were located as well as the extent of their influence. Figure 4 shows the relative magnitude 9Alonso, 9p. cit., p. 138. a go: Flddm MI - >> fl .mCZmA—A. ..Q.m.o GZHmZHZD H8.o oo.o o~.m ~n.e we.n ao.n on.~ on.. um. mo. 0 a I a d J 1 la a 1 a a d d l d 4 oo 5 e e 0 0 o. o o o o o. 0 not! A 1 II J I e I a . Oowm . anmH “no: im.:~ canal A a.» u < em.P u m mm.o u we now «.2 we at mqu: I Q.m.o tomh moz